JP6906507B2 - Selective sampling to evaluate structural spatial frequencies using a specific contrast mechanism - Google Patents

Description

Reference of related application This application is a US provisional patent application named SELECTIVE SAMPLING MAGNETIC RESONANCE-BASEED METHOD FOR ASSESSING STRUCTURAL SPITAL FREENCIES A, filed on October 7, 2015, and filed on October 7, 2015. It relies on the priority of US Provisional Patent Application No. 62 / 382,695, filed September 1, 2016, named ASSESSING STRUCTURAL SPATIAL FREFQUENCIES WITH SPECIFIC CONTRAST MECHANIMS. This application is a partial continuation of US Pat. No. 14,840,327, filed August 31, 2015, and is now US Pat. No. 936,738, May 27, 2016. It further relies on U.S. Patent Application No. 15 / 167,828 of the application. US Pat. US Provisional Patent Application No. 62 / 064,206, filed October 15, 2014, with the same name, and US Provisional Patent Application No. 62, January 25, 2015, named MICRO-TEXTURE CHARACTERIZATION BY MRI. Relying on the priorities of the specification 62 / 107,465, their disclosure is incorporated by reference. U.S. Patent Application No. 15 / 167,828 Patent specification, METHOD FOR ASSESSING STRUCTURAL SPATIAL FREQUENCIES USING HYBRID SAMPLING WITH LOW OR INCREASED GRADIENT FOR ENHANCEMENT OF VERY LOW NOISE SELECTIVE SAMPLING WITH NO entitled GRADIENT 2016 March 2 application It further relies on the priority of US Provisional Patent Application No. 62/302,577. All referenced applications are assignees in common with this application, and their disclosure is incorporated herein by reference.

The methods claimed herein relate to the field of diagnostic evaluation of fine textures in biological systems for pathological assessment and disease diagnosis, and in material and structural assessment in industrial and engineering research. More specifically, the present invention provides the k-value associated with the spatial composition of the texture of biological tissue with respect to the k-value in a narrowly associated neighborhood with a low tilt, with the tilt of the MRI apparatus turned off. Use a method for repeated measurements. This allows the texture of the tissue to be assessed on a time scale of about milliseconds, which allows the patient's movement problems to be ignored. This method enables in vivo assessment of the diagnosis and monitoring of changes in texture induced in tissues by disease and treatment. Typical technical goals are 1) to be able to assess bone health and fracture risk to assess changes to trabecular architecture due to bone disease, 2) eg liver, Assessment of fibrosis development in soft tissue disorders such as lung and heart disease, and 3) in neurological disorders such as various forms of dementia, or as in, for example, traumatic brain injury (TBI) and chronic traumatic encephalopathy (CTE). Changes to microstructures for characterizing and monitoring abnormal neurological conditions such as brain injury and downstream neurological conditions, or autism and schizophrenia. Applications of other medical conditions include assessment of changes in blood vessels, such as in the vascular network surrounding the tumor, or as associated with the progression of CVD (cerebrovascular disease), and assessment of changes in the ducts in response to tumor growth. The present invention also requires application in the evaluation of microstructures for a range of industrial purposes, such as the measurement of material properties in geology for manufacturing or characterizing various types of rocks, and the measurement of microstructures / textures. Has other uses.

Subtle texture changes in tissues have long been recognized as the earliest markers in a wide range of diseases, but robust clinical assessment of subtle textures remains obscure and the time required to obtain data. The main difficulty arises from the blurring caused by the movement of the object through.

Early and accurate diagnosis is important for successful disease management. Although clinical imaging provides a lot of information about the condition, many of the tissue changes that occur as a result of the onset and progression of the disease, or as a result of treatment, are on a very fine scale, often up to tens of microns. be. Changes in the texture of fine tissues have long been recognized by diagnosticians, including radiologists and pathologists, as the earliest precursors to a wide range of diseases. , Remained out of the capabilities of current imaging technology. For example, the differential diagnosis of obstructive lung disease depends on the symptoms of texture in the lung parenchyma, but the robustness of computed tomography (CT) measurements of early disease is limited. The cancellous bone microstructure, a determinant of fracture risk of aged bone, also remained confusing due to image blurring due to patient movement during scanning with magnetic resonance (MR) imaging. .. Post-processing analysis of MR images may be used to try to identify the texture of the surface of structures such as tumors and white matter. (DRABYCZ, S., et al .; "Image texture characterization the discrete orthogonal S-transform"; Journal of Digital Imaging, Vol. 22, No. 6, 2009. characterization from MRI and CT scan images by multi-resolution analysis ”; 29th Annual International Conference of the IEEE Engineering The effect is limited because it does not deal with the underlying problem, the movement of the subject. (MACLAREN, J. et al .; "Measurement and direction of magnetic head motion daring magnetic resonance imaging; magnetic resonance imaging of the brain", PLOS / ONE, Nov. imaging: a review; Magnetic Resonance in Medical, Vol. 69, 2013.)

The main causes of movements affecting MR imaging are pulsatile movements of the heart, respiratory-induced movements and monocontractions. The first two are quasi-periodic, and the usual approach to it is gating in the slowest phase of motion. However, even with gating, there is sufficient variation between acquisitions, which causes a loss of spatial phase coherence at high k values of interest for texture measurements. This problem is exacerbated by the fact that movements may not be perfectly periodic and often result from multiple causes. Since monocontraction is rapid and induces random displacement, it is not possible to maintain coherence at high k values of interest when measuring texture.

Positron emission tomography (PET) provides valuable diagnostic information, but cannot achieve resolutions below about 5 mm and relies on the use of a radioactive tracer for imaging and an X-ray beam for positioning. It raises dose concerns, especially when repeated scans are required. (BERRTNGTON DE GONZALEZ, A. et al .; "Projected cancer risk from Computed Tomographic scientists performed in the United States in 2007"; It is extremely costly and requires a cyclotron nearby. A resolution of CT up to 0.7 mm is theoretically possible, but this resolution is obtained at high radiation doses and tends to be reduced by patient movement during a few minutes of scanning time. Non-negligible risks from associated radiation doses make CT problematic for long-term imaging and limit achievable resolution. X-ray digital resolution is limited, along with serious dose concerns. This is because the resulting 2D image is a composite of absorption throughout the thickness of the tissue through which the beam has passed. Current clinical diagnostic methods for diseases that are the object of the methods claimed herein are difficult to obtain sufficient in vivo resolution or accuracy. In some cases, there is currently no definitive diagnostic method. For other medical conditions, especially in the chest and liver, the diagnosis is biopsy-dependent, the risk of complications and even mortality is not negligible, and reading and sampling errors tend to be high. be. (WELLER, C; "Cancer detection with MRI as effective as PET-CT scan but with zero radiation risks"; Medical Daily, Feb. 18, 2014.)

Bone health is impaired by aging, among other causes, by bone cancer (as a side effect of cancer treatment), diabetes, rheumatoid arthritis, and as a result of malnutrition. Bone disease affects more than 10 million people a year in the United States alone, adversely affecting quality of life and shortening life expectancy. The current diagnostic criterion for assessing bone health is bone mineral density (BMD) as measured by dual-energy X-ray absorption (DEXA) projection techniques. This modality results in surface bone mineral density that integrates attenuation from both cortical and cancellous bones, similar to standard X-ray imaging mechanisms, but intraosseous bone, which is the most relevant marker for bone strength. Only limited information is provided on the beam structure. (KANIS, J. AND GLUER, C; "An update on the diagnosis and assessment of osteoporosis with densitometry"; Osteoporosis International, Vol. Densitometry, and vertebral fractures in male osteoporosis ”; JBMR, Vol.15, issue1,2000.) BMD correlates only loosely with fracture risk. A post-processing technique, TBS (Cancellous Bone Score), has been attempted to correlate with changes in pixel gradation in DEXA images to provide information about bone microstructure. Comparative studies determined that BMD in the hip joint was still a better predictor of fracture. However, TBS does not provide a detailed assessment of trabecular structure. (BOUSSON, V., et al .; "Travelular Bone Score (TBS): available knowledge knowledge, clinical relief, and future projects"; Osteoporosis International. of the spine assed by TBS related to future neck fracture? A Spanish case-control study ": Osteoporosis International, Vol. 24, 2013.) TBS is still evaluated in Osteoporosis International, Vol. 24, 2013.

Measurements of bone microstructure, specifically trabecular spacing and trabecular element thickness, require a resolution of approximately 0.1 millimeters. MRI, ultrasound imaging, CT, and micro-CT all fit this problem. MRI makes it easy to obtain high contrast between bone and bone marrow, but it takes a long time to acquire an image with sufficient resolution to characterize the trabecular network, and the patient moves during that period. The resolution is limited. The finer the texture size of this network, the greater the blurring due to movement. Attempts to mitigate the effects of patient movement by looking only at the skeleton skeletons that have been removed from the cause of cardiac and respiratory movement have been tried using both MRI and micro-CT. However, the correlation between the microstructure of the limb bone and the microstructure of the central bone is unknown. In addition, a large data matrix, and thus a long acquisition time, is still required to obtain sufficient image information to determine trabecular space and element thickness. This long acquisition time changes the level of motion-induced blurring, depending on patient compliance-monocontraction remains a serious problem even when measuring limbs. Proposed MR-based technology, fineSA (JAMES, T., CHASE, D .; "Magnetic field grade data characteristic sussential usage one dimensional (ID) spatial-frequent US (April 26, 2011) provides a much smaller data matrix of continuously finely sampled, one-dimensional frequency-encoded acquisition data, which are then combined to reduce noise. By acquiring, we are trying to avoid the problem of patient movement. In this case, the image is reduced to one dimension, reducing the size of the acquired data matrix and thus reducing the acquisition time. However, tilt-encoded echoes have very low signal-to-noise (SNR), which requires noise averaging. The advantage of resolution is obtained by this method for 2D and 3D imaging, but many repeating spatially encoded echoes over several response times (TR) for signal averaging. Acquisition time is about a few minutes due to the need to acquire-too long to be immune from movement. Therefore, the improvement in resolution that can be obtained by this technique is limited.

There is a need for accurate, robust, non-invasive, in vivo measurements of trabecular spacing and trabecular element thickness that can assess the bone of the central skeleton. Because they are key markers for assessing bone health and predicting fracture risk. So far, no clinical technique has been able to bring about this ability.

Fibrotic disease occurs in response to widespread biological injuries and injuries in internal organs, and the development of collagen fibers is a healing response of the body. In more advanced fibrous diseases, the density of fibers in the affected organ is high. Fibrotic pathologies range from lung and hepatic fibrosis to heart fibrosis and cystic fibrosis, pancreatic fibrosis, muscular dystrophy, bladder and heart disease, and a number of myelofibrosis disorders in which fibrous structures replace the bone marrow. appear. Fibrotic progression is associated with several cancers, including breast cancer. Different progressions are seen in prostate cancer, where the disease destroys healthy, organized fibrous tissue. In all cases, texture spacing that is emphasized in tissue changes in response to disease progression as collagen fibers occur along the underlying tissue structure. In liver disease, the texture wavelength changes because the texture of healthy tissue in the liver is replaced by a longer wavelength texture that results from the "decoration" of the lobular structure of the organ. In other organs / diseases, texture changes reflect upsets in healthy tissue with the progression of textures that indicate fibrous intervention.

Assessment of texture changes from fibrotic progression requires 0.1 mm scale resolution to cover the extent of disease progression in most fibrotic pathologies. One of the most common such conditions, liver disease, is a representative example in which the fibrous structure is difficult to assess. Currently, the typical criteria for assessing a condition are tissue biopsy-a non-negligible complication frequency-and mortality-at risk of highly invasive and often painful procedures (patient biopsy). You need to stay in the hospital for hours to overnight for later observations), and sampling errors are prone and reading variability tends to be large. (REGEV, A;. "Sampling error and intraobserver variation in liver biopsy in patients with chronic HCV infection"; American Journal of Gastroenterology; 97,2002.BEDOSSA, P.et al;. "Sampling variability of liver fibrosis in chronic hepatitis C Hepatitis, Vol. 38, issue 6, 2004. VAN THIEL, D. et al .; Another modality with ultrasound, which is often used to assess, can only provide a proper assessment later in the disease-it is used to diagnose cirrhosis. Magnetic resonance-based elastography (MRE), which has been under development for some time to be used in the evaluation of liver disease, cannot be evaluated early-reading error before significant fibrous infiltration (advanced disease) is too large. Further, this technique is a very costly procedure because it requires the presence of additional expensive hardware, skilled technicians, and takes a total of about 20 minutes for configuration and scanning time. The ability to image fibrous textures directly by MR imaging is impaired by both the patient's movement over the time required to obtain the data and the lack of contrast between the fibers and the surrounding tissue. Data acquisition during a single breath-hold is severely impaired by the pulsatile movement of the heart and breath-hold failure, which causes significant movement in many organs such as the liver and lungs. Also, since the SNR is low enough, motion correction by combining re-registered MR intensity profiles obtained from continuous echoes is extremely problematic. Similarly, the assessment of the amount of cardiac fibrosis in early disease using MRI is severely disturbed by heartbeat during the time of measurement. Unlike Gaussian noise, motion cannot be averaged due to its non-linear effect-it must be at a sufficient signal level to allow re-registration before averaging to reduce electronic noise. More sensitive (higher signal-to-noise) non-invasive techniques are needed to be able to assess texture changes throughout the range of fibrotic progression from onset to advanced pathology in order to be able to diagnose and monitor response to treatment. It is said that.

The onset and progression of numerous neurological disorders is associated with repetitive, microscopic changes in the structure / texture of nerves and blood vessels. However, the ability to assess such changes in the brain is only available after death. At present, the definitive diagnosis of Alzheimer's disease (AD) is by postmortem histology of brain tissue. Due to other forms of dementia such as AD and Levy body dementia, motor disorders such as muscular atrophic lateral sclerosis (ALS), Parkinson's disease, chronic traumatic encephalopathy (CTE) and other medical conditions or trauma. Conditions caused by traumatic brain injury (TBI), such as those with damage to brain structures such as multiple sclerosis (MS), cerebrovascular disease (CVD), and other neurological disorders are in advanced stages. Often diagnosed only by changes in behavior and memory, which impedes the ability to intervene early. In addition, conditions such as epilepsy and autism have been associated with abnormal changes in subtle neural structures, which, if clinically diagnosable, allow target selection to test response to treatment. Will do.

Various in-vivo diagnostic techniques are available for AD and other dementias, but none of them are definitive. These techniques range from written diagnostic tests, which tend to result in large evaluation errors, to PET imaging, which assesses amyloid plaque density or glucose metabolism (FDG PET). As mentioned above, PET imaging is extremely expensive, cannot provide high resolution, depends on the use of radioisotopes and X-ray beams for positioning, and is licensed for long-term use due to dose concerns. Make it complicated. Moreover, neither amyloid imaging nor FDG PET have been shown to be definitive indicators of AD. (MOGHBEL, M. et al. "Amyloid Beta imaging with PET in Alzheimer's disease: is it faceible with currant radiotracer.

The use of CSF biomarkers for the diagnosis of dementia is painful and highly invasive, and indistinguishable signal levels by anatomical location in the brain as possible with imaging biomarkers. This is a serious drawback of using a liquid biopsy, as various forms of dementia have been found to have different spatial / temporal progression in the brain. Another disorder associated with various forms of dementia is CVD (cerebrovascular disease), which induces cognitive impairment as a result of reduced blood flow by blocking blood vessels leading to brain tissues. .. There is a need here for high-resolution assessment of pathologically-induced changes in microvessels.

Tissue shrinkage due to atrophy in many forms of dementia, including AD, can be measured in MRI by carefully registering long-term acquired data, but for diseases, this shrinkage is measurable. By the time it becomes, it is progressing. Early stages of the disease are indicated in postmortem histology by degradation of the column sequences of cortical neurons, and the normal spacing of these columns is about 100 microns in most cortical regions. (CHANGE, S. et al .; "Microanatomical cortex of cognitive availability and declination: normal aging, MCI, and Alzheimer's disease"; Cerebral. alien cognitive cortex of the human brain ”; Journal of Clinical Neuroscience, 15, 2008.) Due to the size of this texture and the extremely thin cortex, data can be collected with very little patient movement, even at excellent acquisition speeds. Make it impossible. Evaluation of texture changes on the order of tens of microns is extremely problematic in vivo, but when possible, it can target a range of subtle texture changes in the diagnosis and monitoring of neurological disorders, and is important in the development of treatments. Play a role.

Another possible neurological application of the claimed method is to determine in vivo the boundaries of various control regions of the cerebral cortex or the different Brodmann areas in which they are contained. Such capabilities greatly support data interpretation in the study of brain function, such as performed using FMRI (Functional Magnetic Resonance Imaging).

The three types of diseases mentioned above, bone disease, fibrotic disease, and neurological disease are not a comprehensive list. Other pathologies with changes induced by microstructured pathologies, such as angiogenic growth of the vascular system surrounding the tumor, or changes in the vascular system and ducts in response to fibrosis and tumor progression in the breast are also pathological. The ability to resolve fine tissue textures allows for early detection of disease and monitoring of response to treatment.

The ability to measure changes in fine textures will be of great value for the diagnosis of disease. Non-invasive techniques that do not rely on the use of ionizing radiation or radioactive tracers leave a great deal of discretion in early diagnosis and repeated measurements, allowing monitoring of disease progression and response to treatment. Magnetic resonance imaging (MRI), which provides tuned tissue contrast, is just such a non-invasive technique, with no radiation dose concerns. However, in order to avoid the problem of signal degradation due to patient movement, data needs to be taken on a time scale that was not previously possible.

A selective sampling method for assessing tissue texture using magnetic resonance (MR) is by applying a contrast mechanism that enhances the contrast between component tissue types in the polyphase biological sample being measured. Achieved. Therefore, the volume of interest (VOI) is selectively excited with a plurality of time-varying radio frequency signals and applied slopes. Encoding A gradient pulse is applied to induce a phase wrap to create a spatial encoding for a particular k-value and orientation, which is determined based on the texture in the VOI. .. A time-varying series of acquisition slopes is initiated to produce a time-varying locus through the three-dimensional k-space of the k-value encoding, although the resulting k-value set is required to produce an image of the VOI. It is a subset. Multiple series of samples of NMR RF signals encoded using the k-value set are recorded simultaneously. The recorded NMR signal sample is then post-processed to give rise to a signal-to-k value dataset for the k value within the k value set determined by the locus to characterize the tissue within the VOI.

The features and advantages of the embodiments disclosed herein are better understood by reference to the following detailed description when considered in conjunction with the accompanying drawings.

This is a simulation showing the number of data samples required for averaging to achieve an output SNR ≥ 20 dB, depending on the input SNR. This is a simulation showing the number of data samples required for averaging in order to achieve SNR ≧ 20db depending on the position in k-space. FIG. 6 is an exemplary timing diagram of a pulse sequence for a patented method showing the timing of a single TR. It is an enlarged view of the exemplary timing diagram of FIG. Claimed to obtain multiple measurements of a selected set of k-values, where different numbers of samples are taken at each k-value so as to prevent the decrease in energy density at increased k-values. It is an example of a timing diagram for the method to be done. It is shown that the ability provided by the claimed method to obtain many repetitive target k values within a single TR allows robust signal averaging to increase SNR. It is a simulation. It is a simulation showing the result of trying to obtain 90 samples to be averaged using the conventional frequency encoded echo method, and the acquisition of a specific k value with only a few iterations is a long recording for each echo. Due to time, it is possible in each TR. It is an exemplary timing diagram of a patented method designed to result in data acquisition across multiple refocused echoes within a single TR. Two possible shapes of interest acquisition volume (VOI) are shown. FIG. 6 is an exemplary timing diagram of a pulse sequence for a patented hybrid method showing the timing of a single TR. It is a detailed view of the hybrid element of the enlarged scale method. It is a further detailed view of the very low SNR acquisition mode part of FIG. It is a further detailed view of the low SNR acquisition part of FIG. It is a further detailed view of the high SNR acquisition part of FIG. FIG. 6 is an exemplary timing diagram of a pulse sequence for a patented hybrid method showing data acquisition within a single echo. It is a further detailed view of the very low SNR, low SNR and high SNR acquisition part of FIG. It is an exemplary timing diagram of a pulse sequence for low SNR acquisition. It is a further detailed view of the low SNR acquisition mode of FIG. It is an exemplary timing diagram of a pulse sequence for high SNR acquisition. It is a further detailed view of the high SNR acquisition mode of FIG. It is a graphical description of healthy bone structure, osteoporosis bone structure. It is a graphical description of fibrous tissue in the liver. It is an exemplary timing diagram of a pulse sequence that performs the first diffusion contrast. It is a timing diagram of the pulse sequence which carries out the second diffusion contrast. It is a graphic description of VOI dispersed in fibrous tissue. It is a graphic description of the cortical mini-column in the brain. It is a graphic description of the arrangement of VOI in the brain. It is a display of three histological structural images showing progressive pathology with the progression of AD. It is an exemplary representation of the placement of VOI in the cerebral cortex and the application of multiple slopes to the k value.

Hereinafter, definitions of terms used in the present specification are provided.
180 ° reverse pulse An RF pulse that inverts the spin in the tissue region to allow the MR signal to be refocused.
180 ° pulse Tilt the net magnetic field vector _{antiparallel to B 0} RF pulse 90 ° pulse Tilt the net magnetic field vector to the _{cross section with respect to B 0} RF pulse 3T 3 Tesla A / D analog-digital converter AD Alzheimer Disease ADC Diffuse-enhanced imaging average diffusion coefficient adiabatic pulse excitation Adiabatic pulses are a type of amplitude- and frequency-modulated RF pulses that are relatively insensitive to the six heterogeneity and frequency offset effects. ..
AWGN Additive White Gaussian Noise Additive White Gaussian Noise (AWGN) is a fundamental noise model used in information theory to mimic the effects of many naturally occurring random processes.
Biopsy A biopsy is a sample of tissue taken from the body for further examination.
C / N contrast vs. noise, image quality measurements based on signal differences between structural elements rather than overall signal level CAWGN Complex Numeric Additive White Gaussian Noise CBF Cerebral Blood Flow Chemical Shift A molecular environment with different nuclei that contributes to MR signals A slight change in MR resonance frequency due to.
CJD Creutzfeldt-Jakob disease crusher tilt A tilt applied to both sides of a 180 ° RF refocus slice selection pulse to reduce the spurious signal produced by pulse imperfections.
The CSF Cerebrospinal Fluid DEXA dual-energy X-ray absorption method is a means of measuring bone mineral density using two different energy X-ray beams.
DSC Dynamic susceptibility Contrast DTI Diffusion-weighted imaging DWI Diffusion-weighted imaging Echo An RF pulse sequence in which a 90 ° excitation pulse is followed by a 180 ° refocus pulse to eliminate the effects of field heterogeneity and chemical shifts during echo.
Frequency encoding Gaussian noise, where the frequency encoding of spatial positions in MRI is achieved by using an auxiliary magnetic field induced by a mechanical tilt coil, is a probability equal to the probability density function of a normal distribution, also known as a Gaussian distribution. Statistical noise with a density function (PDF).
Mechanical magnetic field gradient pulsing gradient set to change the encoding of the gradient pulse k-value MR scanner primarily used to spatially encode a signal or to set a specific phase lap in a selected direction. A set of coils around the bore of the GRE tilted cold echo interleaved acquisition Signal acquisition from multiple VOIs that are continuously excited within a single TR Isochromat A group of microscopic spins that resonate at the same frequency ..
Two-dimensional or three-dimensional Fourier transform of a k-space MR image.
k-value coefficient The coefficient of the Fourier series or transform that reflects the relative load of each particular k-value in the Fourier series.
Two-dimensional or three-dimensional Fourier transform of a k-space MR image.
k-value One of the points in k-space that reflects the spacing of structural elements in the texture field.
k-value selection pulse A library of gradient pulses k-space values used to select a particular k-space encoding along the sampled direction k-space acquired in a particular region of the tissue for tissue characterization Net Aggregation of Factors Mechanical Inclination Magnetic Field Inclination MRE Magnetic Resonance Elastography-Using Acoustic Shear Waves and Imaging Their Propagation Using MRI, which Can Be Imposed by Using a Set of Tilt Coilers in MR Scanners Imaging technology to measure the stiffness of soft tissues.
MRI Magnetic Resonance Imaging MS Multiple Sclerosis Noise Floor In theory, a noise floor is a measurement of a signal resulting from the sum of all of the noise sources and unwanted signals in the measurement system NMR Nuclear Magnetic Resonance PET Positron emission tomography is a functional imaging technique that uses a positron-radioactive radioactive tracer to produce a three-dimensional image of a functional process in the body.
Phase Coherence (Spatial) When referring to multiple measurements or multiple k values within a common VOI, a particular phase-encoded lateral magnetization vector indicates that the sample has the same position with respect to the reference system. Phase encoding is used to give the phase angle. The particular phase angle depends on the position of the transverse magnetization vector within the phase encoding gradient, the magnitude of the gradient, and the duration of the gradient application.
Phase wrap Helical perturbation pitch of phase of lateral magnetization along phase encoded sample Tightness profile of phase wrap along k value encoding direction One-dimensional plot of signal strength RF radio frequency electromagnetic signal half Crystalline texture Texture slices showing regular spacing along one or more directions Slices (slabs) Slice-selective refocusing RF pulses used indiscriminately to show non-zero-thickness planar parts Spin refocusing by a combination of slice-selective tilt and RF pulse, such that the bandwidth selects thickness along the direction of tilt, and the RF pulse tilts the net magnetization vector from its equilibrium position. Only spins that process at the same frequency as the RF pulse are affected.
SE Spin Echo SNR Signal-to-noise ratio Spoiler tilt See crusher tilt T2 Defined as the time constant for attenuation of lateral magnetization resulting from natural interaction at the atomic or molecular level.
T2 ^{* In} one of the actual NMR experiments, the transverse magnetization decays much faster than predicted by the natural atomic and molecular mechanisms. This rate is indicated by T2 ^* (“T2-asterisk”). The T2 ^* can be considered an "observed" or "effective" T2, while the first T2 can be considered a "natural" or "true" T2 of the tissue to be imaged. T2 ^* is always T2 or less.
TBS cancellous bone score is a technique for looking for texture patterns in the DEXA signal to correlate with bone microstructure to assess bone health. TbTh bone trabecular thickness for bone measurement.
Trabecular spacing for TbSp bone measurements.
Number of trabecular bones for TbN bone measurement.
The TE spin echo sequence has two parameters. The echo time (TE) is the time between the 90 ° RF pulse and the MR signal sampling corresponding to the maximum echo. The 180 ° RF pulse is applied at time TE / 2. The repetition time is the time between the two excitation pulses (the time between the two 90 ° RF pulses).
Texture Frequency The number of repetitions of the texture wavelength per unit length in the texture Characteristic spacing between structural elements in the texture wavelength texture The TR spin echo sequence has two parameters. The echo time (TE) is the time between the 90 ° RF pulse and the MR signal sampling corresponding to the maximum echo. The 180 ° RF pulse is applied at time TE / 2. The repetition time is the time between the two excitation pulses (the time between the two 90 ° RF pulses).
Vector Coupling Tilt The magnetic field tilt VOI of interest volume window function resulting from any vector coupling of a set of tilted coils In signal processing, the window function (also known as the apodization function or tapering function) has several selected intervals. X-ray diffraction, a mathematical function that has zero values outside X-ray diffraction is a tool used to identify the atomic and molecular structures of crystals.

The embodiments disclosed herein allow fine biological textures to be non-invasively measured and evaluated in vivo at high resolution and in response to the onset of disease and the progression of a range of medical conditions. Provide MR-based technology that enables monitoring of formation and / or change. This same method can be applied to the characterization of fine textures in other biological and physical systems. MR-based resolution of fine textures can be scaled to a size scale previously unattainable in in vivo imaging. Although the methods are described herein for biological systems for tissue inspection, of microstructures for a range of industrial purposes, such as the measurement of material properties in manufacturing or in geology to characterize various types of rock. It is equally applicable to evaluation and other uses that require microstructure / texture measurements.

The method claimed herein is the in vivo resolution of fine textures by acquiring essential data fast enough to neglect the effects of subject movement, which is a factor limiting MRI resolution. Achieve this significant improvement in. This fast acquisition is achieved by acquiring data in increments-in a single location, in an orientation, and at one time, or at a selected set of k-values, within a TR. After applying the encoding tilt to select the k-value of interest, the data is acquired with the tilt turned off, the encoded k-value is acquired multiple times, and subsequently to reduce electronic noise. Allows averaging and thus allows robust measurements of individual k-values before motion blurring can occur. Acquisition of TR can be repeated as many times as necessary to make up a larger set of measurements at the selected k-space towards the generation of a continuous spectrum of texture spacing present in the tissue or in the tissue. Change the encoding as needed to extend to the desired range of real space and k-space required. The set of one or more k-values output from each TR is here high SNR due to the ability to average repetitions without the influence of motion, and the interest measurement is an image rather than an image generation. Because of the texture spacing of, we are not interested in the lack of phase coherence between TRs.

In its simplest form, the methods claimed herein are of particular interest, such as lesions, organs, certain parts of an organ, specific areas of bone, or some areas of affected organs for sampling. It consists of acquiring an MR signal from within the internal volume including the tissue region. This internal volume is limited, but to name a few, intersecting slice-selective refocusing, selective excitation using phased array transmission in combination with appropriate tilt, and from tissue outside the region of interest. Of several methods, including adiabatic pulse excitation to scramble the signal, an external volume suppression sequence, and other methods of selectively exciting spins within the internal volume, including physical isolation of the tissue of interest. Can be excited by one of them.

After defining the volume of interest (VOI), multiple samples of the signal centered on a particular k value are obtained, with the tilt turned off and the spread defined by the receiver BW and sampling length. This measurement is repeated only in a particular direction within the VOI, rather than attempting to map the entire k-space as needed to generate the image. One or more samples of a particular k-value are acquired within the acquisition block during a single TR, and the k-value is subsequently incremented or decremented so that other k-values are desired during the same TR. Allows even more samples of. This method allows multiple samplings of each k-value of interest over a time period of several milliseconds, avoiding movement of the subject. The process can then be repeated in additional TRs, and the only condition of movement between acquisitions of the k value is that the VOI remains within the tissue region of interest. The composition of the spatial frequency amplitude spectrum can be achieved without the need for spatially coherent acquisition. Since the amount of interest is the relative intensity of the various k-values (texture spacing) present in the sample volume, any movement between blocks does not impair the measurement as long as the acquired volume remains within a typical tissue sample. While constructing the spectrum of k-values contained in a tissue, the use of fairly robust and real-time pilot and acquisition algorithms is used for movements with an amplitude large enough that the internally excited volume can move to other tissue volumes. It can be used for total repositioning of internally selectively excited volumes, and rejection of datasets that did not stay in the proper tissue.

Repositioning of the VOI, which allows sampling of textures at multiple locations within or across an organ or anatomy, can determine changes in the pathology of an organ. The acquired data can be spatially mapped with respect to the positioning image. The VOI can be driven within a continuous TR, or the interleaving acquisition row within a single TR by exciting an additional volume during the time the signal recovers before the next TR. It is either one of them. The condition is that the continuous VOI is excited with the new tissue and does not overlap with the previous slice selection. Spatial variation of the medical condition can be determined by this method. It can also be used to monitor the temporal progression of organ pathology if the measurements are repeated over time.

Customization of the pulse sequence for the prewind phase in the sample volume positions the highest k value of interest at the echo peak, where the signal is the strongest and provides the best SNR measurement.

Sampling of k-values within the volume under study, along multiple directions at various angles, and along various paths, either linear or curved, is specific to textures, especially mini-columns of neurons. It can give rise to important information about textures that have a semi-ordered structure in the direction. Measurements of the k-value associated with column spacing are extremely sensitive to sampling path alignment, as small changes in sampling direction on either side of the perpendicular indicate a rapid decrease in signal magnitude with respect to that k-value. expensive. Fluctuations in the acquisition path on either side of the signal maximate give rise to pathologically evoked randomness measurements, which can be indicated by the width of the peak.

With the slope turned off for data acquisition, tailoring bandwidth to a particular species can enhance structural information when the chemical composition of the structure under study is known.

The method claimed herein can be used in combination with a time-dependent contrast scheme that targets blood flow. Some of these contrast techniques are blood oxygen level dependent (BOLD) imaging, arterial spin labeling (ASL) imaging, and dynamic susceptibility contrast (DSC) imaging. Because these methods use a variety of techniques to emphasize the vascular system, changes in the texture of the vascular system associated with many pathologies, including CVD (cerebrovascular disease) and tumor growth, can be measured.

The methods claimed herein are also used in conjunction with other MR-based measurement techniques, including DWI and DTI, with front-end information for parameter selection on diffusion techniques, and their tissue health measurements. Can provide correlation.

High-speed repetitive measurements of a single k-value, with a total time of about millisecond to acquire one block, reduced the blur caused by patient and machine movements to a negligible level, previously It enables robust evaluation of fine textures that could not be accessed in vivo. (In contrast, standard MR image acquisition times are required to last much longer, during which time the patient remains completely stationary.) The SNR for each k value measured is within the block. It is significantly improved by combining individual samples at each k value of. Here, this averaging can be performed without worrying about the movement of the object, and this movement is removed due to the high-speed sequential acquisition of the individual samples in the block.

This significant improvement in SNR is that the method claimed herein determines the pathological features of fine textures rather than obtaining the entire number of spatially encoded echoes required for image formation. Therefore, it is possible because it focuses only on the acquisition of the k value of interest. A significant decrease in the data matrix can increase the number of iterations at the targeted k-value and thus can significantly improve the SNR.

The energy density within a range of texture spacing is proportional to the texture wavelength or inversely proportional to the k value-that is, the higher the k value, the lower the associated signal strength. The fast acquisition made possible by using the method claimed herein acquires the k value, which is initially a low signal, before the effects of _{T 2} and T ₂ ^{* reduce the signal amplitude.} Therefore, the number of repetitions of acquisition at a specific k value can be adjusted. In this way, the SNR of each iteration averaged for noise elimination (or space-topological-corrected before being combined with the subsequent k-value measurement from TR) exceeds this threshold. The movement between acquisition cycles at different k values is not a problem as long as each acquisition is within the tissue's volume of interest (VOI). The patented method only targets the assessment of pathologically-induced changes in tissue texture, eliminating the need for phase coherence throughout the data acquisition cycle as required by imaging methods.

Some advantages arise from the acquisition of data after the slope is turned off for a single k-value sampling in reduced volume (VOI). With proper pulse sequencing, the echo recording window can be designed so that recording is initiated at the highest k value of interest because the signal level is highest at Echo Peak. This makes it possible to record fine structures that cannot be achieved by the in-vivo MR imaging method at present.

Further, since T2 ^* becomes longer when the inclination is turned off, the SNR is improved by allowing a longer acquisition time. This makes it possible to obtain an increased number of samples N.

The coil combination is also simplified by having a higher SNR for each k value, thus significantly improving the overall SNR. This is particularly beneficial as the MRI trend is towards coil arrays composed of many small element coils. Due to the small acquisition volume targeted by the methods claimed herein, no phase correction over the sample volume is required. Only one phase and gain value for each coil is needed to combine multiple element channels. These can be combined using the maximum signal-to-noise ratio (MRC) method, which weights the coils most heavily with the highest signal-to-noise ratio, or other multi-signal combination methods. (The phases and gains of the elements of a given coil array can be determined from the phantom once and then applied to the patient's data.)

Signal acquisition and data sampling in a standard MRI scan is done by taking complex value samples of multiple echoes, while applying tilt sequences simultaneously and echoes in sequence. Imaging methods rely on frequency encoding for one of a plurality of dimensions. This is because it is possible to get one line in k-space by each phase encoding, not just one point. With respect to 3D imaging, the two dimensions in k-space usually rely on phase encoding to generate the target k-space fill, and the third dimension is frequency encoded. Acquiring phase encoding in imaging is usually a relatively time consuming process, involving the acquisition of about 256 k-values in each direction of phase encoding. In clinical MRI scanning, it takes about 10 to 15 minutes to generate an image. The goal in image composition is to obtain sufficient k-space coverage and fill in all the coefficients of the 2D or 3D Fourier series, so the standard MR resolution is limited by the motion of the object.

The method claimed herein is the exact opposite of standard MR data acquisition, which focuses on image generation. Image formation faces the blurring that results from the long-term movement of the object required to obtain the large data matrix required. Since the target of the method claimed herein is the texture, not the image, the only condition for the movement of the subject is that the sampled volume is within the region of similar tissue properties while the data is acquired. To stay. This is much less stringent, and because the scale of possible movements is large enough, and because the temporal order is easily corrected by real-time movement evaluation and correction techniques, it is a goal over structural phase coherence conditions. Is easy to achieve. The acquisition speed of the methods claimed herein is such that in most cases no real-time motion correction may be required. While other methods focused on image post-processing to attempt to extract texture measurements, the method claimed herein eliminates the need for image generation and instead In addition, the emphasis is on measuring the texture directly, thus enabling more sensitive and robust measurements.

Often, k-space sampling is considered synonymous with echo sampling in the presence of tilt sets. In the method claimed herein, the technique for k-space filling is to obtain only the set of k-values needed to evaluate the texture in the targeted medical condition, and the data are tilted off. Obtained after being made. This method allows such fast acquisition of a single k-value iteration for averaging for noise reduction so that the motion of the subject does not degrade the data.

The acquisition after the slope is turned off, along with the large improvement in SNR resulting from sampling the k-values individually, with many iterations of the selected set of k-values acquired in a single TR, results in an SNR. It can be further significantly improved, thus increasing the robustness of the measurement. This will be described below.

MR echo sampling provides a specific sample-to-time-dependent echo time. Echoes are not only included in the gradients applied simultaneously (relative to the frequency-encoded axis) and previously (relative to the phase-encoded axis), but also the isochromat associated with different species of the sample, as well as spin-spin Includes envelopes (T2 & T2 ^{*) associated with the interaction.}

The acquisition of conventional frequency-encoded spins imposes a time-varying slope on the sample, which effectively travels in k-space along a predetermined path. In linear sampling, the path is along a straight line.

Frequency encoding produces only one measurement at a given k-space-at a given point in time, the echoed sample represents one value corresponding to the Fourier coefficient at a position in a particular k-space. .. The next echo sample represents a value at a position in a different k-space, and the next k-value depends on the slope of the slope applied at the same time. This technique works well as long as there is sufficient signal for the corresponding k value. However, if the signal of interest is close to or even below the noise floor, additional samples and subsequent post-processing are usually required.

One way to reduce the noise floor in frequency-encoded tilt readout is to reduce the slope intensity and reduce the receiver bandwidth. Narrowing the bandwidth of the receiver actually lowers the noise level and improves the detection of lower signal levels ( _{proportional to the term k B} TB, where k _B corresponds to the Boltzmann constant and T is. Corresponds to Kelvin temperature, and B corresponds to the bandwidth of the receiver in Hz.) However, this is achieved at the expense of larger chemical shift artifacts.

Chemical shift artifacts occur as a result of different isochromats associated with different species within a biological sample. In a frequency-encoded k-space readout, those species that resonate at slightly higher frequencies appear to be displaced from their actual position in image space in the direction of increasing frequency. If the spatial frequency encoding slope is shallow, the apparent displacement can be quite large.

As such, in order to minimize chemical shift artifacts, the slope of the slope is generally made as steep as possible, with obvious shifts within a narrow range (ie, within one or two pixels of the image area). And minimize. However, this therefore requires a wider receiver bandwidth to accommodate the larger frequency range. This also increases the overall noise floor at a level proportional to the receive bandwidth.

Frequency readout generally results in a forced exchange of tilt intensity, noise level, and chemical shift artifacts.

A common technique for noise reduction in signal acquisition is by iterative sampling of the signal, followed by a combination of multiple measurements. For linear noise sources, such as Gaussian noise, this technique improves the SNR by eliminating random noise in the signal, the elimination effect increasing with the number of samples N.

The noise reduction by this erasing technique is convenient for stationary objects. However, motion-induced blurring is a non-linear effect, so the combination of signals (for the texture / structural wavelength being studied) where individual measurements are shifted through large spatial phase angles is of SNR. It does not improve. A fairly standard technique for correcting motion is to look at MR intensity data in real space and re-register continuous traces / images with each other to maximize overlap. It is assumed that the linear combination of these re-registered signals, like the reduction of white noise, reduces the blur caused by motion. However, this only works if the SNR at each individual acquisition is high enough. Re-registration of low SNR samples results in large variance at the estimated location. Threshold theory defines that when the original signal is below a certain noise threshold, combining the re-registered signal with non-linear blur only increases the error in the signal.

The non-linearity captured by the movement of the object increases at higher k values. This is because the phase shift of the texture induced by movement is increased by k-that is, as the size of the structure of interest decreases, the adverse effects of movement become more severe. This implies that the multiple samples to be combined must be derived from the same acquisition sequence acquired in a sufficiently short period of time, ensuring that there is negligible movement between the samples.

The Cramer-Rao Lower Bound provides insight into the number of samples required for the lower bound of the residual variance of estimates, namely SNR vs. sample size in additive white Gaussian noise (AWGN). .. For low source SNRs in AWGN, a large number of samples need to be averaged to obtain useful SNRs. The initial assumption is that higher SNR can be achieved on average after multiple acquisitions. (CRAMER, H;. "Mathematical Methods of Statistics"; Princeton University Press, 1946.RAO, C.R, "Information and the accuracy attainable in the estimation of statistical parameters";. Bulletin of the Calcutta Mathematical Society37,1945.)

Explaining with reference to the drawings, the graph of FIG. 1, which is a comparison of the output SNR shown in trace 102 with the required number of samples shown in trace 104, shows that at high input SNRs, low noise measurements occur. Show that a single sample is sufficient. At lower SNRs, multiple samples are needed to "average" the noise contribution. The ability to combine samples explicitly assumes that the basic signal of interest is relatively constant during the multiple sample acquisition process (ie, the only component that changes is noise).

The graph of FIG. 2 is a simulation by a signal model (trace 202), provides an input SNR (trace 204), and assumes that the input noise level is 3 mV rms, an SNR of 20 dB depending on the position in k-space. The number of k-value samples (trace 206) required to generate the above is shown. Since the spectral energy density is generally ^{proportional to k-1} , more input samples are needed at higher spatial frequencies (higher k values) to maintain proper SNR. The noise level for simulation is adjusted to an SNR of about 10 dB at k = 2 cycles / mm (λ = 500 μm).

As pointed out above, this type of averaging is possible with purely static samples without target tissue displacement or deformation that occur over a temporary period of data acquisition. However, in vivo applications result in natural movement, even when the patient is cooperative. The negative effects of movement are more severe as the texture spacing of interest is reduced. More importantly, this type of averaging is based on the assumption that the basic signal is the same throughout the acquisition, and that only the zero-mean complex-valued additive white Gaussian noise (CAWGN) changes. If the signal itself changes, the result is not only the average of the noise, but also the average of N different versions of the basic signal, which does not improve the SNR much.

Estimating and correcting motion using a low SNR sample results in a high variance of the estimated position. This also results in a large variance in the "corrected" acquisitions, and does not produce the expected increase in SNR when these acquisitions are averaged. This implies that multiple samples must come from the same acquisition sequence, where the movement between the samples is very small. This is made possible by the methods claimed herein.

The problem becomes more serious at shorter structural wavelengths. Considering the two acquisitions of noiselessness for the time being, one of them is displaced by the amount d. Attempts to average them for a given k value
Y (2πk): = S (2πK) [1 + e ^-2πkd ] / 2 (0.1)
(In the equation, S (2πk) is a complex value signal, and Y (2πk) represents the average of the two acquisitions)
Produces.

に制限する。 this is,
Y (2πk): = S (2πK) e ^−jπkd cos (πkd) (0.2)
It can be expressed as both the amplitude attenuation and the phase shift due to the displacement d. By limiting the magnitude attenuation to the floor value a (where 0 <a <1), d

Limit to.

This indicates that for an error of a given magnitude, the allowable displacement decreases as the value of k increases. This is because the smaller the texture spacing of interest, the smaller the permissible movement during data acquisition.

To address this issue, an alternative approach has been taken in the method claimed herein, which omits frequency-encoded reads and samples points in a particular k-space. Yes, one or more measurements are taken at one time for each interest k-value in a single spatial position and orientation.

Within a given acquisition in a standard MR run, there are M samples acquired by echo. Instead of taking samples at each k-value, those samples with N≤M can be used to estimate the (complex number) basic signal value at a particular k-value. Since multiple samples in one acquired data are much closer in time, they can be combined with much less concern about movement than across multiple acquired data.

When measuring a single k value using the entire echo, the receive bandwidth can be adjusted to pass through and attenuate frequencies above the richest resonant peaks in the underlying NMR spectrum.

Taking a linear MRS spectrum (no structural phase encoding) _{is, H} 2 O (chemical shift [delta] = 4.7 ppm), and carbon results in fat - hydrogen bonds (e.g. CH3, CH2, CH = CH The main spectrum is generated from the peak corresponding to (etc.), and the carbon-hydrogen bond has different chemical shifts ranging from 0.9 to 5.7 ppm, respectively, and the most abundant resonance occurs at δ = 1.3 ppm. It comes from CH2 in the fat chain.

Assuming that a 3T machine is used, the gyromagnetic ratio of hydrogen is γ = 42.576 ΜΗz / Τ, so the chemical shift value is in the range of 166 Hz (for CH2) to 600.3 Hz (for H2O). As long as the receiver bandwidth (one side) above 600.3 Hz is used, the H2O peak will pass. Note that assuming baseband sampling, this implies a sampling rate> 1.2 kHz (when using composite baseband sampling, this is theoretically reduced by about 1/2. The point here is that a narrow bandwidth can be used by this method and the sample rate is about 800 μs. Thereby, the noise of the signal is reduced, and the k-value acquisition data of multiple repetitions is acquired in a few milliseconds, whereby the acquired data is made to avoid the movement of the patient. In contrast, a single image acquisition is performed at TE about 30 ms and TR about 500 ms-2000 ms. It can take several minutes to iteratively obtain the data required for signal averaging-a time range in which respiratory, cardiac, and monocontractive movements are limited in resolution by motion-induced blurring. The patented method allows the acquisition of values in multiple regions in k-space with very low signal levels, such as those found at higher k-values (shorter texture wavelengths) -so far. A fine texture range that remained difficult to capture.

A large number of non-zero frequencies are selected to maximize the signal. In general, this does not correspond to a mere average of all the samples taken. Instead, it resembles a matched filter that is "tuned" to the frequency of interest that corresponds to a particular species of interest.

As an addendum, the entire NMR spectrum can be extracted (without any phase encoding gradient, only by volume selection) to obtain a baseline of basic signal strength (and associated frequencies), which is as follows: In addition, it is spatially modulated and provides insight into the texture wavelengths of the texture element under study with knowledge of the expected species.

The isochromat of interest can be extracted by taking N samples of echo and then taking a Fourier transform. Since the echo is performed without tilt, the intensity of the resulting signal in the isochromat of interest corresponds to the (complex number) k-value coefficient of interest.

Given that the goal is to extract the relative amplitudes of the texture wavelengths, the amplitude vs. texture wavelength measurements are just the information we need. However, the complex phasor value needs to be preserved to the end in order to extract sufficient signal strength and distinguish it from the underlying noise floor.

（式中、σ２は、ノイズの分散を表し、｜Ａ｜２は、関心アイソクロマットの振幅二乗を表し、および０＜ε＜１は、推定値の許容可能な誤差を表す）
と概算され得る。さらに、雑音は、ほとんどが生物学的サンプルから供給されると仮定すると、これはさらに、

（式中、ＮＦ_ｅｆｆは、受信機の実効雑音指数であり、ｋ_Ｂは、ボルツマン定数であり、Ｔは、生物学的サンプルのケルビン温度であり、およびＢは、受信機の帯域幅である）
と概算され得る。この場合、Ｎは、合理的な推定値を生じるために所与の取得データ内で取得される必要があるサンプル数のガイドとして使用され得る。 The relationship between noise floor, signal strength (in a particular species-rich isochromat), required sample size, and maximum margin of error is:

(In the equation, σ2 represents the variance of the noise, | A | 2 represents the amplitude square of the isochromat of interest, and 0 <ε <1 represents the acceptable error of the estimate).
Can be estimated. Furthermore, assuming that the noise is mostly sourced from the biological sample, this further

(In the equation, NF _eff is the effective noise figure of the receiver, k _B is the Boltzmann constant, T is the Kelvin temperature of the biological sample, and B is the bandwidth of the receiver. )
Can be estimated. In this case, N can be used as a guide for the number of samples that need to be acquired in a given acquired data to produce a reasonable estimate.

If the number of samples required exceeds the number available in a single acquisition, the combination of measurements from a single acquisition maximizes the signal before spatial re-registration between acquisitions. May be needed to do. Reasonable estimates and displacement corrections between two or more acquired datasets are required. A combination of measurements at a single k value from a single TR block can now be used to increase SNR and re-registration between successive TRs can be much more successful. Can be done.

The entire set of samples or the entire TR obtained in the echo can be assigned to an estimate of one factor in k-space, but an acceptable value can be estimated using fewer echo samples than the maximum. If so, it expands the possibility of obtaining more than one coefficient in k-space within a particular echo or TR.

FIG. 3 shows an exemplary timing diagram of a pulse sequence for data acquisition using the method claimed herein. The RF pulse contained in trace 302 is used to excite a selected volume of tissue under investigation, as in typical MR imaging methods. The first RF pulse 304 is transmitted simultaneously with the gradient pulse 308 at the first magnetic field gradient shown in the trace 306. It excites a single slice of tissue, or slab, at a position dependent on the orientation and amplitude of the first tilt and the frequency contained in the RF pulse. A negative gradient pulse, the pulse 310 rephases the excitation within a slice or slab of a defined thickness.

The second RF pulse 312, which is twice the amplitude of the first RF pulse 304, is transmitted at the same time as the tilt pulse 316 at the second tilt represented by the trace 314 to excite the slice-selective refocusing of the spins. , This second tissue slice intersects the first tissue slice. (This second RF pulse 312 _{tilts the net magnetic vector antiparallel to B 0} , causing spin inversion and subsequent refocusing, thus with a 90 ° RF pulse and a 180 ° RF pulse. A signal echo occurs at a point after the 180 degree RF pulse equal to the time between.) The initial higher value of the tilt pulse 318 is such that at the start of the tilt pulse 316, it induces a large phase lap over the tissue volume. Designed for crushers, or "spoiler" tilts. A similar gradient pulse 322 at the end of pulse 316 comes after a 180 degree RF reverse pulse, thus breaking this phase wrap. In this way, none of the excitations that do not exist before the 180 degree RF pulse, such as the excitation resulting from the imperfections of the 180 pulse itself, do not have this pre-encoding and are not refocused by the second crusher. Therefore, it does not contribute to the signal. In summary, the second RF pulse combined with the applied second slope is a slice of the signal in the region defined by the intersection of the first slice and the second slice set by this second slope. Provides selective refocusing.

The encoding gradient pulse 326 on the trace 314 sets the initial phase lap, and thus the k-value encoding, along the direction of the gradient pulse 326. Generally, the k-value encoding is directed in either direction by the vector coupling of the mechanical slopes, but is represented as a second slope for ease of visualization.

The third refocusing RF pulse 328, applied in combination with the slope pulse 332 on the third slope represented by the trace 330, defines a third intersecting slice-selective refocus to define the VOI. The tilt pulse 332 also uses the crusher tilt.

The negative prefading gradient pulse 326 ends the phase and is a signal echo following the second 180 ° RF pulse, where signal acquisition begins at a high k value, which is then described below. , Can subsequently be decremented (or incremented or reoriented) for further data acquisition. Since the energy density of the signal is generally ^{proportional to k-1} , this method ensures that the k value with a low SNR is first before the effect of _{T 2 causes a large overall signal loss.} Try to get it.

With all tilts turned off, the receive gate 333 is opened to receive the RF signal, which is shown in FIG. 3 as pulse 334 on trace 336. The RF signal of trace 336 is a display showing only the signal present in the receive gate window, without showing the actual details of the RF signal outside the window. Trace 324 As shown above, sampling begins with an initial k value of 340a and occurs as represented by trace 338. Note that on the scale of the drawing, the sampling rate was high enough that the individual triggers of the analog-to-digital converter (A / D) merged into trace 338. (The enlarged time scale of FIG. 4 described below shows the individual A / D triggers.)

In multiple regions of k-space with sufficiently large corresponding coefficients, which can be fully estimated using a small subset sample of one echo, to select a new k-value during the time the echo is being recorded. , Acquisition of another k-value obtained by applying the gradient pulse 342a shown in trace 314 is achieved. After a suitable set time, another set of samples of echo (here derived from the new k-value factor) can be collected. This process can be repeated, taking multiple samples for each of the selected set of k values within a TR. A plurality of samples are taken with an initial k value of 340a. The k-value selective tilt pulse 342a is then applied, and the resulting k-value 340b is sampled. (In the figure, it is shown as a negative pulse decrementing the k value on the second slope, but in reality, this negative pulse and the subsequent multiple k value slope pulses can be any k value or It can be designed by any vector coupling of these slopes to select the orientation.) Similarly, the k-value selection slope pulse 342b selects a third k-value 340c sampled by the A / D. Each gradient pulse changes the phase lap and selects a new k value. The application of the k-value selective gradient pulses (342c-342f), followed by multiple samplings of the resulting k-value coefficients, is repeated as many times as desired. Data continues to be acquired, but samples of interest are acquired when all slopes are off. Inclined orientation relative to the slice and k values selected may be coincident with the machine inclination, mechanical tilt are aligned so that the right angle or contrast are consistent with B ₀ field. Alternatively, the acquisition direction and k-value encoding can be selected using a slope, which is a vector combination of all three mechanical slope axes.

In situations where it is desirable to measure a low SNR k value, the prewinding encoding gradient pulse can be set so that the first k value to be measured is the desired low SNR k value. Alternatively, the prewinding gradient pulse can be set to zero, so the first k value measured is k0. The measurement of k0 determines the sensitivity of the system's receiver to a particular VOI, to determine the relative dominance of the isochromat regardless of texture at the VOI (eg, water vs. liquid), or It may be desired to establish a reference value for normalization of other k-values measured in the VOI, or for comparison with k-values from other VOIs. In addition, strategies for collecting a particular set of k-values for VOI measure low SNR k-values (generally higher k-values) in multiple TRs in the first set and then remain in the same VOI. In the meantime, it may include measuring k0 and other higher SNR k values in other TRs.

As shown graphically in FIG. 3, the signal reaches its maximum at the time of spin echo. It is also shown in the chart that the signal changes throughout the acquisition period of multiple RF measurements of the k value, and more so during the period between the quasi-next measurement blocks of the k value. Temporally matching the low SNR k-value measurement with the highest echo signal enhances the SNR of the k-value measurement, or matching the higher SNR k-value with the lower echo signal is the echo period. Allows you to collect additional useful k-value acquisitions inside.

FIG. 4 shows an enlarged view of the pulse sequence of FIG. 3 during the initial portion of the RF sampling window 338 from 7.25 to 8.00 ms. Multiple samples of the same k value taken in quick succession with all slopes off provide an input for signal averaging and reduce the AWGN when the SNR is low. In the first block 344a of the sampling window 338, a plurality of samples 346a are taken with a first k value of 340a. A transition sample 348a is taken during the period of application of the k-value selective tilt pulse 342a. When the k-value selection tilt is turned off, multiple samples 346b are taken with a second k-value of 340b. The application of the k-value selective tilt pulse 342b then occurs with the associated transition sample 348b, and subsequently, after the tilt is turned off, a sample 346c with a third k-value of 340c is obtained. The basic signal is minimally affected by motion due to the very short time window used to acquire the data at each given k value. Because data is acquired off the slope, no problem on the chemical shift, and the effective T ₂ ^* is longer, increasing the signal value.

The sampled value of the echo acquired while the k-value selective tilt pulse rises and remains stable and then falls to zero is necessarily affected by the slope applied. These transition samples may provide other interesting information, but are not used in view of the linear measurement of the k-value factor. Only samples recorded when there is no currently active slope are used for this.

For each k-value, a consistent number of samples can be obtained, or an alternative sequence can be used in which a small number of samples are taken because the k-value decreases and thus the amplitude of the signal increases. A pulse sequence designed for this type of acquisition is shown in FIG. Multiple samples of each k-value targeted in the acquisition are acquired in quick succession with the slope off. These iterations provide an input for signal averaging in low SNR signals. Due to the very short time window in which data is acquired for a given k value, as in the pulse sequences shown in FIGS. 3 and 4, the effect of motion on the underlying signal is minimal. ..

The samples in the plurality of portions of the sample windows 344a-344g, which are outlined in FIG. 5, have a given k-value of 340a-340g evoked by the unwinding pulses 342a-342f of the k-value selection tilt, respectively. Corresponds to the number of samples obtained. The number of samples associated with a given k value, N _k , can be selected based on expected SNR, tissue contrast, contrast vs. noise, pathology, texture size, and / or texture bandwidth. In the example shown in FIG. 5, it can be shown that the reduction in sample size is taken for a gradually decreasing k value (larger texture features). This is because, as mentioned above, the primary signal amplitude increases with the decrease of the k value-the energy density is generally proportional to ^k-1. For the same reason, when the effect of T2 is minimal, larger k-values are acquired first in this scheme, longer wavelengths, higher signal strength, k-values are recorded later in the acquisition.

Echo and / or new TR refocusing can be used to build a library of k-space samples. Acquisition of multiple k-values within a TR is facilitated by applying multiple refocusing gradients and / or RF pulses, which can increase the amount of time additional k-values can be sampled within a TR. These late echoes can probably be used to obtain lower k-value coefficients in the selected set. This is because their energy densities at a range of values are generally higher, and _{the effect of attenuation of T 2 on} the overall signal does not affect them as severely as the higher k values. In this way, a larger portion of the required k-space filling can be achieved over fewer TRs, allowing faster data acquisition and minimizing the need to reposition the VOI.

FIG. 8 shows an extension of the basic sequence of methods claimed herein that use spin echo refocusing to extend the recording time of TR. Refocusing with the associated tilt pulse 804 Applying RF pulse 802 results in slice-selective refocusing. After an appropriate set time, the second sampling window 806 is opened by the receiving gate 808. Multiple k-value selection tilt pulses 810 are applied to increment the selected k-value, and after turning off each successive tilt pulse, multiple samples of the selected k-value are taken in the sampling window. NS. A second slice-selective refocusing RF pulse 812 with the associated tilt pulse 814 was again opened by the receive gate 818 after reversing the spin and applying each in a number of k-value selective tilt pulses 820. Data is acquired in the third sampling window 816. As shown in the drawings, the increased number of k-values can be sampled by each refocusing. Refocusing can be repeated until the signal level drops from T2 and other effects disable further signal acquisition. Another way to extend the recording time by exciting multiple signal echoes is to use one or a series of tilted recall echoes (GREs). Since GRE differs from SE in that the effect of resting heterogeneity cannot be refocused, the effect of T2 ^* limits the number of iterations.

In addition to the available tissue contrast, the k value associated with a particular medical condition is part of the determination of the _{number of samples, N k, required for signal averaging.} In liver fibrosis, suitable texture wavelengths are in the 400 micron range, i.e. in the k value of 2.5 cycles / mm. This is similar to the texture spacing found in the development of fibrosis in many other diseases such as cardiac fibrosis. The spacing of the elements in the cancellous bone varies considerably, but the minimum spacing of interest is the width of the trabecular element, which is about 80 microns, setting the maximum k value to 12.5 cycles / mm. In neurological conditions, many of the textures of interest are very fine, on the scale of 50 microns, which is equal to the k value of 20 cycles / mm.

Each medical condition dictates exactly what is needed as quantitative data, i.e. which part of a series of k-values needs to be monitored, and at what resolution and sensitivity. In some medical conditions, short (long) wavelength features increase at the expense of long (short) wavelength features (eg, cirrhosis). In other pathologies, reduced amplitude and expanded short wavelength features indicate a deterioration in the progression of the disease-eg, the regular formation of cortical neuron minicolumns (approximately 80 micron spacing) that advance dementia. In bone, as we age, the highest k-value features first disappear in the structural spectrum. The major structural peaks then slowly shift towards lower k-values to advance osteoporosis, and the pace of this shift accelerates the increase in the proportion of thin trabecular elements to the point of fracture.

The signal level that can be obtained depends to some extent on the anatomy. For example, the required resolution is highest in the brain, but the cortex near the surface of the head ensures that the use of surface coils provides a significant signal boost to the cortical structure. In the liver, the required resolution is low because the structure of interest is on the order of hundreds of microns instead of tens of microns. However, the organ is deeper (further from the coil) and reduces the signal being measured. The use of in-table coils for spinal data acquisition results in modest signal levels and good stabilization. Also, since bone is a high contrast target, the SNR conditions are not so severe. For all these reasons, the exact number of iterations required for averaging depends beyond the targeted k-value range.

FIG. 6 shows that the ability provided by the claimed method to obtain many iterations of the target k value within a single TR allows robust signal averaging and boosts the SNR. The simulation demonstrating is shown. Assuming that the displacement rate of the subject is 30 μm / sec (actually measured clinically over several scans) and the sampling rate = 33.3 kHz (ΔT sample = 30 μs), for averaging 90 repetitive samples can be taken quickly enough that the acquisition remains immune to motion, even up to a k value of 20 cycles / mm (texture wavelength = 50 μm).

FIG. 7, in contrast, is a relatively fast tilt refocus sequence that results in a sampling rate of about 67 Hz (ΔT sample = 15 ms) using traditional techniques for obtaining spatially encoded echoes. Even if it is assumed, it is shown that the movement of the object over the time required for 90 repetitions significantly deteriorates the signal and the ability to improve either SNR by averaging the signals. The situation is actually bad due to the following facts: Acquiring 90 iterations using traditional spatially encoded echoes would require several TRs, the acquisition time would be significantly longer, and the signal degradation due to motion would be much greater. Becomes fierce. Except for the lowest k-value, the potential signal-to-noise ratio due to the combination of multiple samples was negated by the effect of motion.

Acquisition of a sufficiently large selection of k-values allows for the construction of structural profiles in one or more dimensions. As mentioned above, a library of k-space samples can be constructed using refocusing echoes within a single TR or multiple TRs. Phase coherence may not be maintained between different k values if the k values are acquired within the temporally separated TRs and displacements occur between them. This is not a problem if our basic interest is in the relative strength of the signal at a particular k value. If profile or image formation from a library of this value is desired, the required post-processing is, as input, high SNR measurements obtained within each TR using the methods claimed herein. Have. These measurements can then provide a robust input for any necessary re-registration between echoes or TRs towards profile construction. As an example, selection and measurement within a first TR of a selected set of k-values can be achieved by at least one having a low k-value. In subsequent TRs, selection of the same set of k values allows re-registration of data between the two TRs. The reason is that the phase shift can be smaller and correlate between the two TRs than for higher k-value textures, even if significant movement causes a phase change at low k-values. Basically, the higher the k value, the larger the phase shift due to the movement of the object. Obtaining signals from continuous encoding with significantly different k-values allows better estimation of phase shifts by careful comparison of the apparent phase shifts for each.

This is very similar to X-ray diffraction, in which the resulting amplitude-only information (no phase) is the best estimate of the corresponding structural profile based on this amplitude-only information. Present the tasks that determine. There are algorithms for solving the problem, and the chances of success are the range of k-value coefficients obtained, the SNR of the averaged coefficients, respectively, and the values contained in the acquisition of the nominally single value of the k-value. Depends on width. Opportunities for success in this effort are significantly increased by using the claimed method due to the avoidance of subject movement.

The ability to reconstruct a profile from k-value data depends somewhat on the breadth of the spectrum of each single k-value acquisition. It is affected by the size and shape of the VOI (volume of interest), but is also affected by the k-value and pathology as tissue degradation results in more random texture within the tissue.

Choice of VOI-Shape, dimensions, orientation, and position within an organ / anatomy influences the data being measured and its interpretation. The shape of the VOI can be selected to maximize the usefulness of the acquired data. Data can be obtained in different directions and at different texture wavelengths (k values) within the VOI, allowing evaluation of texture anisotropy. Textures can be sampled with multiple VOIs either interleaved within a single TR or in sequential TRs for assessment of changes in pathology across organs. The standard interleaving process for VOIs is used within the TR and is added by applying additional encoding pulses to the vector coupling gradients and the associated k-value selection gradient pulses for the k-values in the interleaved VOI. Data can be provided. As mentioned above, additional excitation RF pulses with related slice selection tilts can be repeated within the same TR by exciting the volume of interest with a tilt set for each iteration, and the tilt at each iteration. Has a different orientation than the first gradient pulse 308 initially applied to the TR to define an additional VOI for excitation in the new tissue that does not overlap with any of the previous VOIs in the TR. It has at least a first tilt (fourth, fifth and sixth tilts in the first iteration, followed by an incremental tilt of subsequent iterations). This response is averaged by taking several measurements, whatever can be mapped or appropriate for the targeted medical condition. This is similar to multipositioning tissue biopsy. However, in the case of tissue biopsy, the number of repetitions is limited due to the highly invasive nature of the technique. The minimum number of structural vibrations sampled at a particular k value indicates the minimum VOI dimension in the direction of sampling-the required length varies inversely with the targeted k value.

To ensure proper sampling of the structure when targeting a range of k values, the dimensions of the VOI in the sampled direction can be kept constant for all k values in the target range. As a result, the sampled structural frequency changes with the k value. This is a simple solution and the sampling dimensions need to be set by the lowest k-value (longest wavelength structure). Using this technique, the sampling dimensions of the VOI are larger than those required for the highest k value in the range and therefore do not result in more localization within the tissue than would otherwise be possible.

Alternatively, data at widely different k-values can be acquired in continuous TRs using VOIs that vary with the particular k-value targeted. Alternatively, the dimensions of the VOI can be selected so that acquisitions in different directions within the VOI are matched to sampling in a particular texture frequency (k value) range.

Similarly, the VOI can be kept constant, and the vector coupling gradient and k-selection pulse for encoding can be changed from TR to TR to assess the size of the feature.

In some cases, it is desirable to densely localize within the spatial domain and broaden the localization within k-space. By defining a non-cubic acquisition volume, it is possible to acquire data from different k values along different (orthogonal or other) directions within the VOI within one TR. The VOI 902 of the elliptical cross section of FIG. 10 is one such possibility. Acquisition along any radial direction and along the axis of the shape is possible within one TR.

In addition, the flexibility of the methods claimed herein can be used to sample k-spaces in linear or curved trajectories. For example, textures can be sampled along radial lines or along arcs or spirals to elicit information on the size of the texture along different spatial directions. These methods are used to determine the anisotropy of textures, or the sensitivity to alignment of semi-crystalline structures such as cortical neuron columns, or a library of k-values within the targeted range of tissues in organs. Can be used to build more quickly.

In one TR (ie, one 90 degree excitation), the k-value encoding can be applied in multiple directions by changing the vector coupling gradient applied for the encoding, and the k-selective pulsing. The exact form of VOI and sampling direction can be used to generate a lot of texture information. For example, in healthy tissue, the tissue composition of cortical neuron fiber bundles is semi-crystalline because the bundles form columns. For this reason, measurements of texture spacing perpendicular to the bundle are very sensitive to orientation. When the orientation is exactly perpendicular to the column, a very sharp signal maximum is expected, and if the orientation changes in any direction of rotation away from that maximum, the signal drops rapidly. One way to measure spatial and organizational integrity (marker of pathology) is to "sway" the acquisition axis around this maximum value to look for resonance in signal strength. .. This technique of looking for "texture resonance" by looking for the signal maximum can be applied in any tissue region. If the condition reduces tissue integrity, the sharpness of this peak is reduced and the maximum signal is reduced.

Similarly, the randomness of spacing in some textures is assessed by varying the length of tissue sampled in a particular direction or in multiple directions, followed by a change in the length of the acquisition. Can be done. The value chosen for that length can be varied across multiple TRs to test the sensitivity of the measured coefficients to this parameter.

VOI can be used in several ways, for example, to scramble signals from tissues outside the region of interest, for example, cross-slice selective refocusing, selective excitation using phased array transmission in combination with appropriate tilt. Can be selectively excited by adiabatic pulse excitation. Parameter selection for various methods can be made by paying attention to SNR optimization. For example, the VOI was generated by slice-selective excitation and two additional mutually orthogonal slice-selective refocus pulses, as in VOI 904 of FIG. By carefully designing the RF pulse, the shape of the VOI can be designed so that the edges are smooth and the window function is closer, as shown in FIG. These window functions provide volume selection without adversely affecting spatial frequency. Recall that in Fourier theory, each spectral line is obscured by the convolution of the Fourier transform of the window function. It is desirable to minimize this ambiguity in the basic spectrum, as it reduces the energy spectral density and adversely affects the SNR.

As mentioned above, it is important that VOIs can be moved from one location to another within the organ or anatomy under study to measure changes in texture / pathology. This response can be mapped or some measurements can be taken and averaged to suit the targeted medical condition. This is similar to multipositioning tissue biopsy. However, in the case of tissue biopsy, the number of repetitions is limited due to the highly invasive nature of the technique.

Different diseases and conditions affect tissues in different ways. In general, disease progression involves both 1) loss of energy density in a particular region of k-space and / or 2) a shift in texture energy density from one part to another in k-space. 3) Changes in the width of existing peaks in the k-values of a series of textures are accompanied. Using trabecular structure as an example-As bone health deteriorates, the average separation of trabecular elements becomes wider (texture shifts to lower k values) and becomes more non-crystalline (k space). (Peak becomes wider), while structural elements become thinner in parallel (shift to higher k values in different parts of the spectrum.) Other tissues / organs are their own individual signs in the texture of the tissue. Affected by diseases with.

Using the methods claimed herein, the k-space is probed to reveal the texture so as to eliminate the loss of signal resolution resulting from the blur due to the movement of the object. Instead of measuring the large set of k-values needed to create the image, we now focus on the acquisition of several k-values selected per TR and yield high SNR by sufficient repetition of each. Each of the individual acquisitions is concentrated on a single k-value. In the first order, the spatial encoding is a single spatial frequency sinusoidal encoding, but there are several factors that have the effect of widening the spatial frequency selectivity of the k-value measurement. One important factor affecting the breadth of k-value measurements, or bandwidth, is the length of the sampled tissue area. A longer sampling length involves more texture wavelengths along the sampling direction, which has the effect of narrowing the bandwidth of the k-value measurement. (This is an inverse proportional relationship between the actual measurement range and the k-space measurement range.) Therefore, the aspect of the claimed method is determined by the VOI dimension or by the acquisition dimension. The ability to set the bandwidth for k-space measurements by choosing the sampling length appropriately. Using this method, the bandwidth of the measurement can be set with the desired k-space resolution appropriate for the tissue being evaluated. (A high k-value for good texture resolution and a high resolution in k-space for sensitive monitoring of pathologically induced changes are required.) For higher-order structures, the predictive range While a set of narrow bandwidth measurements distributed over the texture wavelength can be selected, for more disordered structures, such as the progression of fibrous textures in liver disease, use a single or several broadband k-value measurements. Can be selected to monitor the progression of fibrous textures.

It is necessary to measure both the relative intensity of the k-values of the various textures present in the tissue and the width of the peaks along the k-values of the series of textures present in the texture under study. As such, the data acquisition does not result from experimental parameters, but probes specific one or more regions of k-space using parameter selections that can measure the relative width of peaks originating from the underlying tissue. Can be designed to. It is necessary to recognize the interaction of the two components and design experiments in order to produce the best measurement of tissue changes evoked by the condition.

It is desirable to get a good measurement of the texture by taking multiple measurements of the signal amplitude at multiple specific k values that are close in time, before the data becomes blurry due to motion, in the shortest amount of time. Repeat measurements are made to allow correlation between the best measurements for averaging. An alternative example of repeatedly acquiring a large number of measured values at one point in a three-dimensional k-space is to acquire inclined data, in which the k-value changes continuously during acquisition and the range in the k-space changes. It should be determined by the height of the slope and its pulse width. In addition to changing the magnitude of the k vector, its direction can also be changed during data acquisition. The combination of directional and magnitude changes over acquisition results in a curved trajectory through k-space. If this deviation is small enough, the k values remain somewhat correlated and they can be combined more efficiently to increase the SNR than if they were simply averaged. Therefore, the slope at the time of acquisition smoothes the direction and magnitude of the k-vector to smooth the signal speckle-which appears as a time-varying signal during data acquisition resulting from interference of various phases and amplitudes of the individual spin signals. Can be used to intentionally change. The variations selected in the direction and magnitude of the k-value during acquisition are selected to provide well-combined measurements to obtain a representative power estimate in the vicinity of k-space.

Changing the k value to reduce speckle can be achieved within a single or multiple echoes. For a sphere in k-space defined by the magnitude of the k-value under study, the k-value can be sweeped over the surface of the sphere while maintaining the magnitude of the k-constant, or in the same angular direction. Can be changed by maintaining and the size of k can be changed, or by changing both at the same time.

To reduce the effect of speckle, these variability are usually small enough from either the magnitude or direction of k, where there is a significant correlation between the measurements for the particular tissue under investigation. be.

The principal components of spatial frequency are the same for all of those measurements (they are correlated) unless the tissue being measured has a highly crystalline texture. However, the usual diffraction patterns of microcrystalline or amorphous structures are often speckled. Therefore, by sampling multiple points in the same region of k-space, they are combined in various ways selected to provide optimal smoothing to reduce speckle patterns. The result is a better and more robust measurement by averaging the variability.

A number of techniques can be used to "dither" the k-value to reduce speckle or match width in k-space. The first method is to sweep over a range of angles while keeping the slope on for a period of constant k and acquisition, and to eliminate speckles, measurements using correlation information. Is used in combination with. Alternatively, the same direction in k-space can be maintained, but resized, by leaving the slope on as it is during the acquisition period and combining the measurements using the expected correlation. As yet another alternative, both magnitude and orientation can be changed simultaneously or over a series of acquisitions, essentially reducing noise in structures in "narrow" regions around a particular k-value and representative. Two other alternatives are performed simultaneously to provide both a better evaluation of the magnitude of k, i.e. to reduce speckle. )

When combining measurements of different k magnitudes for noise reduction averaging, there is a phase shift from one radius (magnitude) to the next radius from the wind-up of the slope. .. Fading can be achieved before averaging.

Combining measurements of different sizes in an amorphous structure is better known than combining different angular measurements. Here, in addition to the scheme of reducing thermal noise by high-speed sampling, fluctuations due to speckles (even if the real signal confuses the good evaluation of spatial frequencies) can be reduced.

Therefore, dynamic k-space acquisition is used. The acquisition mode is dynamically selected based on the signal-to-noise ratio (SNR) of the signal at various k-space locations. The slope applied to signal acquisition, the post-acquisition receive bandwidth, and the estimation algorithm used are dynamically adjusted based on the predicted SNR value in k-space to optimize the acquisition time and post-processed SNR. To. In the high SNR region, a single sample at a given k value can be a good estimate. This requires a relatively large receive bandwidth to accommodate relatively fast signal changes in the receive chain when k changes rapidly (due to a large slope).

In the region where the SNR is medium to low, the magnitude of the slope is reduced so that subsequent samples are not taken at the same k value but are correlated, which in turn is within the k value in that range. It can be used to improve the estimation of the value of a basic signal.

Correlation can be introduced into k-space due to window formation selected in profile space. In order to be able to combine a series of samples from the ADC to improve SNR, the correlation between successive samples is enhanced by the proper selection of window formation in the profile space, resulting in shorter windows. Drives a set of values in k-space to a larger correlation distance, and a longer window causes a lower correlation between samples. Inducing the correlation of adjacent points in k-space by window formation in profile space is often a mathematical tool that can facilitate the measurement of basic textures in low SNR environments. Basically, window formation obscures the data, so the power spectrum of the k-value is obscured through the k-space, and the series of measurements can be more easily averaged / combined to increase the SNR. obtain.

In very high SNR environments, it is desirable to measure the power distribution of the actual texture over a range of k-spaces, so the largest possible windows are used. The longer the sampled region in the real (profile) space, the more accurate the measurements when measuring the amorphous texture. Reducing the sampled region by window formation to induce correlation in k-space actually obscures certain desirable measurement points.

However, while facilitating the measurement, the induction of correlation by window formation more or less obscures the basic relative power density profile in k-space that results from the basic texture that is the target of the measurement. The narrowing of the spacing between samples in k-space (determined by the speed and magnitude of the slope of the analog-to-digital converter) increases the correlation, which is used in the post-acquisition process for better estimation. Can be done. In addition, the receive bandwidth in these areas can be reduced, which further reduces the noise floor.

In regions where the SNR is significantly lower, multiple acquisitions of the measured signal level at a particular k value can be taken with a zero (and / or non-zero) slope during the acquisition. Therefore, multiple acquisitions can be optimally combined to provide an estimate for a particular k-value.

An exemplary hybrid pulse sequence is shown in FIG. This particular sequence is an example of fast acquisition with a refocused echo (RARE) type sequence, in which case three different levels of tilt are used for acquisition over three separate echoes, as shown. The illustrated pulse sequence for selecting the desired VOI and the initial phase wrap for k are as described in FIG. 3 and are labeled with the same reference numerals in FIG. This exemplary pulse sequence for establishing VOI is used in the various examples disclosed herein, but the determination of VOI is, for example, a time-varying RF pulse to which a corresponding time-varying gradient is applied. It is done by one of a number of methods, including. Similarly, during or after the pulse sequence used to determine the VOI, the encoding gradient pulse can be applied to the selection of the initial k value. Data recording begins by acquiring a value in the region | k | >> 0, assuming that the signal is minimal there and needs to be acquired first. The second echo sample value was associated with | k |> 0, but its signal level is still relatively low and requires a combination of multiple measurements to provide a robust SNR. The final echo sample value was associated with | k | in the vicinity of | k | to 0, where the corresponding signal was maximum. Note that this is just one example of how this hybrid approach, which uses both zero and non-zero slopes in a single acquisition, can be used. Windups of different amounts of k-values (as determined by the magnitude of the slope and pulse duration) can be obtained in one echo rather than in multiple echoes, as described below. Multiple combinations of windups with different k values can also be obtained in one echo. Further, in the drawings, refocusing such as using RF pulse is disclosed, and tilt refocusing may also be used.

Details of the pulse sequence / signal acquisition on an enlarged scale are shown in FIG. In this sequence, the first echo from the RF pulse 328 and the simultaneous tilt pulse 332 is a pulse sequence similar to that previously described with respect to FIG. 3, and acquisition of multiple samples at slope = 0, and k. The increment of the k value by the value selection tilt pulses 1100a, 1100b and 1100c is used. This samples multiple values at a given location in k-space, after which the values are optimally combined. This is suitable for regions in k-space where the values are significantly smaller and therefore have a significantly lower SNR. This generally occurs when | k | is large.

The second echo from the RF pulse 1102 and the simultaneous tilt pulse 1104 is acquired with a small non-zero slope 1106 that serves the function of time-dependent phase encoding. Small slopes are defined as slopes that provoke samples in k-space, and these samples can be spaced sufficiently close together so that the samples are highly correlated. These samples can then be post-processed by an estimator that utilizes high sample-to-sample correlation to improve the resulting SNR. Quantitatively, the exemplary "small" gradient is up to 20% of the magnitude of the encoding gradient pulse. As shown in the figure, samples of a plurality of values are obtained in a relatively small neighborhood Δk in k-space. The spacing of Δk can be selected so that there is a high correlation between nearby samples due to the window of VOI. Correlation is utilized in estimation algorithms to generate optimal estimates of signal levels over the vicinity of k-space. This is suitable for regions in k-space with low SNR, but its value changes slowly over k-space in small neighborhoods due to the correlation evoked by the window function.

The third echo from the RF pulse 1108 and the simultaneous tilt pulse 1110 is acquired with a relatively large time-dependent phase-encoded slope 1112. The larger slopes used herein result in a series of measurements with lower correlation over the vicinity of the k-space under study in the k-space. In this case, the sample-to-sample correlation available for improving SNR is lower. Larger slopes can be used to sample multiple points in k-space where the values have a high SNR to start (such as those found in lower texture frequency (lower k-value) regions). As shown in the drawings, samples that are relatively widely spaced over k-space and well outside the sample-to-sample correlation as imposed by the window function are generated over the entire pulse. This is suitable for high-speed acquisition of values in k-space, where the signal level is relatively high and high SNR can be recorded. In this case, a single sample at a given point in k-space provides a sufficiently high signal. However, in such higher SNR regions, larger slopes can be used and selected bursts of data samples can be recorded at high speed. Each of these bursts has a significant inter-sample correlation within the burst and allows the calculation of results similar to those described in Obtaining lower slopes above. This can be seen as sampling blocks of data at closely spaced k-values to maintain correlation while migrating along lines in a larger steep K-space.

The acquisition of a nonzero slope can be swept over a curved path in k-space. A wise choice of time increment Δt, slope magnitude G, w (X), and total number N of samples obtained can be adjusted so that the neighborhood of the “high” correlation is M <N. This then allows the estimation of various individual values in k-space by using a subset of M samples for each output estimation. Texture data from within the tissue region defined by the VOI can be obtained by non-zero slope to allow the local distribution of the power density of the k-value within the neighborhood within the k-space to be determined. The range of k-space sampled by the tilt pulse executed at the time of acquisition is determined by the tilt height and the tilt pulse width (pulse duration). Spacing between signal samples in k-space is determined by the magnitude of the slope and the sampling rate (limited by the maximum speed of the analog-to-digital converter). Therefore, the correlation between a series of samples in k-space is determined by the spacing between the samples, the movement of the object, the windows and the basic textures used to group the acquisitions into physical space.

A useful way to select the acquisition parameters is to refer to the degree of correlation required within the set of values to be combined. The wavelength of the repeating structure (texture) is defined as the reciprocal of the k value associated with the texture, i.e. λ _texture = 1 / k _texture . The basic texture signals should not be phase-shifted by a significant percentage of _{λ texture} with respect to each other so that pairs of values [measurements] can be combined to result in improved signal-to-noise ratio. In an exemplary embodiment, the phase shift across the set of samples to be combined must be no more than 80% of 2π.

The resolution in the MR imaging method is limited by the movement of the object at the time of image acquisition. This limitation can be quite significant in uncooperative patients. In addition to patient movement / cooperation, the resolution achievable in MR imaging, an exemplary area of the invention, depends on several factors such as tissue contrast, organs, coil type, and proximity to the coil. Dependent. Robust images of structures in the range less than about 5 mm are problematic, and any one less than about 1 mm is outside the scope of conventional clinical imaging. This is an obvious drawback because many tissue textures in the range of about 5 mm to 10 μm progress and change as the condition progresses, and therefore measurements of these textures can provide a lot of diagnostic information. -These tissue changes are most often the first precursor to the disease. It is this texture wavelength range of about 5 mm to 10 μm that is targeted by the methods disclosed herein.

To measure the texture of tissue, the wavelength range in real space that can be resolved, i.e. the wavelength of the texture for a particular medical condition, ranges from a few millimeters to a few microns. This range is the range in which the image becomes unavailable (blurred) due to the movement of the patient. In the exemplary embodiment, since k is defined to be 1 / wavelength, it is used to k value ranging from about ^{0.2 mm} -1 100 mm ^-1 to define interest textures. It summarizes the region of the k-space of interest, defines the magnitude of the gradient and the duration of the encoding gradient pulse, induces a phase wrap, is selected for a particular k-value and orientation, and initially. Create a spatial encoding for non-zero tilt applied to measurements in the vicinity of the k-value. The methods of the embodiments herein for sample acquisition and post-processing can all be performed in k-space. The only localization in real space is the positioning of the VOI. To measure texture-that is, to determine power distribution within the k-space neighborhood around a point selected in real space, only the neighborhood around one point in real space is sampled.

The exact range required will vary depending on the target medical condition. for example,
Progression of osteoporosis in the microstructure of bone. As an example, the change in average trabecular spacing (TbSp) from healthy bone to osteoporotic bone falls within the wavelength range of about 0.3 mm to 3 mm. The equivalent range of k values is 0.34 mm-1 to 3.4 mm-1. Monitoring in hepatic fibrotic disease shows that in hepatic tissue texture, from vascular-to-vessel spacing with emphasized healthy collagen to pathological conditions in which leaflet-to-lobular spacing becomes prominent tissue texture. It changes with. The blood vessel-to-vessel range is 0.4 mm to 1.5 mm, with ^{a k value of 0.67 mm -1 to 2.5 mm -1} , while the spacing between leaflets is approximately 1 mm to 4 mm. , 0.25 mm ^{-1 to 1} mm -1 k value. The progression of the angiogenic vasculature around the tumor site generally varies from the texture spacing of healthy blood vessels of ^{about 100 μm; k = 10 mm -1.} Due to its chaotic nature, spacing in the angiogenic vascular system ranges from about 10 μm to ^{1 mm or 1 mm -1 to} 100 mm -1. A diagnostic assessment of dementia-related changes in cortical neuron spacing involves measurements of high k values, with a healthy structure of approximately 100 μm spacing or k = 10 mm ^-1 . A variation of about 10-20% of this value with increased structural randomness indicates disease.

Therefore, the acquisition parameters generate a range of k-encodings in which (1) the magnitude / duration of the slope extends near the k-space where it is desirable to inspect the power density present in the target tissue texture, and (2) the combination. The samples to be combined must occur close enough in time so that there is no significant blurring due to the movement of the subject over the acquisition time of the blocks of the combined samples, and (3) the allowable amount of movement is investigated. It can be selected to depend on the neighborhood (ie, wavelength) of the k-space inside.

Acquisition of texture data from within a non-zero slope target VOI allows the determination of local variation in the power density of the k-value within the vicinity of the initial k-space in k-space. The range of k-space sampled for each tilt pulse is determined by the height of the tilt and the pulse width. Spacing between samples in k-space is determined by the magnitude of the slope and the sampling rate.

These parameters (1) allow the acquisition of sufficient data to combine for significant SNR improvement before the motion of the subject can significantly blur the data for the texture being measured. It is desirable that (2) sufficient correlation be ensured across blocks of k-values from the combined acquisitions to maintain SNR ≥ 0.5 dB, and (3) there should be a power density of k-values in the texture. Is selected to set the range of.

The blocks of the series of signal samples that are recombined to improve the SNR may not overlap, or as described below, a selected number of points or, for example, measurements 1-4, 2-5, 3– 6 may be overlapped by a sliding block or the like used to combine them. In addition, the number of samples in each block can vary from block to block over the range of acquisition in k-space, and this variation in the number of samples combined is sufficient to maintain sufficient SNR to provide robust measurements. Determined by the conditions of correlation. Approximate noise levels can be determined independently by several methods well known in the industry, including measurement of noise in the absence of signal inputs.

Acquisition of multiple tilted data of different magnitudes within a single echo, TR, or scan is selected to allow the best signal-to-noise ratio of the combined signal in different target regions of k-space. It can be achieved by the magnitude of the continuous slope. To allow a series of samples from the ADC to be combined to improve SNR, the correlation between successive samples can be increased by proper selection of window formation in the profile space, shorter windows. Drives to a larger correlation distance over a series of values in k space, and longer windows make the sample-to-sample correlation lower. The width of the window selected affects the shorter windows, the desire for correlation over many samples in k-space, and in real space to provide robust measurements, especially when measuring highly amorphous textures. Defined by both the need to sample a sufficient range of textures in.

The combination of post-acquisitions of signals acquired in blocks in k-space, the number of samples to be combined, is determined by the condition that the correlation between the individual signals to be combined is sufficient to achieve SNR ≥ 0 dB (correlation). The level of is determined by the movement of the object, the magnitude of the slope, the sampling rate, the shape of the window, and the basic texture).

The use of non-zero gradient acquisition results from the interference of fluctuating phases and amplitudes of individual spin signals during data acquisition to smooth out the signal speckle-which appears as a time-varying signal during data acquisition. , Can be used to deliberately change the direction and magnitude of the k-vector over a range. The variations selected in the direction and magnitude of the k-value at the time of data acquisition provide well-combined measurements to estimate typical power in the vicinity of the k-space with an SNR of 0 dB. Here, the neighborhood is within 20% of the three-dimensional orientation and size of the center of gravity of the neighborhood.

Compensation for the change in k over a set neighborhood caused by the application of a non-zero slope can be achieved by using a forbidden k-encoding for a particular set of k measurements. In addition, the correlation within a period and within the set of k measurements obtained from the selected VOI is that the phase shift of the data induced by patient movement corresponds to the k value range of the target texture. It can be evoked by choosing a period so that the biological movement is sufficiently small to be less than 50%. Alternatively, the window function may be chosen so that there is a sufficient correlation between the individual measurements and the set estimates that the desired SNR can be achieved.

Details of the very low SNR acquisition mode on a further magnified scale are shown in FIG. In this part of the sequence, the sample gate is open, resulting in an initial value of 1114a in the region 1116 that gives rise to sample 1118, a second value of 1114b induced by the k-value selective tilt pulse 1100a, and a k-value selective tilt pulse of 1100b. At a third value of 1114c induced by, and a fourth value of 1114d induced by the k-value selective tilt pulse 1100c, the k-value is constant. This is the pulse sequence described above in which multiple signal iterations of the same k value are sampled at high speed and all of them are combined into one estimation.

Details of the low SNR acquisition on a further expanded scale are shown in FIG. In this part of the sequence, the k value is indicated by trace segment 1120 due to the non-zero time dependent phase encode gradient 1106 that exists when recording region 1122 when the sample gate is open. Note that it changes slowly, albeit. However, the range of sample 1124 over k-space is a relatively compact neighborhood where the values are highly correlated due to the window function.

Details of high SNR acquisition on a further expanded scale are shown in FIG. In this part of the sequence, the k-value again changes as shown in trace segment 1126 due to the non-zero time-dependent phase-encoded slope 1112 present during the period when the sample gate in region 1128 is open. do. The range of samples 1130 over k-space is still a relatively compact neighborhood, but outside the intersample correlation imposed by the window function.

The low SNR and high SNR acquisition modes at non-zero slope differ from standard frequency-encoded MRI sequences. Because the gradient applied is not used to establish the position, i.e. frequency encoding, but a time-dependent phase to quickly obtain a number of individual samples over a relatively wide neighborhood in k-space. This is because it is used as an encoding.

As described above, the tilt acquisition can be acquired by one echo instead of multiple echoes, as shown in FIG. Again, the initial phase wrap for k and the illustrated pulse sequence for selecting the desired VOI are as described in FIG. 3 and are labeled with the same reference numerals in FIG.

As shown in FIG. 16 and FIG. 17 on a larger scale, the k value is an initial value 1614a, a second value 1614b triggered by the k value selective tilt pulse 1600a, and a th. It is constant at the value 1614c of 3. Note that the k value is decremented as opposed to being incremented in the example of FIG. This is also the already described pulse sequence of 1622a, 1622b and 1622c (all of which are combined into one estimate) in which multiple iterations of the signal at the same K value are sampled at high speed.

In the second part of the sequence within the same echo, the k-value is slowly, as shown by trace segment 1620, due to the non-zero time-dependent phase-encoded gradient 1606 present during the sampling period. There is, but it changes. However, the sample range 1624 over k-space is a relatively compact neighborhood, where the values are highly correlated.

Similarly, a high SNR acquisition is performed in the third part of the sequence, which is still in the same echo. The k-value also varies as shown by trace segment 1626, again due to the non-zero time-dependent phase-encoded slope 1608 present during the period the sample gate is open. The range 1628 over k-space is still a relatively compact neighborhood, but outside the intersample correlation imposed by the window function.

As shown in FIG. 18 (and on a larger scale in FIG. 19), here the initial phase wrap for k and the illustrated pulse sequence for selecting the desired VOI are illustrated in FIG. 3 and consistent in FIG. A low non-zero magnitude gradient 1802 is applied, which is coded but serves the function of time-dependent phase encoding, and data sample 1804 is found in trace segment 1808, as described above. Is obtained from the initial k-value 1806 for the k-value, which has a high correlation with and slowly changes over time. The initial phase wrap may be selected to provide an initial k value of magnitude corresponding to the low signal-to-noise ratio.

Similarly, as shown in FIG. 20 (and on a larger scale in FIG. 21), a pulse sequence for selecting the desired VOI and initial phase lap to set the k-value region is shown, and is described in FIG. And have the same reference numerals in FIG. A larger non-zero slope 2002, which serves the function of time-dependent phase encoding, is applied, and the data sample 2004 is from the initial k-value 2006 for the faster time-varying k-value found in the trace segment 2008. can get. The initial phase wrap may be selected to provide an initial k value of magnitude corresponding to the higher signal-to-noise ratio. Encoding gradient 326 can be used to wind up to the lowest or highest k-value in the target texture, and non-zero magnitude gradient pulses are imposed in the required direction (increase or decrease k). , Reach the other limit for defining textures in k-space.

The samples obtained can be outside the intersample correlation imposed by the window function. However, the signal level for the k value is relatively high and has a high SNR. Further, as already described, fast acquisition of samples in subsets 2010a, 2010b, 2010c and 2010d, for example, may leave the samples in the subset well correlated, and predetermined or expected textures. It can be achieved to be able to provide the desired data in the structure it has. Many sets of values can be combined, but they are well correlated to improve SNR (averaging is one simple form of combination). Therefore, a set of combined data points is used to characterize the power distribution throughout the acquisition to obtain better measurements of the basic texture within the VOI.

As mentioned above, the paraphrase separates the low k and high k values based on the low phase change in the second 90-180-180 excitation (TR). The SNR is maximized in the tilt ON acquisition by a smart combination of continuous k-value samples by re-registering the continuously acquired signal. The data is acquired over a range of k-spaces where the wavelength is long enough and the motion of the object can be easily corrected by re-registration-ie, the phase shift induced by measurements in this k-range is the texture wavelength. Much smaller than. Low k-value signals are sampled in an alternative refocusing sequence, or series of excitations (TRs), and signal acquisition is from the higher k-value range of interest. Wavelength measurements of TE length are used to determine motion-induced phase shifts over the measurements. The phase shift is then applied to the higher k data before re-registration.

A number of correlations are implied by spatial window formation.

として与えられる。これは、フーリエ対（Ｆｏｕｒｉｅｒ ｐａｉｒ）として、
ｇ（ｘ）⇔Ｇ（２πｋ） （２）
として表わされることが多い。 When g (x) corresponds to a one-dimensional (real value) signal, the corresponding function in k-space is subjected to the Fourier transform.

Given as. This is as a Fourier pair.
g (x) ⇔ G (2πk) (2)
Often expressed as.

Window formation is a process that limits the range of g (x) to a finite region of compact support, but performing window formation in such a way is a discontinuity introduced (artificially) by truncation. Is to minimize spectral artifacts caused by.

に起因する。 Regardless of the particular shape used for the window function, there is an inverse relationship between the width of the window and its spectrum. This is a Fourier relational expression

caused by.

Multiplying two functions has the effect of convolving their respective spectra, i.e. f (x): = g (x) h (x) ⇔ F (2πk) = G (2πk) * H (2πk) ) (4)
Is.

Convolution can be thought of as linear filtering of the spectrum, even if the spectrum is an input signal.

は、Ｇ（２πｋ）スペクトルに対する低域フィルターのような機能を果たし、信号を平らにする傾向を有する。αの値が大きいほど、低域フィルターは狭くなる。これは、Ｆ（２πｋ）の隣接する値間に有意な相関を生じる。 Term

Acts like a low frequency filter for the G (2πk) spectrum and tends to flatten the signal. The larger the value of α, the narrower the low frequency filter. This results in a significant correlation between adjacent values of F (2πk).

Estimators that observe noise samples of filtered inputs are well studied and can be applied to generate optimal estimates; Wiener filters, Kalman filters, etc.

Dynamic acquisition mode may be used, where
X corresponds to a three-dimensional vector in the image space,
g (X) corresponds to the value of the image at a given 3D spatial position.
K corresponds to a three-dimensional vector,
G (K) corresponds to the value of the image g in k-space.

For the sake of simplicity in the beginning, the time dependence of this signal is ignored, which in ^{turn is due to T1, T2, T2 *} , as well as different Isochromats (different species in volume), etc. Depends on the contribution of the signal to be made. After all, these effects are taken into account.

The basic principle on which it relies is as follows.
In general, the SNR of G (K) is highest at | k | = 0, and then decreases as | k | increases. The rate at which the SNR decreases is generally SNR ^∝ | k | −α ( In the formula, α is in the range of 1 to 3).
The sampling rate, combined with the magnitude of the slope, sets the sample spacing (Δk) density in k-space for a given VOI.
As the magnitude of the slope decreases, the sample density increases (ie, Δk decreases).
There can be corresponding correlations depending on the size of the window formation in the image space.

（式中、
ｒは、メートル（ｍ）単位での実数値の３次元の空間座標を表し、
Ｉ（ｒ）は、空間座標ｒの非負数実関数である像を表し、
ｋは、サイクル／メートル（ｍ^−１）単位での実数値の３次元のｋ空間座標を表し、
Ｓ（ｋ）は、Ｉ（ｒ）のフーリエ変換を表し、かつ一般的に、ｋの複素数値関数である）
によって与えられる。および積分は、３次元の空間平面全体にわたる。 When generalized, the simplified MRI relationship between spatial coordinates and k-space is

(During the ceremony,
r represents the three-dimensional spatial coordinates of real values in meters (m).
I (r) represents an image that is a non-negative real function of the spatial coordinates r.
k represents three-dimensional k-space coordinates of real values in cycles / meter (m ^{-1) units.}
S (k) represents the Fourier transform of I (r) and is generally a complex-valued function of k)
Given by. And the integral spans the entire three-dimensional space plane.

In short, S (k) represents the corresponding value of the image function I (r) in three-dimensional k-space.

（式中、

は、値４２．５７６ＭＨｚ／Ｔのプロトンの磁気回転比であり、
ｇ（ｔ）は、Ｔ／ｍ単位での傾斜の強度を表す、時間の実数値の３次元関数である）
を有する。この関数は、プロトンスピンを何らかの望ましい方法で操作することを目的とするパルスシーケンスの一部としての設計入力である。 Similarly, k-space coordinates are a function of time, and general expressions.

(During the ceremony,

Is the gyromagnetic ratio of protons with a value of 42.576 MHz / T.
g (t) is a real-valued three-dimensional function of time that represents the intensity of the slope in T / m units)
Have. This function is a design input as part of a pulse sequence aimed at manipulating proton spins in some desired way.

と表わすことが好都合であることが多い。 The integral in equation (7) shows that the value of k (t) for a given value t is calculated as the integral of all previous histories of the gradient function. Technically correct, but this

It is often convenient to express.

Here, _{if t 0} represents a convenient starting time, k _{(t 0)} is the k value corresponding at _{t 0,} and the lower limit of the integration begins at _{t 0.}

と表わすことができる。Ｓ（ｔ）は、ｇ（ｔ）においてエンコーディングされる傾斜シーケンスと関連して行われるＭＲＩエコー実験中に取得するであろう複素数値の基底帯域信号を表す。 In order to make the dependence on time clearer, equation (6) is

Can be expressed as. S (t) represents a complex numerical baseband signal that would be acquired during an MRI echo experiment performed in connection with the tilt sequence encoded at g (t).

と表わされる。 Without loss of generality, k, g and r are
k = [k _x k _y k _z ] ^T
g = [g _x g _y g _z ] ^T
r = [r _x r _y r _z ] ^T (10)
Can be decomposed into Cartesian components. In addition, equation (9) is

It is expressed as.

In general, k (t) represents the path of a curve in K-space as a function of time.

に単純化される。ここで、

および式（８）は、

へと単純化される。

であると定義する。これは、時に、
Ｒ（ｋ）⇔ρ（ｘ） （１６）
と表わされる。簡潔に示すために、Ｒ（ｋ）およびρ（ｘ）は、フーリエ変換対である。（１２）と（１５）とを比較することにより、Ｓ（ｔ）は、単に、
Ｓ（ｔ）＝Ｒ（ｋ（ｔ）） （１７）
で表わされる様々なフーリエ係数にわたる時間依存性のプログレッションであることが分かる。ここで、時間値ｔと対応するＫ空間座標とのマッピングは、式（１４）によって与えられる。 First, to facilitate the explanation of the initial concept, the evaluation is limited along one dimension by assuming _{k (t) = [k x} (t) 0 0] ^T. Therefore, the equation (11) is

Simplified to. here,

And equation (8)

It is simplified to.

Is defined as. This is sometimes
R (k) ⇔ ρ (x) (16)
It is expressed as. For brevity, R (k) and ρ (x) are Fourier transform pairs. By comparing (12) and (15), S (t) is simply
S (t) = R (k (t)) (17)
It can be seen that it is a time-dependent progression over various Fourier coefficients represented by. Here, the mapping between the time value t and the corresponding K-space coordinates is given by the equation (14).

Generally, a desired signal and noise combination is received by the antenna in order to model the received signal In in a real MRI machine. The signal is then filtered, amplified, down-converted, sampled, and quantized.

Although the specific details are machine-dependent, a simple model can be developed to represent the output of a machine as follows: Y (t) is the combination of the interest signal and the noise signal.
Y (t) = S (t) + W (t) (18)
(During the ceremony,
S (t) is given in equation (17).
W (t) is _{a zero-mean additive white Gaussian noise process of complex values with variance σ w} ² , i.e. E {W (t)} = 0 and E {W (t) W ^* (t + τ)} = σ _w ² δ (τ))
To be expressed as.

によって与えられる）
と表わされ得る。 After that, the received signal Y (t) is uniformly sampled.
Y _n = Y (t) │ _{t = n · Δt} = R (k (t)) │ _{t = n · Δt} + W (t) │ _{t = n · Δt} (19)
This is simpler,
_{Y n = R (k n)} + W n (20)
(During the ceremony,
The sequence _{k n,}

Given by)
Can be expressed as.

を定義する。その後、（２１）は、より単純に、
ｋ_０＝ｋ（ｔ）│_ｔ＝０
ｋ_ｎ＋１＝ｋ_ｎ＋Δｋ_ｎ＋１ （２３）
と表わされ得る。

To define. After that, (21) becomes simpler,
k ₀ = k (t) │ _{t = 0}
k _{n + 1} = k _n + Δk _{n + 1} (23)
Can be expressed as.

In this case, in short, the sequence k _n is defined by a sequence of increment determined by integration between samples tilt function.

Equations (20), (22) and (23) can be used to describe signals under different gradient conditions disclosed herein.

Although "pre-phased" by some prior tilt activity, the tilt is no longer held at zero for FIGS. 12, 14 and 15 as described above. However, the collection of echo samples can be analyzed as follows.

として与えられる。 In that case, the signal of the formula (20) by _{Y n = R (k n)} + W n (24)
Given as, and _{k n} are the formula (21),

Given as.

Measurements, because it occurs in a non-zero slope regime, the integral term is no longer zero, which is that the sequence k _n is no longer constant, and hence the sequence R (k _n) is no longer constant Imply.

Since the basic structure of I (r) not make assumptions, implied a value specific structure or relationship with any Between the R (k _n) can not be made. This clearly puts Applicants at a disadvantage when they want to estimate useful signals in a very low SNR environment.

By applying multiplicative window function in the image area, it can structure imposed on the value of R (k _n). This is achieved by leveraging two Fourier transform identities. Multiplication in one region corresponds to convolution in the reciprocal domain. Define the following Fourier pair.
v (x) ⇔ N (k)
ρ (x) ⇔ R (k)
ζ (x) = Z (k) (26)
Therefore, the product in one region corresponds to the convolution in the reciprocal region.
v (x) = ρ (x) ζ (x) ⇔ N (k) = R (k) * Z (k) (27)

である。 Scaling in one region corresponds to inverse scaling in the reciprocal region. When ζ (x) ⇔ Z (k)

Is.

The window formation function is commonly used to limit the image space to a finite compact region of interest, while at the same time minimizing the adverse effects of the window itself on the corresponding image spectrum. Those skilled in the art will recognize that there are a variety of window functions, each developed with a unique set of features.

を検討する。対応するフーリエ変換は、

によって与えられ、これは、フーリエ対
Π（ｔ）⇔ｓｉｎｃ（ｆ） （３１）
と表わされることが多い。等式（２８）を使用して、わずかに一般化されたバージョンおよびそのフーリエ対は、

である。等式（２７）を使用して、窓のプロファイルおよびフーリエ対は、

である。基準として等式（２４）を使用して、サンプリングされるＭＲＩ信号は、
Ｙ_ｎ＝Ｎ（ｋ_ｎ）＋Ｗ_ｎ （３４）
と表わされ得る。これは、（３３）を使用して、

と表わされ得る。ここで、コンボルーション積分は、明確に展開された。 For the sake of explanation, the most basic window function:

To consider. The corresponding Fourier transform is

Given by, this is Fourier vs. Π (t) ⇔ sinc (f) (31)
Is often expressed as. Using equation (28), a slightly generalized version and its Fourier pair are

Is. Using equation (27), the window profile and Fourier pair are

Is. Using equation (24) as a reference, the sampled MRI signal is
_{Y n = N (k n)} + W n (34)
Can be expressed as. This uses (33),

Can be expressed as. Here, the convolution integral was clearly expanded.

The value of the convolution integral taken by k _n is no longer a function of only one point of R (k _n). For each point k _n, convolution integral, calculating a weighted sum of the values of R (k) around the k _n. The range near the k-space is inversely proportional to the parameter X. When the value of X is small, the width of the neighborhood in k-space increases.

を定義する。これらは、従って時間間隔Δｔおよび関数ｇ（τ）の関数である。 For embodiments herein, the range of regions of interest corresponds to a set _{of values k 0} , k ₁ , k ₂ , ... k _{N-1 in k-space.}

To define. These are therefore functions of the time interval Δt and the function g (τ).

For example, g (τ) = G (wherein, G is a positive constant that is) and, when a simple assumption, k _n is only uniform sampling over a portion of k-space, and k _{n =} k ₀ + nGΔt (37)
Given by. In this case, _kmin and _kmax are
k _min = k ₀
k _max = k ₀ + (N-1) GΔt (38)
Given by.

Simple sampling of k-space may be selected, but is not particularly necessary. In fact, there may be applications where non-uniform and / or even non-monotonic sampling strategies can be useful.

Ideally, parameters X (and window function), the result of the weighted sum over the vicinity of the k _n is,
N _{(k n)} ≒ C
(During the ceremony,
C is a complex numerical constant, but not wide enough to lose significant spectral resolution)
It is selected so that it has a "sufficiently wide range".

である。 For the purposes of the embodiments disclosed herein, the "small" non-zero tilt can be determined based on the desired window. From equation (10)

Is.

となる。これは、ｋ空間における各点が、オフセットｘ_０に比例する複素空間において回転されることを示す。 Suppose the nominal center point of the profile, it is shifted to the center point x _0. as a result,

Will be. This indicates that each point in k-space is rotated in complex space proportional to the _{offset x 0.}

を生じる。ここで、初期位相オフセットθ_０および位相インクリメントΔθは、
θ_０：＝−２πｋ_０ｘ_０
Δθ：＝−２πＧΔｔｘ_０ （４５）
によって与えられる。 Since the slope is a positive constant, according to equation (39),
k _n = k ₀ + nGΔt (43)
Can be assumed to be. By substituting into (42)

Produces. Here, the initial phase offset θ ₀ and the phase increment Δθ are
θ ₀ : = -2πk ₀ x ₀
Δθ: = -2πGΔtx ₀ (45)
Given by.

の各取得サンプルに乗算する。 Due to the proper application of the specified window formed function, if R (k _n) ≒ C is a complex constant in the neighborhood, the rear acquisition estimator and combined into the final estimated value generating it Before, first, the offset phase increment e ^{jn Δθ} ,

Multiply each acquisition sample of.

An estimate of Δθ can be obtained from a sequence of k-space samples taken in the window profile over lower k-values (higher SNR).

Correlation can be elicited by window formation as one parameter as already described. The multiplication of the profile ρ (x) by the real-valued window function ζ (x) is based on the Fourier relational expression v (x) = ρ (x) ζ (x) ⇔ N (k) = R (k) * Z (k). (46)
Corresponds to the convolution in k-space.

Z (k) is processed as an impulse response of a linear filter applied to the complex value signal R (k) in k-space to generate the complex value output signal N (k).

として表わされ得る。 The autocorrelation function of the output signal R _NN (K ₁ , K ₂ ) is as a function of the autocorrelation of the input signal R _RR (K ₁ , K ₂ ), and the impulse response Z (k) is.

Can be expressed as.

によって与えられる。 Equation (47) is inconvenient because the autocorrelation function _{of the basic signal RRR} (K ₁ , K _{2) is usually unknown.} R (k) is a stationary process in the broad sense of white noise, and the autocorrelation is R _RR (K ₁ , K ₂ ) = σ _R ² δ (K _1- K ₂ ) (48).
The simple assumption is made that it is expressed as. In this assumption, (47) is
R _NN (K ₁ , K ₂ ) = σ _R ² R _ZZ (K _2- K ₁ ) (49)
It is simplified to. Here, _RZZ (K) is an autocorrelation function of the impulse response Z (k), and

Given by.

は、基本的なサンプル点が相関される程度を測る。低ＳＮＲレジームでは、サンプルの全てにわたって高相関が望まれるため、下界：

を設定する。 The k-to-time mapping is mapped as follows.
k _n = k ₀ + nGΔt (51)
Normalized correlation

Measures the degree to which the basic sample points are correlated. In the low SNR regime, a high correlation is desired across all samples, so the lower bound:

To set.

Equations (50) and (53) provide a defined relationship between the window function impulse response Z (k), the slope intensity G, the sample spacing Δt, the number of samples N, and the correlation lower bound η _min. offer.

であると定義されると仮定する。ここで、Π（ｘ）は、下記で定義する標準的ないわゆる矩形関数であり、およびＸは、定数である。

For example, the window function

Suppose it is defined as. Here, Π (x) is a standard so-called rectangular function defined below, and X is a constant.

によって与えられる。 The impulse function Z (k) is a Fourier transform.

Given by.

The corresponding normalization correlation function η (K) is
η (K) = sinc (XK) (57)
Given by.

と近似され得る。これは、反転され、かつ（５８）に適用されて、

を生じ得る。ここで、これは、傾斜の強度Ｇ、サンプル間隔Δｔ、およびサンプル数Ｎの積の上界を明白に表す。 In this case, if the correlation is limited to the lower bound by _{η min = 0.95 according to (53),}
η _min ≤ sinc (X ・ N ・ G ・ Δt) (58)
The condition arises. This is done by using the first two terms of the Taylor series.

Can be approximated as. This is inverted and applied to (58),

Can occur. Here, this clearly represents the upper bound of the product of the slope intensity G, the sample spacing Δt, and the number of samples N.

によって与えられる。 In general, the sample interval Δt and the number of samples N are determined by other considerations. Given these as given, the maximum slope level is

Given by.

In this case, with respect to non-zero slope data acquisition, as long as the slope G is below the calculated upper bound, the sample obtained will have a defined correlation level. This condition is defined as a "small slope" level for the purposes of this specification.

Sampling above this limit results in lower sample correlation, resulting in lower potential post-acquisition SNR gains. A "greater" tilt can be defined as operating under this condition. Determining the tilt is as follows: (1) selection of window functions that affect the shape (and to some extent, width) of the "main lobe" in the impulse response (eg, rectangle, Takey, humming, etc.), (2) selection of window range ( The larger the range of the profile region, the narrower the "main lobe" in the impulse response), (3) the impulse response that can produce an autocorrelation function, (4) the effective width in k space that must contain the sample. The desired correlation level to be, and (5) the sampling rate ^* number of samples ^* slope size (as long as this number is limited by the number contained within element (4)), which determines the actual sampling proximity size. Affected by a number of parameters, including), the slope remains in the "lower slope level" regime.

An exemplary embodiment maintains a constant ratio of texture wavelength to length of VOI acquisition axis. As the targeted k-value changes, the length of the VOI acquisition axis changes so that the corresponding texture wavelength to acquisition length ratio remains constant. The purpose here is to keep the number of texture "cells" sampled constant. In this way, it is expected that the differential expansion observed at a particular point in k-space Δk will come from a source other than the length sampled in real space, such as the finite width of the RF pulse or the edge of the gradient pulse. ..

MR-based diagnostic techniques can be combined. Some MR-based techniques designed to look at very fine tissue structures are difficult to interpret in some medical conditions as they only provide indirect measurements of the basic structure. Provide data that can be. Diffusion-weighted imaging and magnetic resonance elastography (MRE) are two such techniques. Since this provisional application method measures directly, it often provides better measurements of fine textures and, in some cases, complementary data that enhances diagnostic ability. Combined acquisition techniques can provide a more robust measurement of texture and thus pathology.

The disclosed embodiments can be used in combination with magnetic resonance elastography (MRE). At present, the main application of MRE is as a diagnosis for liver disease, which determines the response to treatment, progression, need for biopsy, and the like. The targeted medical condition is fibrotic progression, which the technique indirectly measures by measuring tissue sclerosis. In many cases, it is difficult to distinguish fibrosis development from other sclerotic-induced conditions such as portal hypertension and inflammation. In addition, hepatic iron overload, which often results from a compromised liver, results in inadequate visualization of hypointensity and thus induced mechanical waves.

The methods disclosed herein provide a direct measurement of fibrotic progression in the liver, as such, in the case of various triggers of hepatic fibrotic disease, additional with respect to disease progression or response to treatment. Data is provided. Local measurements within the targeted anatomy are provided to calibrate other indirect measurements, such as MRE, DTI, DWI, and the like.

The disclosed embodiments can be used in tumors in combination with or in place of diffusion-weighted imaging. The ability to detect tumor edges with high accuracy facilitates accurate surgical removal. In VOI, data can be acquired using the methods disclosed herein along the direction through the tumor region to look for edges in regions of the angiogenic vasculature.

The ability to measure the inside of a tumor to determine its response to treatment helps with the desired intervention. As an example of the latter, immunotherapy for melanoma tumors induces tumor swelling due to T cell infiltration, which looks similar to malignant tumor growth on structural MR scans. Therefore, it is difficult to decide whether to continue treatment. The ability to examine the condition of the vascular system within a tumor makes it possible to distinguish whether the growth was cancerous or due to an immune system response. The methods disclosed herein also provide local calibration of DTI measurements currently in use, which are often difficult to interpret.

The disclosed embodiments can be used in oncology as bone degradation measurements. It is well known that radiation and / or chemotherapy often impair bone health. Measuring bone changes resulting from cancer treatment helps determine if interventions are needed to coordinate treatment and protect bone health.

Currently, following surgery and treatment for breast cancer, patients are regularly subjected to MR scanners to image breast tissue. The presence of the sternum in the field of view of such tests makes it easy to apply a short additional sequence of this method to measure changes to cancellous bone and thus to obtain a measure of bone health.

As a further example of the possibility that the methods disclosed herein can be used in oncology, the embodiments disclosed are used in tumor formation and progression to stage the progression, type, and response to treatment. In response, the hyperplastic progression of ductal growth can be measured and quantified, or the angiogenic growth of the vasculature surrounding the tumor can be measured and quantified. Continued treatment after breast surgery is often accompanied by reduced estrogen levels, further impairing bone health, and as such, MR scan references for bone monitoring are common. The use of the methods disclosed herein allows for a robust and detailed assessment of bone health by directly measuring cancellous bone structure.

Disclosure embodiments also complement big data and machine learning schemes. Disclosure methods also tend to be used for comparisons between large collections of medical data to learn more about the disease, to be more predictive of individual patients and specific diseases, and to be aware of trends in different populations. Complement. The advantages of using this application method in conjunction with big data / machine learning are:
Instead of the 20 examples of output of the methods disclosed herein compared in populations of unknown medical conditions, for example, k-value distribution vs. femoral fracture can be assessed, or changes in k-value in cortical neuron bundles. Can be measured and correlated with performance in the Alzheimer's Mini-Mental State Examination or other assessments of AD, or local k-values in the liver compared to other inferences of liver disease in large populations,
The use of machine learning in large populations allows the determination of specific biomarkers in the medical condition,
The ability to create useful correlations in big data is much better with high SNR measurement inputs as provided by the methods of the disclosed embodiments.
Such machine learning may include, for example, indicating the case where the disease is defined by the appearance of a particular k-value appearing in the lesioned tissue.

As the disease progresses, changes in the condition occur near the cellular level in the affected tissue prior to the visible condition. For example, in bone disease, fractures are often the downstream effect of ongoing progressive thinning of the trabecular element. In soft tissue diseases such as liver disease, fibrous structures develop over time in the affected organ, eventually resulting in cirrhosis. And in neurology, tissue texture in the brain changes in response to the onset and progression of disease in both white and gray matter. The ability to measure early changes in disease, such as those that affect the texture of fine tissues, allows for earlier treatment, targeted targets for inclusion in clinical trials, and response to treatment to enable early diagnosis. Enables highly sensitive monitoring.

The methods introduced herein are this direct and high of tissue texture, despite their ability to provide clinically robust measurements of changes in tissue texture with disease onset and early progression. Allows measurement of sensitivity and provides the required diagnostic capabilities.

One of the most valuable features of the disclosed method is that it can be used in combination with the most contrasting methods applied to MRI. Since this method measures the texture as opposed to the image, it only needs to have contrast between the tissue texture elements. This contrast can be generated by many methods selected to optimize tissue contrast in a particular medical condition. Measurements of tissue texture result in high spatial resolution due to the relative avoidance of movement of the subject. The acquisition time of the methods previously disclosed herein is short and the data can be acquired by incorporating image acquisition in various sequences.

For example, in bone, where there is virtually no signal from the cancellous bone element itself, a standard T1 pulse sequence that produces high signal from fat provides the required high contrast between bone and bone marrow. Therefore, texture measurements using the methods herein can result in high sensitivity to bone when applied at T1 contrast. The T2 contrast emphasizes the fluid that determines whether a bone lesion is soluble or sclerosing when there may be little fat left around the calcified bone to provide a signal. Can be used for. Because T2-weighted imaging has numerous applications, including imaging of peritoneal lesions, imaging of iron deposits in the brain, and imaging of the heart, these organs / pathologies can be used in combination with the methods disclosed herein above. To be able to emphasize the tissue texture in.

MRI contrast generation is becoming more and more sophisticated over time. In addition to gadolinium and standard ^{exogenous contrast media such as T1, T2, T2 *} , there are, for example, proton density contrast, fat suppression, and inversion recovery sequences. Many new contrast techniques often rely on functional contrast and have been developed to highlight the different tissues involved in the medical condition. MR angiography, a method of visualizing the vasculature and blood flow, utilizes MR signal saturation or phase contrast induced in flowing blood to assess the density and permeability of blood vessels. .. BOLD (Blood Oxygen Level Dependent) Contrast uses metabolic changes in the blood to image active brain regions. Both DWI (Diffusion weighted Imaging) and DTI (Diffusion tensor Imaging) are used to assess pathology in an increasing range of disease and provide a signal that reflects the microstate of the target tissue. .. To assess the condition, ASL (arterial spin labeling) traces the diffusion of magnetically labeled blood (intrinsic contrast) through the brain. Perfusion imaging is used to evaluate other measurements of blood microcirculation, functional response in capillaries. In both of these contrast schemes, the temporal course of blood flow is followed by an assessment of the condition of the vasculature near the tumor. This is because it is an important feature of the diagnosis of glioma. Blood vessels are more abundant in tumors than normal brain tissue and tend to have a larger volume. Higher grade tumors also tend to have higher blood volumes, and degradation and remodeling of extracellular matrix macromolecules results in a loss of blood-brain barrier integrity, which is seen as a leakage of contrast agent. Be done. These measurements can capture the degree of tumor angiogenesis, which is an important biological marker of tumor malignancy and prognosis, especially in gliomas. The application of the methods disclosed herein near the peak of signal contrast results in a direct measurement of the density and size of blood vessels, a correlation that robustly measures the degree of angiogenesis in or near tumors. The data provides a direct measurement of vascular texture on a fine scale. Combined in this way, robust measurements of pathological angiogenesis can be made to stage neurological conditions such as stroke and tumors.

As an example, diffusion weighting, in its simplest form, DWI, uses the random Brownian motion of water molecules to produce contrast in MR images. The correlation between pathology and diffusion is complex, but in general, concentrated cell tissues exhibit a lower diffusion coefficient. Obstacles such as macromolecules, fibers, and membranes also affect the diffusion of water in tissues. Thus, the water molecule diffusion pattern can reveal microscopic details about the tissue state. By measuring the difference in water diffusivity over the area of tissue, a diffusivity map can be created that reflects the local medical condition. Diffusion weighting is particularly useful in tumor characterization, vascular system classification, and diagnosis / monitoring of cerebral ischemia, among other medical conditions. Ischemic infarctions, abscesses, and some tumors in the brain result in very limited spread. Cysts and edema have little limitation on diffusion.

Diffuse images cause some problems in data interpretation, the most prominent of which are: 1) Long diffusion gradients increase echo time TE and reduce SNR 2) Need The high signal-to-noise ratio, which is said to be, causes eddy currents on the metal surface in the scanner, which causes distortion of the signal. 4) The sequence is sensitive to motion by design, so data recording needs to be very fast. Therefore, the ability to increase SNR by averaging signals from multiple acquisitions is limited. In addition, sensitivity to motion can cause difficulty in interpreting the data throughout the acquisition process, as at least six different directions are required to determine partial anisotropy (FA). 5) Interpretation of diffusion-weighted signals is not easy. The measured diffusion coefficient can come from many sources, as the exact mechanism that regulates the water diffusion process in tissues, especially the brain, is not clearly understood. What is inferred from measurements of barriers and restrictions on free diffusion is based on some assumptions about basic tissue pathology. It can take many forms, including cell membranes, organelles, cell spacing, axonal density, glue density, myelin status, and the like. 6) Each DWI voxel represents an average, and the standard voxel size is about 2 to 2.5 mm on a side. To interpret the change in mean diffusion coefficient (ADC) within a voxel, we make some assumptions such as tissue homogeneity and the type of structure that causes diffusion variation.

The methods disclosed herein necessarily involve acquisition within a single VOI or within multiple interleaved VOIs within a TR. The data is acquired without using spatial encoding tilts to form the image. This significantly shortens the acquisition time and, in combination with targeted acquisition in k-space, allows the acquisition of essential data fast enough to avoid movement of the subject. Multiple measurements from a single volume acquisition can be mapped across the anatomy under study, but each measurement is rapidly acquired within a single volume. High SNR is guaranteed by this single volume technique because there is time to repeatedly measure each target k value before the effects of motion enter. Due to the need to avoid movement, conditions to keep the acquisition fast enough limits the number of iterations and the number / range of k-values from which the data is acquired.

The combination of high sensitivity, indirect or speculative measurements to standard DWI movements and their low signal-to-noise ratio does not make the measurements as robust as desired in the clinical setting. The use of the methods disclosed herein for data acquisition when using diffuse contrast still has a long echo time, but the data acquisition is sufficient to minimize blurring due to signal motion. Due to its high speed, movement problems can be alleviated. In addition, additional data can be obtained by the disclosed method during scanning, depending on contrast such as T1 or T2 enhancement. All data obtained by standard DWI images, diffusion-weighted acquisition by the methods disclosed herein, and methods using contrast are input to machine learning algorithms to determine the correlation between measurements. And all three measurements can be correlated with pathological data and outcomes. Correlation with outcome data is time consuming, but will provide the best assessment of the capabilities of the methods disclosed herein.

Due to the long echo time and sensitivity to patient movement, most diffusion weighting is performed using fast echo-planar imaging pulse sequences, resulting in relatively fast data acquisition. However, a single EPI (shift of about 10 pixels) emphasizes chemical shift artifacts.

In addition, fat suppression is often used as part of DWI data acquisition, as there is less movement of water through fat, producing bright signals that can obscure lesions.

In one embodiment of the method disclosed herein, described in more detail below, crusher tilts on either side of the 180 ° tilt to eliminate the concentration of noise signals generated during the 180 ° pulse period. To use. Replacing these crusher slopes with diffusion weighted slopes allows the acquisition of both diffusion weighted signals followed by a restricted k-value signal. As such, diffusion weighting is a measurement in VOI.

DTI (Diffusion TENSOL Imaging) -In highly oriented tissues such as nerve and white matter tracts, diffusion preferably occurs in one direction and along the nerve / axon tract is orthogonal to the track. Much more preferred than spread-track orientation. The degree of orientation or anisotropy within the tissue is an indicator of the medical condition. Because many neurological conditions degrade the normal state of neural structures, such as the arrangement of minicolumns of cortical neurons, or demyelination of axons that form the white matter tract in the brain leads to deterioration of the normal state. Is. In anisotropic diffusion, the value of diffusion contrast changes depending on the direction. Since this anisotropy is a measure of the progression of the pathology, measurements of diffusion contrast in multiple directions can be used to result in "partial anisotropy" resulting from the tissue structure, and thus the progression of the pathology. Provide measurements. In practice, partial anisotropy is measured using at least 6 non-colinear tilts, resulting in a symmetric 9x9 matrix, the "diffusion tensor" whose eigenvalues are in three orthogonal directions. Produces a main diffusion axis.

Diffusion tensors mapped across the brain can be used to delineate the pathways of the white matter tract, along with the use of diffusion anisotropy to diagnose and monitor pathology in the brain. This is called tractography. Possible applications of the methods disclosed herein are multiple sclerosis using standard T1 or T2 contrast, or using the methods disclosed herein in conjunction with diffusion weighting. The texture in the white matter tract affected by (MS) is measured to determine the anisotropy of the measured value with respect to the input for machine learning, which correlates with standard DTI acquisition.

The methods disclosed herein provide the ability to obtain tissue texture using contrasts that may be applicable to the particular tissue under examination. Contrast is applied using any one of the mechanisms described above that enhances the contrast between component tissue types in a polyphasic biological sample being measured. As will be described in more detail thereafter, the contrast mechanism and its application examples can occur at various points within the NMR-induced pulse sequence. Using pulse sequencing, such as those described with respect to FIGS. 3 and 8, the volume of interest (VOI) is selectively excited with multiple time-varying radio frequency signals and applied gradients. Similarly, an encoding gradient pulse as described with respect to FIGS. 3 and 8 is applied to induce a phase wrap to create a spatial encoding for a particular k-value and orientation. Determined based on the expected texture of the tissue within the VOI. A series of time-varying acquisition slopes is initiated, producing, for example, a time-varying locus through the three-dimensional k-space of k-value encoding as described above for FIGS. 8, 11, 16 or 18. In this case, the k-value set is a subset of it needed to produce an image of the VOI. Multiple series of samples of an NMR RF signal encoded with a particular k value are recorded simultaneously. The recorded NMR signal sample is then post-processed to produce a signal-to-k value dataset for the k value in the set determined by the locus.

The first tissue for the application of the methods herein is bone. Although the impact on quality of life is enormous, there is no accurate and non-invasive method for a sensitive assessment of the likelihood of fracture. DEXA, the current typical standard measurement that relies on X-ray resorption, measures the area density of bone. A major determinant of bone strength in predicting fractures is trabecular microarchitecture, the measurement of which is not currently available in vivo. The methods disclosed herein allow this measurement.

Bone breakdown occurs due to several factors, including disease, cancer treatment, eating disorders, and aging / lifestyle. Since trabecular structure erodes bone, it is a redundant number that can be determined from three major morphological measurements-thickness of trabecular element TbTh, repeated spacing of trabecular cells TbSp, and TbSp and TbTh. The number of trabecular bones TbN − changes. TbTh decreases continuously with bone decomposition. Finally, when the trabecular element or strut bone breaks, there is a discontinuity in the local value of TbSp. Bone disintegrates ectopically over time, and anisotropy is largely driven by the effects of load-bearing stress. As bone disease progresses, TbSp increases along the main load-bearing direction faster than in the direction perpendicular to this direction, and the variation between the two measurements becomes a marker of bone degradation. In addition to the occurrence of bone morphometry anisotropy, fluctuations in trabecular spacing TbSp measurements increase to the point of fracture due to thinning of the trabecular strut and TbSp. Causes discontinuities in the measured values of.

The most reliable measure of bone health is the thickness of the trabecular element, which is currently not possible to measure directly in vivo due to the spatial resolution required for the measurement.

The methods disclosed herein allow this measurement. The reason is that this method provides the resolution required to measure TbTh in thinning trabeculae up to tens of microns near a range of sudden discontinuities in TbSp that can be measured to assess further degradation. Because it lowers. Since the trabecular bone is thin to the point of fracture, a sudden shift to increase TbSp should be visible in the signal distribution vs. k value due to the broken trabecular element. In addition, the degree of anisotropy of TbSp can be used as a correlation measure of fracture potential. The data is by positioning the VOI within the target bone region and acquiring data within a single TR or across multiple TRs, allowing acquisition over the appropriate range of k values over TbTh and TbSp. , Can be obtained using the methods disclosed herein. The data rapidly acquired within one TR can be averaged to improve the SNR, in which case the only condition is that the data be acquired from similar bone tissue.

The required contrast is between bone and bone marrow. T1-weighted image provides hyperintensity in the bone marrow and bone provides a negligible MR signal. Alternatively, an IR sequence that produces T1 contrast enhancement can be used. Some work is done using diffusion weighting to image bone-this is a possible contrast mechanism when using the methods disclosed herein.

In healthy bone, TbSp and TbTh are closer in size than in lesioned bone. This can be seen from FIGS. 22A and 22B by comparing the image of healthy bone (FIG. 22A) with the image of severe osteoporotic bone (FIG. 22B). The exact morphology of this relationship varies somewhat, and the difference between these two morphometric parameters is greater over the pathology, eg, in the spinal column than in the hip joint. Increasing the difference in measurements of these two morphometric parameters provides a marker of disease progression.

Because the distinction between the two measurements is so great, it involves the acquisition of signal data in two spatially separated regions of k-space to measure both TbSp and TbTh in osteoporotic bone. In healthy bone, when using multiple slopes to select a region in k-space, it is possible to define regions in some skeletal regions that contain both TbTh and TbSp in the signal vs. k-value distribution. It is possible. As the bone disease progresses, the fluctuation of the measured value of TbSp becomes large, such as the change in the percentage of TbTh. In addition, TbSp increases (wider spacing) and TbTh decreases as the disease progresses. Therefore, the centers of each of these distributions gradually separate as the condition increases. The overall shape of these distributions can be determined using either the magnitude of the slope or the acquisition of slope on or slope off, or a combination of the two, which is expanded by VOI window formation. These widened distributions can be used in real time to determine regions in k-space and sample more finely in continuous TR.

If desired, data can be acquired using multiple interleaved VOIs within each TR. This method allows the variability of the signal at a particular k value within the region of bone to be determined. Data obtained from different VOIs can be mapped and the value / color / icon is assigned to the signal or peak k value in the distribution, or to the difference in k values associated with the peak position of TbTh and the peak position of TbSp, respectively. Be done. The particular k-value and k-value set established by the time-varying slope may include TbSp and TbTh in the range 0.01 mm to 5 mm in exemplary applications.

The methods disclosed herein can be applied, for example, within and around bone lesions identified by conventional T1, T2 or proton density contrast, or flow or diffusion contrast MR imaging, and within that region. Evaluate the condition of cancellous bone. Of interest here is the determination of lesion type. Does the lesion indicate an erosive tumor, or does the lesion in the area of inflamed / decomposed bone surround the fracture? Some lesions are dangerous and erosive tumors, and some lesions are benign. Acquisition of data by the methods disclosed herein in multiple VOIs in the area of the lesion is whether the trabecular structure is degraded in the vicinity and how the degradation is both spatially and temporally. Be able to determine if it has progressed. Additional biomarkers can be derived by inputting MR images of lesions into machine learning algorithms and correlating them with trabecular bone data of TbTh, TbSp, TbN, anisotropy, and measured variability. Since the appearance of lesions on the image can be associated with a particular degree of bone pathology, this novel method can improve the diagnostic content of MR images.

T2 contrast can be used in combination with the methods disclosed herein to emphasize fluid in bone lesions in tumors and classify it as either soluble or sclerosing. In such conditions, it is possible that little fat / bone marrow remains around the calcified bone to provide a signal. Bone lesions in tumors usually have a fluid and a mixture of bone marrow in various inflammatory conditions. Proton densities can be used to generate signals from the outside of hard bone. Alternatively, diffusion weighting returns a signal based on the diffusion of water molecules in the fluid-impregnated bone marrow phase.

The use of the methods disclosed herein to obtain signal vs. k value data in bone, applying the various methods disclosed above, results in several biomarkers for assessing bone health. .. Measurements of TbSp, TbTh, and TbN in multiple VOIs at different locations in the bone and in different orientations of the texture-encoded slopes include the magnitude and variation of morphometric parameters, those in the direction of the bone's load-bearing axis. Produces information on variability in measurements and variability in measurements over local and larger bone regions. For example, the measurement of TbTh in one local orientation in advanced disease clearly reflects the condition, but by combining the sum of the acquired data, markers that are more sensitive to bone health are derived. It can be.

However, correlation data is needed for the development or extraction of features / biomarkers. The best predictor of the likelihood of a fracture is the history of the fracture. Bone biopsy is also very sensitive, but this procedure is rare because it is a highly invasive biomarker. DEXA is the current representative criterion for bone health, which measures surface bone mineral density and is insensitive to trabecular architecture. Therefore, this is the best predicted value for the possibility of fracture. However, for sufficiently large samples, this metric method can enhance the correlation with respect to the diagnostic definition. Taken together, DEXA data and patient fracture history provide a high level of learning framework for correlating with the overall output data derived by the methods disclosed herein and are disclosed herein. Allows the definition of sensitive diagnostic tools from the methods used.

Rather than trying to derive biomarkers from the acquired data by comparing them individually to look for feature extracts, machine learning algorithms provide the best correlation data between multiple measurements.

Furthermore, in addition to the correlation between TbTh and anisotropy data and measurements of TbSp and its changes, fracture data from the subject are correlated with these measurements and also using machine learning algorithms. Develop markers for possible fractures. The biomarker inputs it into a machine learning algorithm from the signal-to-k value data and correlates it with, for example, fracture occurrence data from the same patient, DEXA measurements, or bone biopsy report data. Can be derived directly.

For liver tissue or other tissue exposed to fibrotic invasion, a second use example of the method disclosed herein is provided. The underlying causes of liver disease are complex and diverse, but a hallmark of the disease is the development of fibrosis in the liver. The onset and progression of the disease is characterized by increased protein deposition in the liver structure, primarily collagen fiber accumulation. Fibrotic progression promotes healing in the short term if the disease remains untreated, but the healing response itself becomes a problem, and excess proteinaceous material interferes with the normal functioning of organs. In the case of liver disease, this process progresses to cirrhosis if left untested-resulting in carcinoma and / or liver failure. For this reason, early diagnosis of liver disease is best when a range of management options are available. The use of the methods disclosed herein to assess disease-induced pathological changes to liver tissue is inexpensive and non-existent for routine MR examinations ordered to examine liver health. Providing invasive and rapid additions. Here we focus on liver fibrosis, but the pathology and application of the methods disclosed herein are similar to those of a range of diseases characterized by infiltration of fibrosis. A partial list of these diseases is cardiac fibrosis, cystic fibrosis, idiopathic pulmonary fibrosis, pancreatitis, and renal disease. In addition, in medical conditions such as prostate disease, protein deposition disappears in response to disease progression. The mechanism is reversed, but the tissue texture assessment required for diagnosis and monitoring is the same.

Biopsy is the current representative criterion for diagnosing liver disease, but intra-organ sampling error, significant reading error, and non-negligible complication frequency (morbidity), and even mortality rate biopsy. Make it not the optimal diagnostic method. With sufficient statistics, the development of fibrosis in the liver requires a large number of samples, given the large spatial variability. However, due to the highly invasive nature of biopsy techniques, only a small number of samples can be taken.

An accurate assessment of the progression of fibrosis is needed, but at present, apart from the medical condition, there is no direct measure of this. Fibrosis has progressed considerably by the time liver disease can be diagnosed by liver function tests, ultrasonography, or MR imaging. There is a need for diagnostics that can track disease progression early when it is still reversible. The application of the methods disclosed herein to the measurement of fibrous structure is a direct and non-invasive measurement that allows long-term monitoring of multiple samples of disease onset, progression, and response to treatment. To.

In the liver, like other fibrotic diseases, collagen accumulates in organs in a specific pattern, "decorating" basic structural elements. Liver tissue is composed of a number of adjacent units, or "lobules," whose structure is illustrated by a hexagonal pattern of portal veins in and around the central vein (see figure below). As such, in a healthy liver (no fibrosis progression = F0 stage), large dimensions appearing in the signal distribution of the size of tissue texture features result from spacing between these elements-about 0. Range from .5 mm to 0.7 mm. At the onset of the disease, the development of fibrosis begins in the three portal veins and then progresses to finally form a bridge that connects the three portal veins to the central vein (stages F1-F3). These bridges expand and fuse to form islands of regenerative tissue surrounded by fibrous deposits. In this process, the structural spacing between vessels within the tissue texture is gradually replaced by the spacing between leaflets (steps F3 to F4). Therefore, a clear marker of disease progression is a shift in the distribution of texture wavelengths from shorter wavelengths to longer wavelengths (decrease in k-value), which shifts from about 0.5 mm to about. Often 2 mm and longer. Collagen buildup on the surface of the lobules in response to ongoing disease, and even hepatocytes within the lobules are decorated, so that the lobules themselves become the main texture feature, and repeated spacing between lobules. Is the width of the significant iteration of power distribution in k-space. This change occurs over time as the disease progresses, shifting the power density at the texture wavelength from a healthy range to a feature size of about 2-3 mm (inversely proportional to the k value). The prominent texture features associated with this shift in texture wavelength are well known from histological studies.

In order to diagnose liver disease at its early stage, the shift of the signal from what is expected in a healthy liver to a longer wavelength (lower k value) indicating the onset and progression of the disease is a function of the k value. It can be tracked using the methods disclosed herein. This measurement at a particular time point takes continuous samples at individual k-values across the desired range of texture wavelengths, along the arc of response to disease progression and treatment, or instead tilts. It can be done either by applying it during the period of data acquisition and allowing it to reach the desired range in k-space. A hybrid of these two acquisition methods can also be used.

The contrast between the collagen that decorates various liver structures and the underlying tissue can be obtained using either intrinsic or extrinsic contrast. The signal from fibrosis can be dark on standard T1 imaging and bright on T2 imaging due to the high water content in the fibrotic structure. The use of Gd contrast shortens T1 and at T1-weighted fibrosis appears brighter to the background tissue. Higher contrast helps with more robust measurements. However, because these contrast mechanisms provide contrast between fibrosis and underlying tissue, standard MR imaging methods identify patterns of fibrosis progression in the various liver structures that characterize early disease. It is not possible to obtain sufficient resolution. The patient's movement blurs the image even when using breath-hold imaging or respiratory synchronization. Using standard MR imaging, liver disease can only be assessed at a more advanced stage, when the liver can be irreversibly damaged. Advanced disease can be diagnosed, but what is needed to justify treatment and monitor response is early diagnosis.

Techniques based on a variety of MR imaging methods have been used to assess liver fibrosis in addition to conventional T1 and T2 contrasts with or without contrast media. MRE (Magnetic Resonance Elastography), Diffusion-weighted Imaging (DWI), and MR Perfusion Imaging can yield some information about liver disease, but none can be robustly diagnosed in earlier-stage disease. Their main problem is that they rely on surrogate markers for fibrosis progression. MRE relies on the measurement of sclerosis, perfusion imaging measures blood perfusion parameters, and DWI (diffusion weighted imaging) measures the ADC of water in liver tissue (apparent diffusion cofficient). All of these parameters change in response to many factors, such as inflammation, portal hypertension, steatosis, edema, iron overload, and changes in hepatic perfusion, in addition to the development of fibrosis. At present, there is no direct way to measure the progression of fibrosis in early-stage disease. The methods disclosed herein assess disease status in early liver disease by providing the ability to achieve robust resolution at k values for the measurement of fibrosis progression using standard MR contrast methods. It can be so.

One of the features that novels the method disclosed herein is that it can be used in combination with most contrast mechanisms. One application of this method is in combination with diffusion-weighted contrast using diffusion-weighted contrast (see Figure 23 below), but fiber deposition in early-stage disease rather than the entire image obtained in standard DWI. The signal is acquired only in the k value range of. By using the methods disclosed herein for signal acquisition, data is acquired with much finer spatial resolution than possible with diffusion-weighted MR imaging. The texture being measured is not the average measurement affected by partial volume imaging, but the scale of the fiber-decorated structure between the actual fiber agglomerates. Fibrosis reduces the diffusion coefficient to water, and a lower ADC (apparent diffusion coefficient) in the area of fibrosis results in a brighter signal than the surrounding tissue. The structural signal obtained by using diffusion weighting in combination with the methods disclosed herein measures highly localized water diffusion. Thus, the leaflet unit changes from an undepicted boundary to a hexagonal boundary glue decoration, after which water diffusion at the boundary is impeded to fill the entire leaflet and the diffusion-weighted signal intensity is enhanced. As the condition increases, the use of diffuse contrast can provide the unique characteristics of the texture. Two methods for positioning the diffusion weighted slope are as shown in FIGS. 24 and 25. The illustrated pulse sequence for selecting the desired VOI and initial phase lap for k is as described in FIG. 3 and is labeled with the same reference numerals in FIGS. 24 and 25. FIG. 24 shows the positions of the diffusion-weighted gradients 2402 and 2404 on both sides of the second 180 ° slice selection pulse, whereas in FIG. 25, the diffusion-weighted gradients 2502 and 2504 are before the first 180 ° slice selection pulse. And positioned after the second 180 ° slice selection pulse, it results in a longer diffusion time for the same TE than is available when the pair is placed on either side of the last 180 ° slice selection pulse.

Fibrous texture progression also uses the methods disclosed herein, with or without exogenous agents such as Gd, in combination with contrast, eg, T1 or T2 enhancement. Can be evaluated. By using local sampling in both real and k-space, the methods disclosed herein allow fast acquisition of signals at the required k-value and are pathological at specific points in the liver. Allows Robust Assessment of Tissue Texture-Provides measurements of the texture frequency present at that location, which is induced by the movement of the subject and does not cause the blurring that limits current MR imaging methods. Using the methods disclosed herein, the problem of respiratory movement is avoided by the rate of acquisition of essential data. VOI2602a-2602d using either interleaving acquisition within one TR or measurement at multiple separate TRs to sample the volatility of the condition within the liver, as shown in FIG. Can be positioned at various points in the liver. By using the methods disclosed herein, texture coherence is maintained within each defined VOI through data acquisition in a particular k-value or k-value range, thus maximizing SNR by signal averaging. To enable. If necessary, sampling at the same site may be repeated in subsequent TRs to obtain an average measure of the degree of fibrosis infiltration at that site to assess the stage of disease progression.

Although its application to liver disease has been described in some detail here, assessment of other fibrotic diseases is made possible by the methods disclosed herein. Fibrotic progression is a hallmark of lung disease (eg, cystic fibrosis, idiopathic pulmonary fibrosis), myocardial fibrosis, myofibrosis, pancreatic fibrosis and renal disease. In addition, as mentioned above, some diseases, such as prostate disease, induce a reduction in protein deposition.

Many neurological disorders and neurological conditions have vascular components that can act as markers of disease onset and progression, allowing diagnosis and treatment to be followed, which is the method disclosed herein. A third exemplary implementation is provided. The ability to sensitively assess changes in microvessels makes it possible to monitor the progression of a number of diseases, which are often undiagnosed until the condition has progressed significantly.

Angiogenesis, the formation of new blood vessels from existing microvessels, is required for tumor growth and metastasis. Pathological angiogenesis, rather than the regular formation of blood vessels present in healthy tissue, tends to form chaotic, meandering blood vessels, and is rich in blocked and dead-end structures-see Figure 23. The diameter of the blood vessels and the thickness of the walls are very variable in angiogenic microvessels, and the permeability of the blood vessels is noticeable in places.

For example, tumor aggression is closely correlated with neovascular density, as angiogenesis requires the tumor to be supplied with oxygen and nutrients. The ability to assess the incidence of angiogenesis at the tumor site and characterize vascular morphology allows the aggression of tumors to be assessed. Determining the degree of angiogenic vasculature in and around the tumor makes it possible to determine the boundaries required for surgical removal. Similarly, the response to treatment may be partially traceable by vascular measurements to return to a more standard state. The degree of angiogenic angiogenesis can be assessed to some extent using serum markers or by biopsy. However, biopsies are highly invasive and prone to sampling errors and reading variations.

As a second example, some forms of dementia, most notably Alzheimer's disease (AD), are now recognized to have large vascular components with pathological angiogenesis. Additional forms of dementia, such as Huntington's disease (HD), Parkinson's disease (PD), and frontotemporal dementia, have also been found to have a dysfunctional vascular system. In some cases, a prominent cause of dementia appears to be the pathological vascular system in the brain, such as CVD.

Chronic inflammation is another important factor that can result in abnormal neurovascular structure, exhibiting permeability and bleeding. The pathogenesis of some microvascular systems is linked to the permeability of the blood-brain barrier. Multiple sclerosis, pathological brain damage associated with inflammation, and axonal demyelination result in the destruction of microvessels. Stroke and the resulting ischemia result in the development of angiogenesis, which modifies the capillary network as the body attempts to heal the injury. Because angiogenesis is characterized by increased vascular distribution, it involves both structural and functional changes in the neurovascular system, and this increased density in vascular spacing, and hypervariability, is the deficiency of the brain following stroke. A promising biomarker that can be used to characterize blood status. For all of these conditions-for tumor progression, ischemic stroke, and brain pathology in dementia-for both determining disease progression and assessing response to treatment-of the microvascular system in brain tissue Evaluation means are needed.

Currently, MRI assessments of microvascular health are usually performed by perfusion imaging. Perfusion is the cleaning of tissue through the delivery of blood through microvessels. Such measurements can be used to assess vascular health, as vascular conditions alter the dynamics of blood flow. Perfusion MR imaging uses either intrinsic or extrinsic contrast. Exogenous contrast is usually provided by using a Gd-based contrast agent. Intrinsic contrast is obtained by a technique known as ASL (arterial spin labeling) in which blood flowing into a region of the brain is magnetically labeled. In either case, high-speed imaging techniques perform a series of imaging as the contrast moves into and out of the imaging plane. One of the important features of dynamic imaging, such as perfusion imaging, is that images acquired when the contrast is maximal on the imaging surface are captured on the imaging surface without contrast agent / blood tagging. By deducting, the differential contrast can be obtained.

When a contrast medium is used for this measurement, a mass of contrast medium is injected intravenously and images are continuously acquired as the contrast medium passes through the microcirculation. (The use of T2-weighted produces dark blood when using contrast, and T1-weighted produces bright blood, as opposed to what is observed when no contrast is used.) The contrast is imaged tissue. To enable fast data acquisition to allow multi-image tracking of the flow before leaving the region, the images typically use a variant of the fast MRI acquisition sequence known as EPI (Echo Planar Imaging). And get. Various flow-related quantities are measured to characterize the condition of the blood vessels: (MTT) mean transit time through the boxel, time to peak signal, CBF (cerebral blood flow), and CBV (cerebral blood flow). All of these amounts, which vary with vascular condition, can be measured by perfusion imaging. In addition to sequential acquisition of images acquired while a mass of contrast agent is present in the blood or magnetically labeled blood flows into the imaging surface when using ASL, the blood and surrounding tissue When the contrast between them is minimal, at least one image is taken following the passage of a mass or labeled spin through the microvascular system. This image is then deducted from the earlier image to allow calibration of the absolute signal level from the microvascular system. Time tracking by acquiring multiple images allows for flow characterization and determination of areas of impaired blood vessels.

The angiogenic vasculature is denser than healthy vessels, and the diameter and spacing of the vessels vary more. High spatial variation in vascular thickness and spacing is one of the hallmarks of the angiogenic vasculature and therefore serves as a marker of angiogenesis-related pathology with increased vascular density. However, the image resolution in perfusion imaging is not high enough to determine detailed vascular morphology. Flow contrast emphasizes locally averaged signal variability due to pathogenic flow parameters, resulting in an indirect assessment of vascular morphometry. However, the methods disclosed herein can be used to directly measure vascular density, and variability in vascular spacing, and can provide a direct robust assessment of angiogenic angiogenesis. Using the methods disclosed herein to obtain signal-to-k value data disclosing tissue texture allows for robust resolution of vascular morphometric features. This acquisition can be done in one TR fast enough that the sequence can be injected into a series of multi-imaging perfusions. To provide the best resolution, this morphometry acquisition is performed near peak contrast and is acquired either continuously at one TR or by perfusion image acquisition, either continuously or dispersed at various time points. Data is acquired by any of multiple TRs.

Obtaining a differential measurement of the vasculature by the methods disclosed herein, which obtains signal-to-k value data both with and without contrast, is a textured signal such that it results from the vascular system. Will confirm the origin of. By bringing these two measurements as close together in time as possible, the best spatial and topological correlation between the two data acquisitions allows the signal originating from the blood vessel to be accurately emphasized. ..

The best way to keep the time between two acquisitions short when using ASL contrast is by: 1) Acquiring data by the methods disclosed herein on a particular imaging surface using proton density contrast, 2) directly following the first acquisition, in the bloodstream, very close to the imaging surface. Spin-labeling on the upstream second plane, 3) acquiring spin-labeled data on the imaging surface by the method disclosed herein, the labeling and the second acquisition being the first. To get as close as possible to the time following the acquisition. The signal vs. k-value data is acquired using either the slope-on acquisition, the slope-off acquisition, or a combination of the two, and measures the signal over the desired extent in k-space. Can be provided.

Morphological parameters during measurement are expected to change significantly, so acquisition of signals over a range of k values is required to determine the basic structural features of the vessel. The breadth of the distribution of signal vs. k-values, and where their peaks are in k-space, are important features of interest. Peak assesses mean vessel density, and width assesses variability in vessel spacing. Both markers characterize angiogenesis. Acquisition can be done either by tilting on or off at the time of acquisition. Appropriate window formation in the direction of the acquisition axis of the VOI allows sampling at the targeted extent of the k value, and the exact window function determines the degree of correlation over the range of the k value. Further, hybrid acquisition is possible, where the tilt is on for some part of the data acquisition within one TR and off for the acquisition part. The purpose here is to ensure sufficient repetition of, for example, highly correlated k-value pairs while obtaining a range of k-values, allowing maximization of the SNR by averaging, while being sufficient. It is to ensure high-speed acquisition and avoid the movement of the target.

Alternatively, rather than the time-dispersed data acquisition by the methods disclosed herein in standard perfusion imaging, the methods disclosed herein are vascularized in areas exhibiting pathological flow parameters. It can be used in combination with either blood contrast method to measure morphology directly. In this measurement, both vascular spacing and variability are known markers of angiogenesis, and vascular spacing becomes more random depending on the severity of the condition. For contrast, structural contrasts, such as either T2-weighted or T1-weighted, which produce bright or dark blood, respectively, may be used. In addition, contrasts of both black and bright blood flow can be achieved by various standard methods, including arterial spin labeling. This structural measurement of blood vessels can be performed in as many tissue regions as desired, using as many acquisition directions as desired. Since the angiogenic vasculature is expected to exhibit a high degree of anisotropy, changes in the orientation of the acquisition axis between acquisitions provide another marker of pathology. Correlation of flow data from perfusion imaging with structural vascular data from application examples of the methods disclosed herein can be done by machine learning.

Acquisition by the methods disclosed herein for assessing vascular structure for ischemic stroke or for tracking angiogenesis in the vicinity of a tumor can be performed in the vicinity of the lesion. To achieve this, real-time responses to image scans defining the sites of these lesions are used to target sites for subsequent vascular structure measurements by the methods disclosed herein. Acquisitions using multiple VOIs / locations and multiple acquisition orientations can be performed to correlate with the various lesions that appear in the image sequence.

To study vascular pathologies associated with brain disorders such as dementia, VOIs can be positioned within the vascular system near the cortical areas involved in dementia. Data can be acquired in one or more VOIs in one TR. In addition, in scanners capable of parallel imaging, multiple VOIs can be defined by simultaneous recording of data for sampling the extended regions of the cerebrovascular system. For example, in dementia where multiple cortical areas appear to be damaged, VOIs can be placed in the vasculature supplying these different areas and data can be recorded simultaneously.

The usual path for diagnostic method development is feature extraction for biomarker identification from output data. Although feature extraction can define specific biomarkers, it is often necessary to expand clinical work (years, months, costs) in order to strongly correlate biomarkers with pathological conditions in diagnostic method development. This reliance on the statistical correlation between biomarkers and outcomes derived from individual tests is further hampered by the small initial study population size.

Medical data analysis is changing rapidly. Recently developed analytical techniques enable efficient determination of the overall information content of the data obtained using new diagnostic methods. In contrast to previous "raindrop" methods, current pattern recognition and machine learning techniques allow rapid correlation with other diagnostics. In this way, feature extraction and biomarker development can be done by machine learning rather than by human observation of the data. In fact, one feature (biomarker) is not isolated from the entire signal-to-k value dataset, but is correlated with data from other patients to create a strong pathological correlation. As such, focusing on the results of the acquisition of the methods disclosed herein allows for the highest possible SNR as an input to the machine learning algorithms used in this effort.

Here, computer programs are good at determining patterns within a single image and performing highly efficient data correlation with other medical history / diagnostic information. The data output from the application examples of the methods disclosed herein provides the best information content when taken in its entirety. Rather than reducing the information content by front-end feature extraction, the entire distribution of signal vs. k values obtained from each VOI is a machine learning algorithm along with relevant diagnostic data taken by current standard measurements. Is supplied to. For example, correlation data are the stages of liver disease derived from doctor's reports, liver function tests (liver serum), and physical examinations derived from histological structural images, which are the current standard measurements of liver disease. It can be attached (F0 to F5).

Alternatively, the correlation data can be a separate output from any of such tests. In sufficient numbers, this allows finer gradations to be defined in disease grading, eg, between each stage-between F0 and F1, between F0 and F2, and between F0 and F3. Multiple stages between and can be defined using this method. In addition, outcomes-progression to more advanced medical conditions, or treatment-induced cures-correlate to machine learning algorithms to correlate with texture assessments from application examples of the methods disclosed herein. May provide data.

The evaluation steps obtained by the methods disclosed herein can be mapped onto standard MRI morphological images of the affected liver. (The icon can replace the grading number for easier viewing.) This facilitates visualization of disease variation by organ. In addition, these graded values can be machine-learned to correlate possible correlations with image output obtained, for example, by MRE, standard DWI, or perfusion in the same patient.

The final example of pathological assessment using the methods herein is brain tissue. Brain pathologies are often problematic to diagnose and treat because the organs are sensitive to intervention. In addition, cognitive and behavioral changes occur over a long period of time, so underlying medical conditions can remain untested for years. In AD, there is a long precursory period, the ongoing progression of the condition at the molecular and tissue levels is inherent, and ultimately results in nerve damage. Although some have been tested in large clinical trials, there has been a surprising lack of approval for new treatments for AD or other forms of dementia. As the population grows older, the number of people suffering from the disease increases and the situation becomes dire. Some negative results of these laboratory drugs underscore the need for targeting early in the progression of the condition. However, this requires a targeted diagnosis that is in the precursory phase. Tests that identify such subjects remain elusive.

Studies have shown that gray matter is affected by the onset and progression of dementia earlier than white matter. The cortical structures known to deteriorate the earliest are the hippocampus and the entorhinal cortex, and the condition results in memory and disorientation. Recent studies applying image processing to three T1-weighted MR images of the brain have shown a statistically significant agreement between texture features in hippocampal images and the MMSE (Mini-Mental State Examination) score. .. Textures are not directly measurable by MR imaging due to inadequate resolution, but image analysis metrics, in addition to correlating with lower MMSE scores, have specific texture gradients and in the hippocampus. It correlates with reduced glucose uptake, followed by hippocampal contraction, a marker of AD. These texture features are indistinguishable except as a result of image processing, and their source is unknown. Studies have shown that these texture changes precede cognitive decline and track the development of symptoms. The hippocampus is therefore a good target for the application of the methods disclosed herein for texture measurements for pathological assessment.

Because the exact etiology of dementia in the hippocampus and entorhinal cortex is unknown, the methods disclosed herein provide a detailed assessment of tissue texture within the organ-either the hippocampus or the entorhinal cortex. Used to collect a fully complete dataset of. Both the content and variability of the texture wavelength, as well as the orientation and location dependencies, are measured. Using multiple contrast methods (because the source of the texture is unknown), signal acquisition over a range of k values in at least three (orthogonal) directions is required to fully characterize the texture. Acquisition of data across organs by defining VOI dimensions allows the spatial variability of texture to be determined so that the VOI that fits in the organ can be contained in different locations overall. By acquiring signal data over a range in k-space corresponding to wavelengths from tens of microns to about 1-2 mm, we ensure that various texture signals contribute to the information content of the measurement. The methods disclosed herein can be used in conjunction with any contrast mechanism, such as inversion recovery, T1-weighted enhanced form, or diffusion-weighted.

Predicted values of novel biomarkers provided by the texture data obtained by the methods disclosed herein for assessing the degree of AD pathology correlate with a range of diagnostic information from the same patient. Can be specified by. The main correlation markers are derived from patient outcomes-ie, a definitive diagnosis of AD or other dementia. This is because the definitive diagnosis is well downstream from the medical condition being evaluated by the Applicants, because it has the best diagnostic information content. Additional correlations are derived from MRI imaging data of patients with hippocampal contractions, a proven and continuous marker of progression of AD (as well as other forms of dementia). This correlation occurs over the long term as the disease progresses, if possible. Again, changes in tissue texture in the hippocampus are expected to predate significant cognitive decline and measurable changes in volume by MRI. The third correlated marker is FDG-PET because it is expected that a decrease in glucose metabolism will occur relatively early in the progression of the disease. As a fourth correlated biomarker, the MMSE (Mini-Mental State Examination) provides long-term data on cognitive function and decline. The genetic predisposition of AD provides an additional marker for the correlation with texture measurements obtained by the methods disclosed herein. Earlier markers provide a downstream correlation value (outcome side), whereas genetic markers are present prior to the progression of any medical condition. The correlation between this various set of biomarkers and the data obtained by the methods disclosed herein in the hippocampus and entorhinal cortex across a wide range of patients is described herein in order to predict AD pathology early. Be able to clearly define the diagnostic content using the disclosed method.

Current machine learning algorithms can classify pathological levels of non-specific features as obtained from MR data acquisition by the methods disclosed herein. As such, the correlation data sources disclosed above are input to machine learning algorithms to emphasize texture features and correlations with disease.

Studies have shown that the hippocampus can be the earliest cortical structure affected by AD progression, but its depth in the brain results in low SNR due to distance from the MR sensing coil. Become. The texture within the neocortex provides a target for the assessment of dementia and other brain pathologies, which provides higher SNR due to its proximity to the skull. In a healthy brain, very regular neuronal structures are found in the neocortex. The neurons form multiple bundles with a width of about 50 microns and a spacing of 80 microns, with about 80-100 myelinated neurons grouped together in each bundle. This is the mini-column tissue composition visible in the tissue structure of neocortical tissue. In certain areas of the brain that are thought to be affected early in AD progression by histological studies, this column sequence loses coherence in the prodromal phase. These changes occur prior to atrophy of the entire brain and are frequently used markers for AD progression, better targeting them in early diagnosis. In addition, the temporal progression of minicolumn thinning and loss of coherence over specific brain regions reflects the region-selective progression of confusional pathology in Alzheimer's disease (AD). Therefore, because each form of dementia follows a particular spatial progression across the brain, tracking spatial changes in cortical minicolumns in the brain using the methods disclosed herein allows for the classification of dementia. To.

The structure of these minicolumns in a healthy brain is shown in FIG. 27, and histological structural images are colored to reveal myelin 2702-the coating that coats the neurons. Figures 29A-29C are a series of three histological structural images colored to reveal neuronal pyramidal cells within the bundle. FIG. 29A is a neuron sequence in a healthy brain, and FIGS. 29B and 29C show progressive pathology with the progression of AD-column spacing contracts and regular structures become increasingly random.

Mini-column spacing and sequence changes in these cortical areas are early precursors to disease. Research in this area is early stage, but as with other medical conditions, basic tissue changes should precede symptoms. The problem is that there is currently no technique that can reach the resolution required to assess these early changes in column texture. The methods disclosed herein allow this measurement.

There are several methods in which the methods disclosed herein can be implemented to measure neocortical minicolumn spacing and sequence changes. This measurement, which reflects early changes, is applied to areas of the neocortex that appear to have the earliest effects on behavior in the onset of AD, eg, the temporal cortex. Since these regions are located in the neocortex outside the brain, they produce a strong SNR in the brain coil. FIG. 28 is a diagram showing possible locations of VOI2802a, 2802b, 2802c and 2802d in the neocortex 2804.

T1 contrast can be used to emphasize axon bundles against background tissue, and thus these structures, because the axon components of the neurons that form the minicolumn bundles are covered with myelin, an aliphatic substance. It is a good contrast choice when evaluating.

The difficulty of measuring column spacing arises from the semi-crystalline nature of these structures. In a healthy neocortex, they are highly ordered in the vertical (parallel to the column) direction. As a result, the VOI acquisition axis needs to block some of these columns for measurement, so the differential signal that contrasts the neuron bundles from the background tissue is extremely sensitive to the orientation of the acquisition axis. .. To measure column spacing, the VOI acquisition axis is aligned perpendicular to the length of the column. Slight misalignment reduces contrast-measurements require accurate orientation. Swinging the acquisition tilt angle over an increment angle of about 1 degree or 2 degrees to achieve proper alignment results in signal resonance at the proper alignment, i.e. a series of time-varying slopes. This is repeated to generate a locus through the three-dimensional k-space, and the resulting k-value set is directed around a specific k-value to search for resonance. (It is expected that a slight curvature of the cortex will result in a finite width in this resonance of signal amplitude vs. angle.) A locus through a three-dimensional k-space, and the resulting 10 of the first encoded specific k-value. A k-value set directed within degrees is used in an exemplary embodiment. This resonance is expected to widen as the column structure loses coherence as the condition progresses, and finally disappears when the structure becomes highly random, as shown in the histological structural image of FIG. 29C.

Using the methods disclosed herein, signal acquisition defines a finite range in k-space where the signal is acquired (nominally) at a single k-value or prior to t at the time of measurement. It can be done over a group of k-values.

Since the spacing of minicolumns in the healthy human brain is about 80 microns, sampling from about 70 microns to 110 microns results in resonance. In an exemplary embodiment, this is achieved by placing a VOI within the cortex and providing spatial encoding for k-values in the range corresponding to spatial wavelengths of 40 micrometers to 200 micrometers.

Making measurements in the cortex using the methods disclosed herein involves these basic steps: 1) A contrast mechanism is chosen to enhance the structure. 2) The individual k-value or range of k-values from which the signal is acquired is determined. 3) Timing of acquisition of k value-The number of repetitions in each k value or range of k value is determined. 4) The size of the VOI and the orientation of the acquisition axis in the neocortical tissue are selected. 5) The VOI is positioned at the center of cortical height and aligns the acquisition axis as close as possible parallel to the top and bottom surfaces of the midpoint of the VOI. 6) The signal vs. k value data is then acquired with the slope turned on or off to measure the spacing of the minicolumn. Measurements over a wide range of k values, including the average spacing of minicolumns (about 80 μm) as shown in the literature, guarantee a range of width distribution. The maximum signal strength should occur when the acquisition slope is oriented perpendicular to the column. 6) The acquisition tilt is then oscillated with a small angle increment to look for signal resonance-the sharpness-to-angle deviation of the signal resonance reflects the arrangement of the minicolumns. Sharp resonances show a regular structure. Wide resonances in response to angular shifts indicate that column degradation introduced randomness into the arrangement of the mini-columns. 7) Align the slope to the maximum signal return and sweep the entire k-value range to find the texture wavelength resonance-that is, the peak of the signal vs. k-value distribution from the texture distribution. This resonance can also be used to determine the sequence of the minicolumn. Sharp peaks (high q-values) in the signal-to-k-value curve indicate a regular structure, and the width of the curve indicates the degree of array loss. Finding resonances in the signal-to-acquisition angle and in the signal-to-k value distribution can be accomplished as a process of interaction.

Data can be acquired in the cortex or at other locations near the original VOI, either within one TR or in multiple TRs. Optimal VOI determination for cortical mini-column characterization is 1) that the VOI must fit within the cortex overall in height 2-3 mm, 2) encode axis for accurate evaluation of texture wavelengths. It is determined by the requirement to repeatedly sample the texture along with 3) signal conditions. In addition, the low VOI height along the column direction determines the alignment sensitivity.

The smaller the number of sampling iterations along the encoding axis, the wider the k-space of the signal acquired by the methods disclosed herein. Data can be obtained with tilt turned off, depending on the choice of window width to determine the range of k values that contribute to the signal. Keeping the spread in k-space small enough guarantees a correlation in the signal output.

Changes in neuron column spacing indicate progression / aging of the condition. This can be determined by long-term monitoring of the signal's peak-magnitude vs. K value distribution.

Due to the deterioration of the structure, the directional resonance becomes wider and more diffuse, and spreads over a larger acquisition angle. Also, due to structural degradation, the peaks of the signal vs. k-value distribution are wider and more diffuse, and spread over a span (wavelength) of larger k-values. Ultimately, the peak is not clear in any of the cases in which the degradation of the minicolumn sequence, which is a marker of an increased degree of dementia, progresses. In addition, changes in the width of the mini-column that advance the disease are reflected in the spacing of the column, which is another marker of the disease.

The variation of this disease marker is the degree of anisotropy of the column sequence. As the column sequence deteriorates as the condition progresses, the degree of anisotropy of the column texture also decreases, and the overall cortical tissue texture becomes more isotropic. The degree of anisotropy can be measured by using T1 or other contrast using the methods disclosed herein, and the VOI 3002 has two cortical surfaces 3004 as shown in FIG. 30, as described above. Positioned in the middle of, and comparing the signal-to-k value distribution with the acquisition axis perpendicular to the cortical surface (parallel to the minicolumn), the signal-to-k value distribution is tangential to the cortical surface 3004. Therefore, it is obtained by the acquisition axis aligned (perpendicular to the minicolumn) as shown in FIG.

The use of diffusion weighting in brain conditions, including stroke and brain tumors, is increasing. Diffusion-weighted imaging (DWI) provides indirect measurements of structure at the cellular level by applying a slope that first dephases and then rephases the signal at the target location.

Static water molecules are rephased by the second slope, but those moved between the two slopes are not rephased and therefore do not generate a signal. The difficulty with this technology is that the design makes it extremely sensitive to movement. There is also a low signal-to-noise ratio due to the delayed echo time resulting from the need for a long diffusion slope. The use of the methods disclosed herein for data acquisition when using diffuse contrast is fast enough to acquire data, while echo times are still long, with minimal blurring due to signal loss and motion. Therefore, the problem of movement can be alleviated. The methods disclosed herein can be used in conjunction with diffusion-weighted contrast to assess mini-column spacing and sequence / randomness. This measurement can be made by applying the diffusion slope parallel to the cortical surface (perpendicular to the minicolumn) and then perpendicular to the cortical surface (parallel to the minicolumn direction). These two measurements allow an assessment of anisotropy, which is best in a healthy brain, so that the condition becomes more isotropic as the column deteriorates.

Diffusion weighting is used with reference to FIG. If the mini-column is still intact, the slope as indicated by slope 13006 is applied to produce a low signal. Similarly, if the minicolumn is still intact, a slope as shown in slope 23008 is applied to produce a high diffusion signal.

When the mini-column is deteriorating, the signals in which the two different slopes approach each other increase overall as the mini-column deteriorates to a higher degree, except for the diffusion-weighted signal.

Improvements to this measurement are due to the application of diffusion gradients in multiple directions for data acquisition and the generation of diffusion tensors similar to diffusion tensor imaging (DTI), although data acquisition is the method disclosed herein. by. Diffusion tensor generation must use at least six non-coordinated directions of diffusion slope to generate sufficient data to generate the tensor, the eigenvalues of which reflect the array of minicolumns. , Determine the level of partial anisotropy (FA) in the cortex. FA needs to move and change towards a more isotropic structure as the column structure deteriorates-a FA value of 1 shows the highest anisotropy, and a value of 0 is basic. To show the maximum isotropic property of diffusion, the arrangement of the textures of the columns is clarified.

Measurements and signals when the target k-value is sufficiently regular to define texture wavelength characteristics from the literature and still apparent, as other types of contrast mechanisms are used in combination with the methods disclosed herein. It is selected by a combination of knowledge of the approximate position in k-space of the minicolumn from pre-measurements to determine the distribution of the pair k-value. One way to achieve this is to turn on the slope to provide sufficient spread in k-space during data acquisition, allowing subsequent acquisitions to sample finer, but this measurement is It can also be done using acquisition with tilt turned off.

In the cortex, the average diffusion rate (MD) of water has been found to decrease as dementia increases. Using the methods disclosed herein, it can be determined whether this is due to mini-column failure by measuring the spacing, sequence, and anisotropy of the mini-column as previously described. Signal vs. k value data obtained using the methods disclosed herein, along with cognitive assessment tests such as the MMSE test, which is a downstream neuropathological outcome, and correlation data from serum and imaging data. Can be input to machine learning algorithms.

In addition to Alzheimer's dementia, changes in the mini-column structure, or abnormal morphology in the mini-column structure, include Parkinson's disease, Lewy body dementias, muscular atrophic lateral sclerosis, and autism (Autism spectrum disorders are more (Because it is reflected in wider and denser mini-columns), and dyslexia (normal thinning of the column due to aging does not appear to occur, but results in wider spaced mini-columns), loss It also occurs in dementia and ADHD. Therefore, the methods disclosed herein can be used to assess the degree of medical condition in any of these conditions. Correlation for machine learning to determine the correlation between the measured data and the medical condition can be obtained from segmentation of cortical atrophy, MMSE, doctor's assessment of the degree of medical condition from observational data, and the like.

In addition, the methods disclosed herein are to emphasize static contrast mechanisms for highlighting changes in tissue texture, and changes in the vasculature in the vicinity of lesions that appear in MR images that may indicate stroke or tumor-related pathology. Can be used with flow contrast.

Studies have shown that gray matter is affected by the onset and progression of dementia earlier than white matter. The cortical structures known to deteriorate the earliest are the hippocampus and entorhinal cortex, the condition of which leads to memory and disorientation. A recent study applying image processing to T1-weighted MR images of the brain showed a statistically significant agreement between the texture features on the hippocampal image and the MMSE (Mini-Mental State Examination) score. Textures cannot be measured directly by MR imaging due to inadequate resolution, but in addition to their correlation with the decline in the MMSE score, image analysis metrics have identified specific texture gradients and glucose in the hippocampus. Correlates with reduced uptake of the hippocampus, which is a marker of AD, followed by hippocampal contraction. These texture features are indistinguishable except as a result of image processing, and their causes are unknown. Studies have shown that these texture changes precede cognitive decline and track the onset of symptoms. Therefore, the hippocampus is a good target for measuring texture for pathological assessment when applying the methods disclosed herein.

Since the exact etiology of dementia in the hippocampus and entorhinal cortex is unknown, the methods disclosed herein provide a detailed assessment of tissue texture within the organ-either the hippocampus or the entorhinal cortex. Used to collect a sufficiently complete dataset for. Both the content and variability of the texture wavelength, as well as orientation and location dependence, are measured. Using multiple contrast methods (because the source of the texture is unknown), signal acquisition over a range of k-values in at least three (orthogonal) directions is required to fully characterize the texture. Acquiring data across organs by defining VOI dimensions so that VOIs that fit in the organ can fit in different locations overall allows the spatial variability of the texture to be determined. By acquiring signal data over a range of k-space corresponding to wavelengths from 10 microns to about 1-2 mm, ensure that the various texture signals contribute to the information content of the measurement. The methods disclosed herein can be used in conjunction with any contrast mechanism, such as inversion recovery, T1-weighted morphology, or diffusion-weighted.

To assess the degree of AD pathology, the predicted values of novel biomarkers provided by the texture data obtained by the methods disclosed herein are correlated with a range of diagnostic information from the same patient. Can be defined. The primary correlation marker is derived from the patient's outcome-that is, the definitive diagnosis of AD or other dementia. This is because the definitive diagnosis is well downstream from the medical condition being evaluated by the Applicants, because it has the best diagnostic information content. Additional correlations are drawn from MRI imaging data of patients with hippocampal contraction, which is a proven and continuous marker of progression of AD (as well as other forms of dementia). This correlation occurs over a long period of time, if possible, as the disease progresses. Again, changes in tissue texture in the hippocampus are expected to precede significant cognitive decline and measurable changes in volume by MRI. The third correlated marker is FDG-PET because a decrease in glucose metabolism is expected to occur relatively early in the progression of the disease. A fourth correlated biomarker, the MMSE (Mini-Mental State Examination), provides long-term data on cognitive function and decline. The genetic predisposition of AD provides an additional marker of correlation with texture measurements obtained by the methods disclosed herein. The above markers provide a downstream correlation value (outcome side), but genetic markers are present prior to the progression of any condition. The correlation between this fluctuating set of biomarkers and the data obtained by the methods disclosed herein in the hippocampus and entorhinal cortex across a wide range of patients is disclosed herein for early prediction of AD pathology. To be able to clearly define the diagnostic content of the use of the method.

Current machine learning algorithms can classify pathological levels of non-specific features such as those obtained from MR data acquisition by the methods disclosed herein. As such, the correlation data sources disclosed above are input to machine learning algorithms to emphasize texture features and correlations with disease.

Although various embodiments of the present invention have been described in detail herein as provided by patent law, those skilled in the art will recognize modifications and alternatives to the specific embodiments disclosed herein. Such modifications are within the scope and gist of the invention as defined in the claims below.

Claims (37)

A selective sampling method for evaluating textures using magnetic resonance (MR).
A step of transmitting a first RF pulse selected for the first slice selection and having a first slope having a non-zero magnitude.
A step of transmitting a second RF pulse to which a second slope selected for slice-selective refocusing is applied in a region defined by the intersection of the first slice and the second slice.
The step of encoding a specific k value using the selected gradient pulse,
A step of transmitting a third RF pulse with an applied third slope, the third slope being adapted to slice-selective refocusing, with the first slice and the second slice. The step and the step of defining the volume of interest (VOI) by defining the region defined by the intersection and the third slice selection of.
Steps to turn off all slopes and
A step of recording multiple samples of the RF signal encoded by the particular k value, and
A step of recording a series of k-value samples over the vicinity of the k-value defined by the height and pulse width of the first slope, while applying the first slope as a time-dependent phase encoding.
A step of post-processing a series of k-value combinations recorded within a period of time while the non-zero magnitude slope is applied, the k-values remain correlated and all records. Is in a single TR, including steps and methods. The step of encoding the particular k-value using the selected gradient pulse applies an encoding gradient pulse to induce a phase wrap to create a spatial encoding for the particular k-value and orientation. The method of claim 1, wherein the particular k-value is determined based on the texture in the VOI, comprising a step. Following the recording of a plurality of samples of the RF signal encoded with the particular k-value, a k-value selection pulse is emitted at the selected vector coupling gradient to determine a second k-value.
The step of turning off the vector connection slope and
The method of claim 2, further comprising recording a plurality of samples of the RF signal at the second k value in the TR. In the selected vector coupling gradient, the step of applying additional k-value selection pulses for a plurality of predetermined pulses, with each k-value selection pulse determining the next k-value. ,
With the step of turning off the vector coupling slope after each pulse,
The method of claim 3, further comprising recording in the TR a plurality of samples of the RF signal at the next k-value determined by each k-value selection pulse. The method of claim 1, further comprising the step of transmitting a refocusing RF pulse prior to applying the non-zero magnitude slope. A second slope, which is non-zero and has a magnitude greater than the magnitude of the first slope, is applied as a time-dependent phase encoding, while over the K-space outside the intersample correlation imposed by the window function. A step of recording a second set of samples of a series of k-values in the neighborhood of said particular k-value, widely spaced, all recordings being within a single TR, step. The method according to claim 2, further comprising. The method of claim 6, further comprising transmitting a second refocusing RF pulse before applying the second tilt. The first slope comprises a slope applied over a duration that occurs with the series of k-values in which the recorded sample is closely spaced in k-space, whereby the sample correlates. The method according to claim 1. The specific value of k in the ^range of 0.2 mm -1 100 mm ^-1, and wherein the vicinity of the k values induced by the inclination of the size of the non-zero is a subset of the range, claims The method according to 2. The method of claim 6, wherein the second slope comprises a slope that positions the recorded sample in a k-space outside the vicinity where the k-values are correlated. The specific value of k in the ^range of 0.2 mm -1 100 mm ^-1, and wherein the vicinity of the k values induced by the inclination of the size of the non-zero is a subset of the range, the subset 10. The method of claim 10, wherein has a sample-to-sample correlation in blocks within said range. The non-zero magnitude slope is determined based on the window formation function and

Defined as, here,
G is the inclination strength,
ηmin is the lower limit of the correlation function
X is a constant
N is the number of samples
Δt is a time change,
The method according to claim 1. The method of claim 1, wherein while defining the VOI, the correlation of the recorded samples over a range of k-space created by the non-zero magnitude slope is evoked by the selection of the window forming function. .. The correction of the recorded sample of the change in k-value over the neighborhood of the k-value created by the phase encoding of the first slope is the forbidden height and pulse width of the non-zero magnitude slope. The method of claim 1, which is evoked by the selection of a k-value block using. A selective sampling method for evaluating textures using magnetic resonance (MR).
The step of transmitting the RF pulse and the slope associated with the RF pulse to define the volume of interest (VOI), and
An encoding gradient pulse is applied to induce a phase wrap to create a spatial encoding for a particular k-value and orientation determined based on the texture in said VOI as the initial value of the low SNR k-value range. Steps and
A non-zero magnitude slope is applied as a time-dependent phase encoding, while a series of k-values span the vicinity of the particular k-value defined by the non-zero magnitude slope height and pulse width. Steps to record the sample and
A step of post-processing a series of k-values recorded during the period during which the slope is applied, the k-values remain correlated and the sample size is large enough to result in high SNR. , Steps and methods, including. A selective sampling method for evaluating textures using magnetic resonance (MR).
The step of transmitting the RF pulse and the slope associated with the RF pulse to define the volume of interest (VOI), and
An encoding gradient pulse is applied to induce a phase wrap to create a spatial encoding for a particular k-value and orientation determined based on the texture in said VOI as the initial value of the high SNR k-value range. Steps and
While applying a non-zero magnitude slope as a time-dependent phase encoding, a series of k-values in the vicinity of the particular k-value that are widely spaced outside the intersample correlation imposed by the window function. A method that includes a step of recording a range of samples in a set of. 16. The method of claim 16, wherein the step of recording a series of k-value pairs of samples comprises sampling within overlapping blocks of k-values. The series of k-values within the block are combined such that the combination is sufficiently correlated to result in an improvement in SNR, and the method.
A step of post-processing a series of k-values within the block, wherein the k-values remain correlated as composite data points and the number of samples is large enough to result in a high SNR.
17. The method of claim 17, further comprising the step of using the composite data points to characterize the spectral power distribution over the range of the recorded sample to measure the texture within the VOI. The particular value of k ^is in the range of 0.2 mm -1 100 mm ^-1, and the vicinity of the k values are a subset of said range, The method of claim 15. The particular value of k ^is in the range of 0.2 mm -1 100 mm ^-1, and the vicinity of the k value is a subset of the range, method according to claim 16. A selective sampling method for assessing tissue texture using magnetic resonance (MR).
Steps to apply a contrast mechanism that enhances the contrast between component tissue types in the polyphasic biological sample being measured, and
A step of selectively exciting a volume of interest (VOI) with multiple time-varying radio frequency signals and applied slopes,
Encoding The step of applying a gradient pulse to induce a phase wrap to create a spatial encoding for a particular k-value and orientation, the particular k-value being determined based on the texture in the VOI. To be done, steps and
A step of initiating a series of time-varying tilts to produce a time-varying locus through a three-dimensional k-space of k-value encoding, the resulting k-value set required to produce the VOI image. Steps and
A step of simultaneously recording a plurality of series of samples of an NMR RF signal encoded by the k-value set,
The recorded NMR signal sample is post-processed to generate a signal-to-k value dataset for the k value within the k value set determined by the trajectory to characterize the texture features of the tissue within the VOI. Methods, including steps. The step of selectively exciting the VOI is
A step of transmitting a first RF pulse with a first slope selected for the first slice selection,
A step of transmitting a second RF pulse to which a second slope selected for slice-selective refocusing is applied in a region defined by the intersection of the first slice and the second slice.
The step of applying the encoding gradient pulse and
A step of transmitting a third RF pulse with an applied third slope, the third slope being adapted to slice-selective refocusing, with the first slice and the second slice. Selective for assessing tissue texture using the MR of claim 21, including the step of defining the intersection and the region defined by the third slice selection to define the VOI. Sampling method. The component tissue type is fibrous tissue and the step of applying the contrast mechanism is to select the contrast mechanism from the extrinsic or intrinsic contrast mechanism and between the fibrous structure and the surrounding tissue, T1, T2 or 21. The step further comprises a step of providing contrast including T2 *, IR enhancement, or functional contrast, such as diffusion or flow contrast, BOLD (blood oxygen level dependence), ASL (arterial spin labeling), and diffusion enhancement contrast. A method of selective sampling for assessing tissue texture using the MR described in. The step of applying the contrast mechanism is
Slice-selective refocusing in the region defined by the step of transmitting the first RF pulse with the first tilt selected for the first slice selection and the intersection of the first slice and the second slice. After the step of transmitting the second RF pulse to which the second gradient selected for is applied, the step of applying the first diffusion gradient during the step of selectively exciting the VOI is the step of applying the VOI. Completing the excitation is transmitting a third RF pulse with an applied third tilt, the third tilt being adapted for slice-selective refocusing and said first slice. A step comprising transmitting, defining the volume of interest (VOI) by defining the intersection of the second slice and the region defined by the third slice selection.
A step of applying a second diffusion slope, wherein the first and second diffusion slopes include a step of establishing contrast.
Selective for assessing tissue texture using the MR of claim 21, wherein the step of applying an encoding gradient pulse further comprises setting the particular k value based on the fibrous texture. Sampling method. The step of selectively exciting the VOI is to position the VOI at various points in the liver using either interleaving acquisition within one TR or measurement at multiple separate TRs. Including the step of sampling the variability of the condition in the liver
The spatial coherence of the texture defines the locus through the three-dimensional k-space of the k-value encoding for a particular k-value or k-value range during the step of recording multiple series of samples of the NMR RF signal. Selective sampling for assessing tissue texture using the MR according to claim 23, which is maintained within each defined VOI and allows signal averaging to maximize SNR. the method of. The component tissue type is bone, and the step of applying a contrast mechanism is either proton density enhancement, T1 contrast, diffusion enhancement, soluble or sclerosing bone lesions to enhance diffusion within the bone marrow. The claim further comprises the use of T2 contrast in said intraosseous lesions to emphasize fluid to classify, or the step of selecting a contrast mechanism from other contrast mechanisms to emphasize cancellous bone microstructure. A method of selective sampling for assessing tissue texture using the MR described in 21. The particular k-value and k-value set established by the time-varying series of slopes assesses tissue texture using the MR of claim 26, which covers a range of TbSp and TbTh up to 5 mm. Selective sampling method for. The component tissue type is vasculature, the step of applying the contrast mechanism further comprises selecting the contrast mechanism from extrinsic or intrinsic contrast, the contrast mechanism being ASL (arterial spin labeling), Gd contrast,. The method of selective sampling for assessing tissue texture using MR according to claim 21, comprising T1, T2, T2 *, and proton density contrast, BOLD (blood oxygen level dependent) contrast. With the steps of performing perfusion imaging to provide an assessment of vascular health / pathology,
28. The MR of claim 28, further comprising a step of defining a VOI for selective excitation, wherein the perfusion imaging sequence highlights a possible vascular pathology. A selective sampling method for assessing tissue texture. The component tissue type is a regular neuronal structure in the brain tissue, and the step of applying the contrast mechanism further includes the step of selecting the contrast mechanism from T1 weighted, T2 IR, T2 *, inversion recovery or diffusion weighted. 21. The method of claim 21, for selective sampling for assessing tissue texture using the MR of 1. The component tissue type is the regular neuronal structure in the brain tissue, the VOI is located in the cortex, and the spatial encoding is in the range of k values corresponding to spatial wavelengths from 40 micrometers to 200 micrometers. For selective sampling for assessing tissue texture using the MR according to claim 1, which is in and further aligned tangentially to the surface of the cortex. , The method of claim 21. A step of continuously repeating the time-varying series of gradients to generate a locus through a three-dimensional k-space, the resulting k-value set being directed around the particular k-value and having a signal strength. A method of selective sampling for assessing tissue texture using the MR of claim 30, further comprising stepping to position the resonance of the signal sample. The method of selective sampling for assessing tissue texture using the MR of claim 32, wherein the resulting k-value set is oriented within 10 degrees of the particular k-value. The VOI is centered on the cortex and
A step of aligning the time-varying series of slopes parallel to the top and bottom surfaces of the VOI.
In order to measure the spacing of the mini-column of the cortex, the step of acquiring signal data by turning the slope on or off over a range of k-values centered on the predicted spacing of the mini-column.
A method of selective sampling for assessing tissue texture using the MR of claim 30, further comprising. A method of selective sampling for assessing tissue texture using the MR of claim 34, further comprising swinging the acquisition slope with an angle increment to determine signal resonance. The step of selectively exciting the VOI defines the VOI dimensions so as to allow the VOI to fit in the brain to fit in different parts of the brain as a whole, spatially of the texture feature. The variability can be determined, and the acquisition gradient provides signal data over a range of k-spaces corresponding to wavelengths from tens of microns to 1-2 mm, the range of the k-value set being orthogonal to up to three. In the direction, and said method
Steps to obtain biomarkers from a selected set, including MRI imaging data of hippocampal contractions, FDG-PET, MMSE (Mini-Mental State Examination) and genetic predisposition to AD, and
To characterize the texture features of the tissue in the VOI and the biomarkers acquired into the machine learning algorithm, the signal vs. k-value dataset for the k-value in the k-value set determined by the locus. Steps to enter and
A method of selective sampling for assessing tissue texture using the MR of claim 30, further comprising receiving output from said machine learning algorithm that correlates with texture features and disease. A selective sampling method for assessing tissue texture in the brain using magnetic resonance (MR).
Steps to select a contrast mechanism from T1 weighted, T2 IR, T2 *, inverted recovery or diffusion weighted,
The step of applying the contrast mechanism and
A step of selectively exciting a first volume of interest (VOI) using multiple time-varying radio frequency signals and applied tilts, whereby the first VOI is centered on the cortex of the brain. To align a particular k-value axis parallel to the surface of the cortex at the midpoint of the first VOI.
Encoding The step of applying a gradient pulse to induce a phase wrap to create a spatial encoding for a particular k-value and orientation, said particular k-value, which provides the predicted spacing of the tissue minicolumn. A step that measures the spacing of the minicolumn in the center,
A step of initiating a series of time-varying acquisition slopes to produce a time-varying locus through a three-dimensional k-space of k-value encoding, the resulting k-value set yielding the image of the first VOI. Steps and, which are a subset of what you need to do
A step of simultaneously recording a plurality of series of samples of an NMR RF signal encoded using the k-value set to obtain the signal over a range of k-values.
A step of swinging the acquisition slope with an angle increment that determines signal resonance, wherein the sharpness of the signal resonance reflects the arrangement of the minicolumn.
Steps to select the acquisition angle and
A step of repeating a series of time-varying acquisition slopes to generate a time-varying locus through a three-dimensional k-space of k-value encoding with respect to the range of k-values swept to determine the peak of the signal vs. k-value distribution.
A step of selectively exciting a second VOI using a plurality of time-varying radio frequency signals and applied slopes, wherein the second VOI is either within one TR or within a plurality of TRs. A method comprising a step in the cortex or elsewhere near the first VOI.

Images