WO2005044104A1

WO2005044104A1 - Method of quantifying blood flow through heart muscle

Info

Publication number: WO2005044104A1
Application number: PCT/JP2004/016192
Authority: WO
Inventors: Hajime Sakuma; Takahiro Natsume; Hisato Maeda
Original assignee: Mie Tlo Co., Ltd.
Priority date: 2003-11-06
Filing date: 2004-10-25
Publication date: 2005-05-19
Also published as: JP4446049B2; JP2005137558A

Abstract

A method of quantifying the blood flow through a heart muscle is provided. A dynamic scan is performed on the subject to whom a magnetic resonance contrast agent is injected, and time-series data is obtained. The blood flow is calculated based on the data obtained in such a way that the dynamic state of the flow of the blood passing through the heart muscle is accurately reflected on the calculation result. The result is displayed with a numerical value or an image. The method of quantifying the blood flow through the heart muscle uses electrocardiographically gated first-pass MRI using an MRI contrast agent having a T1 reduction effect. In the method, a means for correcting a body movement, a means for setting a processing of a region of interest (ROI), a means for creating an input curve and an output curve using the signal strength varying depending on the variation with time of the concentration of the magnetic resonance imaging contrast agent, a means for performing signal strength saturation correction of the input curve, and a means for determining the slope (K1) of the linear portion of the Patlak plot obtained from the corrected input and output curves are provided.

Description

Description Quantification method for myocardial blood flow

In the present invention, a dynamic scan is performed by injecting a magnetic resonance contrast agent into a subject, and blood flow dynamics of the magnetic resonance contrast agent passing through the myocardium (perfusion) based on time-series data obtained by the scan. The present invention relates to a method for quantifying myocardial blood flow, which displays a calculation result reflecting the above as a numerical value or an image. Background art

Ischemic heart disease is the leading cause of death in the United States and Europe, and about 70,000 people die each year in Japan from ischemic heart disease. In diagnosing ischemic heart disease, it is important to accurately evaluate myocardial blood flow as well as morphological evaluation of coronary stenosis.

Myocardial blood flow is directly related to myocardial oxygen supply, and reduced myocardial blood flow is a sharp reflection of ischemia. Until now, myocardial blood flow in ischemic heart disease has been performed mainly by nuclear medicine, but it has the disadvantage that it cannot detect subintimal ischemia due to low spatial resolution. Using magnetic resonance imaging (MRI), a clinical study was started in the early 1990s to evaluate the cardiac blood flow distribution by observing the blood flow dynamics of a rapidly intravenously injected MR contrast medium through the myocardium. Was. MRI is an image that excites the nuclear spin of the subject tissue placed in a static magnetic field with a high-frequency signal having the Larmor frequency, and reconstructs an image from the magnetic resonance signal generated by this excitation. It is a diagnostic method.

At the beginning of clinical research, the myocardial imaging range was limited, the temporal resolution was inadequate, there were many artifacts, and the diagnostic accuracy was insufficient. Only With the development of high-speed MRI in the heart region, the combination of k-space partitioned Daladiant Echo and echo-brainer data acquisition enables ultra-high-speed imaging of T1 (longitudinal relaxation time) -weighted images with little image distortion artifacts. It has become.

At present, it is possible to repeatedly acquire a dynamic MR image with a spatial resolution of 3 mm and a slice number of about 8 every two heartbeats. Myocardial perfusion testing by MR I has excellent spatial resolution, and even in studies using visual evaluation, detection of myocardial ischemia is superior to SPE CT (single photon emission computed tomography) using nuclear medicine myocardial perfusion products Therefore, the development of an objective ischemia diagnosis method using a quantitative analysis method is strongly desired.

(Patent Document 1)

JP-A-6-26 9424

(Non-Patent Document 1)

Go Matsuda, et al., “Current status of ischemic heart disease testing using MR I”, Journal of the Japanese Society of Radiological Technology, June 2001, Vol. 57, No. 6, p. 6 64-6 70

(Non-patent document 2)

Hakuma Sakuma, "Practical Use of Contrast Agents for MR I in the Cardiac Area", Nippon Medical Review, October 2004, Vol. 62, No. 12, p. 682-6 89

To date, quantitative analysis of myocardial blood flow by MRI has shown that the maximum slope of the upslope and the peak signal intensity of the time-signal intensity curve (TSC) in the myocardial region are determined. Evaluations have been performed as indicators, but their quantitativeness and diagnostic accuracy are limited. In the field of nuclear medicine, quantitative measurement of tissue blood flow is performed by performing a comparative analysis of blood (input) and tissue (output) dynamics of a tracer. Regarding myocardial blood flow analysis by MRI, it is expected that myocardial blood flow diagnosis can be performed with higher spatial resolution than nuclear medicine by performing a comparative analysis of blood input and tissue output. However, to quantitatively evaluate myocardial blood flow using MR contrast agents, it is necessary to solve the following five technical issues. (1) The problem that the MR signal shows a saturation phenomenon in the high-contrast-concentration region and blood flow quantification is lost. (2) Breathing during acquisition of dynamic MRI images-Influence of heart position fluctuation due to body motion. (3) The problem of loss of blood flow quantification due to uneven sensitivity distribution of MR signal receiving coil. (4) The problem that the difference between the background signal in the blood flow analysis and the arrival time of the contrast agent depending on the myocardial site deteriorates the blood flow quantification. (5) There is no method to effectively and quantitatively display the blood flow gradient from the obtained myocardial blood flow data to the myocardial intima side and epicardial side and the blood flow reserve.

An object of the present invention is to provide a method for quantifying myocardial blood flow, which can accurately evaluate myocardial blood flow. Disclosure of the invention

In order to solve this problem, a method for quantifying myocardial blood flow according to the present invention comprises a method for diagnosing myocardial blood flow by electrocardiogram-gated first-pass magnetic resonance imaging (Gated First-Pass MRI) using a magnetic resonance contrast agent having a T1 shortening effect. In the method for quantifying the flow, a means for correcting body movement, a means for setting processing of a region of interest (ROI), and an input and output curve are created from a signal intensity corresponding to a change with time of the magnetic resonance contrast medium concentration. Means for performing signal strength saturation correction of the input curve, and means for calculating the slope K i of a linear portion of the Patrac plot obtained from the corrected input curve and output curve. It is the gist.

In this case, when measuring the signal intensity in blood with the magnetic resonance contrast agent, a magnetic resonance contrast agent having a low contrast agent concentration is used, and when measuring the signal intensity in myocardial tissue, Using a normal concentration of magnetic resonance contrast agent Good to be.

Further, it is preferable that the means for performing body motion correction sets a joint point between the left ventricle and the right ventricle and moves the processed image so that the joint point becomes a fixed point.

Further, the means for performing the signal intensity saturation correction of the input curve corresponding to the time-dependent change of the magnetic resonance contrast agent concentration prepares a quantitative line relating to the relationship between the contrast agent concentration and the signal intensity (concentration signal curve) in advance. It is good to find a regression line by the least squares method using the data of the low concentration area on the curve, and to set the regression line to the relationship between the contrast agent concentration and the signal intensity (signal intensity saturation correction function) in the entire concentration area. .

Further, it is preferable that the value is obtained in each segment of the myocardium, the base of the heart is arranged around the apex, and the value is displayed concentrically at each angle divided from the side wall.

Further, it is preferable to provide a means for correcting non-uniform signal detection sensitivity in the left ventricular myocardium and the left ventricle lumen.

The myocardial blood flow quantification method according to the present invention having the above-described configuration corrects the MR signal saturation phenomenon in the high-contrast-concentration region, and enables quantitative analysis of abnormal blood flow and blood flow reserve. This is a new method using a contrast agent. It is also a method of correcting left ventricular movement due to breathing. In addition, this method comprehensively corrects the effects of non-uniform signal intensity due to coil sensitivity in the left ventricular myocardium and left ventricular lumen, including the three-dimensional positional relationship between the myocardial and blood regions of interest. Further, this is a method of correcting the signal intensity saturation of the blood input curve. In addition, in quantitative blood flow analysis, this is a signal processing method that automatically sets the approximate range of background subtraction processing and Patrac plot for many regions of interest in the entire myocardium. This is an image processing and display method that integrates and draws information on the myocardial blood flow of the entire left ventricle and information on the blood flow reserve or the myocardial blood flow gradient from the intima to the adventitia. Brief Description of Drawings

FIG. 1 is a diagram showing extraction of a 64 × 64 matrix image including a left ventricular myocardial region from a 256 × 256 matrix left ventricle short axis image captured by a Gated First-Pass MRI examination according to the present invention.

Figure 2 shows the relationship between MR contrast agent concentration (in this case, G d-DTPA concentration) and signal strength.

Figure 3 shows the configuration of a polar coordinate map in which the apex is located at the center and the base is located at the center in a concentric circle at each angle obtained by dividing the myocardial blood flow value of each segment (l to n) from the side wall. It is.

FIG. 4 is a diagram showing extraction of a 64 × 64 matrix image including a left ventricular myocardial region from a 256 × 256 matrix left ventricle short-axis image captured by the Gated First-Pass MRI examination according to the present invention. .

FIG. 5 is a diagram showing a temporal image of the fourth slice of the left ventricular short-axis tomographic image after extracting the 64 × 64 matrix region at the time of the rest examination.

FIG. 6 is a diagram showing an image in which a myocardial region and a region of interest (ROI) are set in a left ventricular short-axis tomogram and a left ventricular normal region.

FIG. 7 is a diagram showing input and output curves obtained from the case of FIG. The input curve [C a (t)] is the T S C in the left ventricular lumen, and the output curve [C b (t)] is the T S C in the left ventricular myocardium (side wall).

FIG. 8 is a diagram showing a Patrac plot created using the input / output curves of FIG. The straight line in the graph is the regression line of the linear part of the Patrac plot (6 points from 2 forces), and the slope is a value reflecting myocardial hemodynamics.

FIG. 9 is a graph showing the relationship between various concentrations of Gd-DTPA (mol / L) mixed with blood and MR signal intensity (5 examples). The straight line on the graph is linear approximation using data up to Gd-DTPA concentration of 0.7 mmol / L, and the obtained regression line is used as the signal intensity saturation function. Figure 10 shows the input curves before and after the correction of signal intensity saturation at rest in a case in which a proportional relationship holds between Gd-DTPA concentration and MR signal intensity at 1/5 dose and no ischemia. FIG. 7 is a diagram showing a correction input curve.

FIG. 11 is a diagram showing a polar coordinate map of an example of correcting the signal detection sensitivity of the MR coil in a case where ischemia was not observed.

FIG. 12 is a diagram showing a value pole coordinate map and a CFR polar coordinate map of a case diagnosed as right coronary artery (RCA) stenosis.

FIG. 13 is a diagram showing a comparison table between the values and the maximum slope of the upslope. BEST MODE FOR CARRYING OUT THE INVENTION

Hereinafter, an embodiment of a method for quantifying myocardial blood flow according to the present invention will be described with reference to the drawings. First, a method is described that corrects the phenomenon of MR signal saturation in the high-contrast-concentration region and enables quantitative analysis of abnormal blood flow and blood flow reserve.

In humans or non-human animals, the extracellular fluid distribution or intravascular distribution with T1 shortening effect was evaluated in order to evaluate abnormal blood flow and blood flow reserve in the myocardium or other organs. Using a magnetic resonance imaging method, which has the function of repeatedly injecting a magnetic resonance contrast agent as shown below into a vein and repeatedly taking images with an imaging time interval of less than 4 seconds, the signal in both the observed blood and tissue We have invented a method that detects temporal variations in intensity and enables quantitative analysis of abnormal blood flow and blood flow reserve.

Magnetic resonance contrast agents containing paramagnetic lanthanide and exhibiting extracellular fluid or intravascular distribution (eg, Gd-DTPA, Gd-DTPA-BMA, etc.) 0.001-0.01 mmol / kg Dose is administered after dilution in saline, dextrose or water to make the same volume as that of the magnetic resonance contrast agent 0.01 to 0.075 mimol / kg, and repeated at an imaging time interval of less than 4 seconds. When using a magnetic resonance imaging method with the function Detect intermittent fluctuations.

In addition, the above-mentioned magnetic resonance contrast agent is administered at a dose of 0.01 to 0.075 mimol / kg, and the blood is collected by using a magnetic resonance imaging method having a function of repeatedly performing imaging at an imaging time interval of less than 4 seconds. It detects temporal changes in signals in both the tissue and the tissue.

When assessing blood flow reserve, administer a drug that has a dilating effect on arteries, arterioles or capillaries in the tissue, and administer the above magnetic resonance contrast agent at a dose of 0.01 to 0.075 mmol / kg. Then, using a magnetic resonance imaging method having a function of repeatedly performing imaging with an imaging time interval of less than 4 seconds, a temporal variation of a signal in both blood and tissue is detected.

Quantitative diagnosis of tissue blood flow and blood flow reserve using magnetic resonance contrast agents requires that the contrast agent concentration and signal intensity be linear. As a result of a basic study conducted by the inventor, various types of magnets that have the function of performing repeated imaging when 0.01 to 0.075 mmol / kg of a magnetic resonance contrast agent containing gadolinium, which is a paramagnetic lanthanide, are rapidly intravenously administered. On the images obtained by resonance imaging, the signals observed in the myocardium and other tissues show a linear relationship with the contrast agent concentration, but there is a saturation between the signal observed in blood and the contrast agent concentration. A phenomenon occurred and the linearity was lost.

On the other hand, when a gadolinium-containing magnetic resonance contrast agent of 0.001 to 0.01 mmol / kg is administered, the blood signal is linear with the contrast agent concentration on the images obtained by various magnetic resonance imaging methods that have the function of performing repeated imaging. Relationship.

Based on these results, dilute 0.001-0.01 mmol / kg of magnetic resonance contrast agent into saline, dextrose solution, or water to make the same volume as that of the 0.01-0.075 mmol / kg dose of magnetic resonance contrast agent. Imaging in the blood by adjusting the circulatory dynamics after administration of the contrast agent and performing imaging using magnetic resonance imaging, which has the function of performing repeated imaging, is performed. The agent concentration can be quantitatively determined from the signal of the magnetic resonance image. In this case, using a magnetic resonance contrast agent of 0.01 to 0.075 mmol / kg dose, keep the high-frequency output and reception sensitivity of the magnetic resonance imaging apparatus constant, or record the high-frequency output and reception sensitivity and perform imaging. It detects temporal variations in signals in both blood and tissue. Furthermore, do not administer a drug that has a dilating effect on arteries, arterioles or capillaries in the tissue, perform imaging using a magnetic resonance contrast agent at a dose of 0.01 to 0.075 mmol / kg, and examine the blood and tissue after loading. Of the signal is detected. This method enables quantitative evaluation of tissue blood flow and tissue blood flow reserve from quantitative measurement of changes over time in contrast agent concentration in both blood and tissue.

Next, a method of correcting temporal movement of the left ventricle due to respiration (body motion correction) will be described. When setting a region of interest (R〇I) in contrast dynamic MRI, the temporal movement of the left ventricle (LV) due to respiration becomes a problem. Therefore, we invented a method to easily correct the movement of LV over time using anatomical indices.

The junction of the right ventricle (R V) and LV (! ^ ー point) is used as an anatomical index. First, an RV-LV point is set for each temporal image of the slice to be processed in the left ventricular short-axis tomographic image. Then, by moving the processed image up, down, left, and right (X-Y) so that the set RV-LV point has constant coordinates, the time-dependent movement of LV due to respiration in each time-series image is obtained. Correction becomes possible. In addition, after the LV movement due to respiration is corrected to facilitate later data processing, extraction is performed to images with an arbitrary matrix number based on the scale point.

After setting the RV-LV point, it is possible to automatically correct the movement of the LV over time due to respiration and extract an arbitrary number of matrices into the image.

Next, the non-uniformity of coil signal sensitivity in the left ventricular myocardium and left ventricular lumen Will be described. When quantitatively measuring myocardial blood flow and coronary blood flow reserve using contrast-enhanced dynamic MRI, the non-uniformity of signal detection sensitivity depending on the distance from the signal detection coil becomes a problem. Thus, in order to quantitatively measure myocardial blood flow and blood flow reserve, a method for three-dimensionally correcting nonuniform signal detection sensitivity in the left ventricular myocardium and left ventricular lumen was invented.

In contrast dynamic MRI, the signal detection sensitivity nonuniformity correction coefficient of the left ventricular myocardial region was calculated using the image of the phase before the contrast agent flowed into the left ventricular myocardial region. Since the T1 value of the left ventricular myocardium does not change significantly in the region of the myocardium and the T1 value can be assumed to be almost the same within the left ventricular myocardial region, the signal intensity distribution in the left ventricular myocardial region before the inflow of the contrast agent The non-uniformity of the signal and the difference in signal strength between slices are considered to almost reflect the signal detection sensitivity of the MR coil system.

First, a region of interest (ROI) is set in the left ventricular myocardial region of the left ventricular short-axis tomogram. The ROI is divided into arbitrary segments to correct signal detection sensitivity nonuniformity, and the relative signal strength to the average signal strength of all segments in the ROI in all slices of the left ventricular short-axis tomogram is calculated in the ROI of each slice. Was calculated for each segment. Using the obtained relative signal intensities, the signal detection sensitivity non-uniformity correction coefficient of each segment in the ROI in each slice was calculated. By multiplying the obtained non-uniform signal detection sensitivity correction coefficient of the left ventricular myocardial region by the measured myocardial blood flow value, it is possible to correct the non-uniform signal detection sensitivity of the left ventricular myocardial region.

The non-uniform correction coefficient for signal detection sensitivity in the left ventricle lumen was calculated using the correction coefficient for signal detection sensitivity in the left ventricular myocardial region. As shown in Fig. 1, within the ROI set in the left ventricular myocardium of the slice in which the ROI was set in the left ventricle lumen, 45 degrees, 135 degrees, 222 degrees, and 31 degrees from the center of the left ventricle Four points ((X1, Y1), (Χ2, Υ2), and the midpoint of the endocardial and epicardial ROIs at each angle of 5 degrees Find the coordinates of (X3, Y3), (X4, Y4)).

The signal detection sensitivity correction coefficient for each pixel in the ROI set in the left ventricle lumen is calculated by interpolation using the correction coefficient of the left ventricular myocardium at the obtained coordinates of the four points. Since the correction coefficient of the left ventricular myocardial region is calculated in consideration of the unevenness of signal detection sensitivity between slices, the signal detection of the left ventricular cavity calculated using the obtained correction coefficient of the left ventricular myocardial region In the sensitivity nonuniformity correction coefficient, the nonuniformity of the signal detection sensitivity between slices is considered. By multiplying the obtained signal detection sensitivity non-uniformity correction coefficient of the left ventricle cavity by the left ventricle space of each temporal image, it is possible to correct the signal detection sensitivity non-uniformity of the left ventricle space.

Next, a method of correcting the signal intensity saturation of the blood input curve will be described. As shown in Fig. 2, the relationship between the MR contrast agent concentration (in this case, G d-DTPA concentration) and the signal intensity is represented by the linear part y1 in the low concentration area and the exponential function part (saturation characteristic part) in the high concentration area. ) It consists of y2. From the intravenous Gd concentration of 0.01 mmol / kg to the estimated left ventricular concentration of 0.7 mmol / l, the linearity of the contrast agent concentration in the left ventricular blood and the signal intensity was confirmed by phantom and volunteer tests using contrast agents. It is clear that is secured.

The relationship between contrast agent concentration and signal intensity (concentration signal) was obtained by mixing blood samples and various contrast agent amounts in advance and imaging samples with various contrast agent concentrations under the same conditions as those used in contrast dynamic MRI. Curve). A straight line approximation by the least squares method is performed using data in the low concentration area that has a substantially linear proportional relationship, and the regression line is used for the case where a proportional relation is established between the contrast agent concentration and the signal intensity in the entire contrast agent concentration region. The relationship between contrast agent concentration and signal intensity (signal intensity saturation correction function).

Before performing a resting or stress test, perform a test (lZn dose test) in which the concentration of the contrast agent is reduced to 1 / n of this test. The value of the dilution factor n is a value that does not cause saturation in the blood MR signal. Next, a rest test is performed, and the peak of the input curve for the rest test is Calculate the signal intensity saturation ratio at the peak of the input curve of the resting test from the signal intensity ratio of the peak of the 1 / n dose test to the peak. The signal intensity saturation ratio is applied to the concentration signal curve and the signal intensity saturation correction function to calculate the contrast agent concentration in the left ventricular cavity at the peak of the input curve. Next, the signal strength reduction in each phase of the input curve is corrected using the density signal curve and the signal strength saturation correction function. In the stress test, signal intensity saturation correction of the input curve is performed using the contrast agent concentration at the peak of the input curve obtained by the rest test. Next, a method of automatically setting the approximate range of back-round subtraction and Patrac plot in blood flow analysis based on the two-compartment method will be described. MR contrast agents, such as Gd-DTPA, pass through the capillary membrane and distribute to the extracellular fluid, but are not taken up into cells. Therefore, a model is assumed in which the protein is taken up from the capillary tube into the extracellular fluid at a constant K i and excreted from the extracellular fluid into the capillary vessel at a constant k ₂ . Here, assuming that the Gd—DTPA concentration in the artery is C a (t) and the Gd—DTPA concentration in the myocardial region is C b (t), the temporal change of C b (t) is Is given by the first order differential equation.

[Equation 1]

d ^ ^Cb ^ y2, [Ca (¾-k ₂ [Cb ()]

dt In this model, in the phase earlier than the average transit time of Gd-DTPA, the concentration of Gd—DTPA in the myocardial region is low, and Gd—DTPA by the k ₂ [C b (t)] term Assuming that the effect of excretion is negligible, a model represented by the first-order differential equation of Equation 2 can be assumed.

[Equation 2] d [Cb (i)] d [Ca {t)]. _T ^ _r _, Here, Vm is a constant indicating the volume of G d-DTPA distribution pooled in the myocardium. Equation 3 is obtained by integrating Equation 2.

[Equation 3]

[Cb ί = Vm [Ca (?)] + «I ^ C (r) T Further, by dividing both sides of Equation 3 by [C a (t)], Equation 4 is obtained.

[Equation 4]

Equation 4 can be regarded as a linear equation like Equation 5.

[Equation 5]

By plotting Equation 5 on a graph, a Patrac plot can be obtained. Also, by performing a straight line approximation by the least squares method on the straight line portion of the Patrac plot, the inclination of the straight line portion, ie, Ki can be obtained. In this case, this is a constant representing the rate of ingestion of extracellular fluid from the capillaries into the myocardial region, and is considered to reflect myocardial blood flow.

In addition, in Gated First-Pass MRI, enormous processing is required to perform the above-described Patrac plot analysis on all slices and all segments of the left ventricular myocardial region. He invented a method to automate the analysis of truck plots for use in actual clinical tests.

The time-signal intensity curve of the left ventricle lumen (input curve) and the time-signal intensity curve of the left ventricular myocardium (output curve) were used for this Patrac plot analysis. Use. In order to quantitatively calculate myocardial blood flow, it is necessary to subtract the signal (background) of each tissue from the input / output curve, so the tissue signal intensity in the left ventricular cavity and myocardial region from the image before the inflow of the contrast agent Is calculated and subtracted as the background of each curve.

Subsequently, the time phase immediately before the time phase at which the differential coefficient of the input curve becomes maximum is set as the rise time of the input curve. Then, until the time when the differential coefficient of the output curve becomes maximum, the slowest time when the increase in the signal strength due to the contrast agent becomes almost 0 is set as the rise time of the output curve. The signal strength of the time phase before the rise time of each curve is 0.

In this Patrac plot analysis, linear approximation by the least-squares method is performed using the data of the linear portion of the Patrac plot created from the input / output curve, and the slope of the obtained regression line is used to determine the myocardial blood flow [ml / min / 100g] is calculated. First, a track plot is created using the input / output curve after the background subtraction. Subsequently, the minimum value of the Patrac plot is detected to determine the straight-line approximation range, and the next phase after the detected point is set as the approximation start point. The myocardial blood flow is calculated by performing linear approximation using five points (five scans) including the approximation start point as the approximation range. The processing for determining the rise time of the output curve enables automatic processing of the Patrac plot analysis taking into account the case where the inflow start time of the contrast agent differs depending on the part of the myocardium.

Next, a method to effectively display myocardial blood flow and blood flow reserve in the entire left ventricle using high-resolution MR images of multiple slices and to overlay and visualize the intimal to epicardial blood flow ratio I do. The myocardial blood flow of all slices and all segments calculated by Gated First-Pass MRI data processing can be displayed on one image by displaying it in polar coordinates. The polar coordinate display method of myocardial blood flow has the advantage that not only can the myocardial blood flow of the entire left ventricle be evaluated at one time, but also that it can be easily associated with coronary arteries.

The method of displaying myocardial blood flow in polar coordinates is used in nuclear medicine examinations, but M Since RI has high spatial resolution, it can separate and evaluate myocardial blood flow on the intima and adventitia, which was difficult in nuclear medicine. Since the myocardial blood flow decrease accompanying coronary artery stenosis occurs strongly on the intima side, the ratio of the intima side to the adventitia side myocardial blood flow is important information for diagnosing the pathological condition of ischemic heart disease. In addition, the myocardial blood flow at rest and during loading is quantitatively evaluated by contrast-enhanced MRI, and the ratio between the two is calculated to calculate the myocardial blood flow reserve quantitatively.

The present inventors have invented a method for quantitatively and comprehensively displaying the distribution of myocardial blood flow and myocardial blood flow reserve, or the distribution of myocardial blood flow and myocardial intima side blood flow Z epicardial side blood flow ratio. .

Using the point set at the center of the left ventricular cavity of each slice as the center point, the left ventricular myocardial region is divided at an arbitrary angle (for example, every 10 degrees) so that the side wall becomes 0 degree, and the no-track analysis is performed. By doing so, the average myocardial blood flow (kl) in each segment is calculated. As shown in Fig. 3, a polar map is created in which the apex is located at the center and the base is located at the periphery in a concentric manner at each angle obtained by dividing the myocardial blood flow value of each segment from the side wall.

By displaying the myocardial perfusion at rest, myocardial perfusion at load, and perfusion reserve on a color scale in each segment, myocardial perfusion and perfusion reserve can be comprehensively displayed. In addition, by displaying the myocardial intimal blood flow and the epicardial blood flow ratio in each segment on a color scale, the blood flow gradient in the myocardium can be quantitatively displayed.

Examples>

The method of quantifying myocardial blood flow according to the present invention was performed on 12 heart disease patients, 8 males, 4 females, and an average age of 64 ± 9 years.

Gd—DTPA (0.05 mmol / kg) was injected at a rate of 4 ml / sec by bolus injection, and a GE 1.5T high-speed MR device for heart (Signa CV / i) and a GE EPI-compatible phased array coil for heart were used. And used for imaging. Hybrid EPI (TR = 6 to 7 msec, TE = 1.4 msec, ET = 4) is used for the imaging pulse sequence Then, images were taken with a slice thickness of 10 mm, a slice gap of 2 mm, FOV of 33.9 cm x 33.9 cm, and a collection matrix size of 128 x 128.

During imaging, electrocardiogram synchronization (2R-R interval) triggered by R wave was performed under respiratory arrest, and 30 time-lapse images of each slice were captured. In the stress test, 0.56 mg / kg of dipyridamole was intravenously injected as a drug load, and hand grip load was also used.

Before the test at rest, the test was performed at 1/5 dose (0.01 mmol / kg: 15 doses below) of the amount used at rest and load test to correct the signal intensity saturation of the input curve. After performing the test, a resting test was performed, and a load test was started 15 minutes after the resting test was completed. The collected and reconstructed MRI images were stored in digital imaging and communications in medicine (DIICOM) format and transferred to a personal computer (PC) via CD.

PC was used to construct the MRI data processing program. Turbolinux 8 Workstation was used as the operating system (hereafter OS), and FORTRAN and C were used as programming languages. G77 and gcc in GNU compiler collection (hereinafter GCC) version 2.95.3 were used for FORTRAN, C language compiler and linker, respectively. To improve the operability of the processing program, a program based on a graphical user interface (GUI) was created. GCC and GIMP tool kit + (hereinafter GTK +) version 1.2.10 library and GIMP drawing kit (hereinafter GDK) library were used to create the program for the GU I part.

First, the correction of the signal intensity saturation of the input curve will be described. In the region where the G d—DT PA concentration is very low (0.7 mmol / l or less; see Fig. 9), the G d—DT PA concentration and the signal intensity are almost proportional, but the G d—DT PA concentration is high. As a result, the proportional relationship between them does not hold. Bolus injected Saturation correction of signal intensity is required in the left ventricular cavity where G d—DT PA passes at a relatively high concentration, and a proportional relationship is established in advance even at the maximum signal intensity. 1 Left ventricle at Z5 dose Signal intensity saturation correction of the input curves for each test was performed using the time-intensity signal curve (TSC) in the cavity.

In order to obtain a relationship between the Gd-DTPA concentration and the signal intensity, 11 samples of phantoms were prepared by mixing Gd-DTPA from Ommol / 1 to 6 mmol / L with blood collected from five cases. By imaging the phantom, the relationship between the Gd-DTPA concentration and the signal intensity in the hybrid EPI sequence was determined. Linear approximation was performed using data up to 0.7 mmol / l where a proportional relationship was established between the Gd-DTPA concentration and the signal intensity, and the obtained regression line was used as a signal intensity saturation correction function.

Furthermore, the experimental values were approximated by a fifth-order polynomial. The Gd-DTPA concentration at the peak of the input curve was calculated from the ratio of the peak in TSC of the lumen of the left ventricle at the 1/5 dose to the peak signal intensity of the input curve. Then, the signal intensity of the input curve is converted from the relationship between the Gd-DTPA concentration and the signal intensity to the Gd-DTPA concentration, and the corrected signal intensity is calculated from the concentration obtained using the signal intensity saturation correction function. Calculated.

Next, the correction of the non-uniform signal detection sensitivity of the coil will be described. Since the signal detection sensitivity in MRI differs depending on the distance from the coil, the signal intensity increases from the anterior wall near the coil to the septum in the left ventricular short-axis tomogram. Since the T1 value of the tissue in the myocardial region is almost uniform, the nonuniform detection sensitivity was corrected using the signal intensity of each region in the myocardium before the inflow of the contrast agent. First, in the image before contrast agent inflow, the average signal intensity in the region of 5 ° to 20 ° within the ROI set in the myocardial region is calculated, and a shallow profile curve is created for the left ventricular short-axis tomographic image of each slice did. Here, in order to reduce the influence of signal fluctuation, the obtained perimeter profile curve is approximated by Fourier series, and the approximate curve force, 2

From 17 the signal intensity in each region of the myocardium was determined. Next, the relative sensitivity of each region to the average signal intensity (S _my ) of the left ventricular myocardium calculated by Equation 6 is calculated, and the reciprocal is calculated as the sensitivity correction coefficient (C _seg (angle, slise) ), The detection sensitivity non-uniformity was corrected by multiplying the calculated value of each area by the sensitivity correction coefficient corresponding to that area (Equation 8).

[Equation 6] w 360

. 2 ∑ ∑ U I) 60: x «)

[Equation 7]

C _seg (angle, slice)-

[Equation 8]

K _cl ^ ngle, slici) = K _bl (a? Igie _y slice) x (angle, slice) where n is the number of processing slices, S _seg (angle, slise) is the average signal strength of the area, and K _bl ( angle, slise) is the value before sensitivity correction, and K _cl (angle, slise) is the K value after sensitivity correction.

Next, the value calculation method will be described. As shown in Fig. 4, first, a 64X64 matrix region including the left ventricular myocardium is extracted from the temporally reconstructed short-axis tomographic image of the left ventricle of the MRI image with the matrix size of 256x256, The correction for the movement of the left ventricle over time was manually performed. Figure 5 shows an example of the image over time. This is a time-lapse image of the fourth slice of the left ventricular short-axis tomogram (64-year-old female, without ischemia) after extraction of the 64X64 matrix area at the time of the rest examination. The value shown below each image is the elapsed time (seconds) from the start of imaging. Bolus injection It is confirmed that the obtained G d-DTPA contrast agent passes through the right ventricle and flows into the left ventricular myocardial region after a further time from the left heart lumen.

G d— DPTA flows into the myocardium and several points are set at the left ventricular lumen and the left ventricular myocardium in the image where the border of the left ventricular myocardium can be confirmed, and a cubic spline function is used between each point. A region of interest (hereinafter referred to as ROI) was set in the left ventricular cavity and left ventricular myocardium by interpolation as shown in Fig. 6. The TSC obtained from R〇I set in the left ventricular cavity was used as the input curve for Patrac plot analysis, and the left ventricular myocardium was divided into regions in the range of 5 ° to 20 °. The TSC obtained from the region was used as the output curve. Figure 7 shows the input and output curves actually obtained.

As shown in the figure, after subtracting the background from the obtained input curve C a (t) and output curve C b (t), a Patrick plot shown in Fig. 8 was created, and a minimum The gradient was calculated for each slice and each region from the approximate straight line by the multiplicative method. No ,. To evaluate the linearity of the linear portion of the track plot, a correlation coefficient with the obtained approximate straight line was calculated. In order to visually evaluate the distribution of values throughout the left ventricular myocardium, a polar map was created using the _Kt value as a parameter, using the polar map display method often used in nuclear cardiology examinations.

Next, the calculation of the coronary artery blood flow reserve will be described. Coronary blood flow reserve (CFR) is obtained from the ratio of resting coronary blood flow to maximum coronary blood flow after drug loading, and is an excellent indicator of the degree of functional stenosis of coronary arteries. Non-invasive CFR measurement includes evaluation of functional stenosis of coronary stenosis lesions, percutaneous transluminal coronary angioplasty (PTCA), evaluation of restenosis after intervention with stents, coronary artery bypass It is expected to be applied to diagnosis of graft stenosis. In this example, the CFR was calculated from the value as an index of myocardial perfusion obtained from MRI and equation (9). [Equation 9] ^ ―-stress

Here, K _stress is the value obtained from the _stress test, and K _rest is the value obtained from the test at _rest . Also, a polar coordinate map was created using CFR as a parameter in the same way as the values.

Figure 9 shows the relationship between the Gd-DTPA concentration and the signal intensity obtained from the experiment (average of 5 cases). In the region where the Gd-DTPA concentration was up to 0.7 mmol / l, the Gd-DTPA concentration and the signal intensity showed an almost proportional relationship, but beyond that, the proportional relationship was not established. Similar results were obtained for each of the five cases. The regression line (signal intensity saturation correction function) calculated using data up to 0.7 mmol / l was y = 731.4x + 164.7, and the correlation coefficient was 0.9978. Results The experimental value was approximated by the fifth-order polynomial is ^{^{found, y = 0.978x 5 - 17.38x 4}} + 122.1x 3 + 447.2x 2 + 983.4x + 143.2 is obtained, with the subsequent two relations A saturation correction was performed on the input curve.

FIG. 10 shows the result of correcting the signal intensity saturation of the input curve in a case where ischemia was not observed. In the input curve at rest before saturation correction, the peak signal intensity was low due to the nonlinearity between the Gd-DTPA concentration and the signal intensity, and the shape was 5 times the TSC obtained with the 1 Z 5 dose and the shape Different. The input curve after the saturation correction almost coincided with the TSC shape obtained at the 15 dose, and the effect of correcting the saturation of the signal intensity was recognized.

The average value and standard deviation of the correlation coefficient with the regression line in the linear part of the Patrac plot of each slice and region in all 12 cases were 0.9962 soil 0.0108 in the resting test (n = 1656). The correlation coefficient was 0.9971 +0.0041 for B-temple examination (n = 1656), and 0.9967 ± 0.0081 for the whole (n = 3312) including both tests, indicating high linearity. Figure 11 shows the value polar map and the sensitivity correction coefficient polar map before and after correction of nonuniformity in detection sensitivity in a case where no ischemia was observed (at rest). Since the positional relationship with the detection coil is shown in the figure, the value was high from the front wall to the septum before the correction of non-uniformity in detection sensitivity, but a more uniform polar coordinate map was obtained after the correction. In this example, the mean value and standard deviation of the values of the entire left ventricular myocardium before correction were 0.018i0.007, the coefficient of variation (CV) was 39.2%, and the values after correction were 0.016 ± 0.002 and 10.5%, respectively. The mean and standard deviation of the CVs for the four patients without ischemia were 34.6 ± 9.0% before correction and 10.4 ± 1.9% after correction.

Figure 12 shows the K 丄 value polar map and the CFR polar map of a case (57-year-old, male) diagnosed with right coronary artery (RCA) stenosis. In the K-value polar map, no local decrease was observed at rest, and under load, a decrease was observed in the region from the side wall force to the lower wall.The CFR polar map also showed a decrease in CFR in that region. Was observed. This area was almost identical to the area controlled by RCA.

The kinetic model of G d—D TPA in the body in this method is finally expressed by Equation 5 under hypothetical conditions, in which case the Patrac plot is a straight line on the graph. All slices and all regions of 12 cases have a straight line portion in at least 5 frames, and good linear approximation results (correlation coefficient 0.9967 ± 0.0081) are obtained, and the above assumption is considered to hold.

The 1/5 dose test and resting test were performed under the same conditions, and the input curve at rest at 5 times the TSC in the left ventricular lumen at 1 Z5 dose was originally It can be assumed that they are almost equal. However, the shape of the input curve actually obtained during the resting test was significantly different from that at the time of the 15 dose due to the nonlinearity between the signal intensity and the Gd-DTPA concentration. Is recognized. By performing the saturation correction as shown in Fig. 10, the shape of the input curve at rest is almost the same as the shape at the time of the 15 dose, and the saturation according to the present embodiment is obtained. Correction is useful, and more accurate Ki value calculation has become possible.

Since the signal detection sensitivity at the MRI differs depending on the distance from the coil, non-uniformity correction of the detection sensitivity is required for accurate value calculation. As a result of the detection sensitivity unevenness correction, a more uniform polar coordinate map was obtained in the case where ischemia was not observed as shown in Fig. 11, and the average CV in 4 cases was from 34.6%. This is reduced to 10.4%, which confirms the validity of the sensitivity correction method of this embodiment and is considered indispensable for high-precision value calculation.

Gated First-Pass MRI was used to calculate myocardial TSC, calculate the maximum slope of upslope from the TS volume of the myocardial region during the first-pass of the contrast agent, and reported a semi-quantitative evaluation of myocardial hemodynamics. I have. However, as shown in the table of Fig. 13, the maximum slope of upslope was determined in four cases in which ischemia was not observed.As a result, the average CV was 15.65%, and the fluctuation was smaller than the value of 8.52%. large. In addition, although it has been reported that the peak signal intensity of TSC in the myocardial region is used as an indicator of myocardial hemodynamics, the value fluctuates greatly and the correlation with the measurement results using positron emission tomography (PET) is low. It has been reported. These semi-quantitative methods for assessing myocardial hemodynamics are directly affected by the decrease in the bolus of the contrast agent, and their quantitative and diagnostic capabilities are considered to be limited.

On the other hand, this method is an analysis based on a kinetic model of G d-DTPA in the body, and its gradient value is calculated from a linear approximation using at least five points. It is less affected by fluctuations, and enables a more accurate myocardial hemodynamic test than myocardial hemodynamic measurement using other indices.

In the case where ischemia was not observed, the K i value polar coordinate map at rest and during loading was almost uniform, and the K i value polar coordinate map in the case of right coronary artery stenosis showed that In addition, a decrease in the Ki value was observed. In other cases, a distribution of values consistent with the findings was obtained. It is considered that the Ki value obtained by using the method reflects myocardial hemodynamics.

CFR is an excellent indicator of functional stenosis of the coronary artery and impaired myocardial microcirculation, and its measurement required intracoronary Doppler flowwire and PET. Insertion of Doppler flowwire into the coronary artery is invasive, and PET is expensive, so available facilities are limited. However, the CFR calculated from the index value of myocardial hemodynamics obtained by the present method showed a decrease in the area of control of stenotic vessels, and the CFR calculated by the present method reflected coronary artery stenosis. However, the possibility of noninvasive CFR measurement was suggested. Except for correction and R_〇 I set of body motion due to breathing in all process, it is possible to automate the creation of polar coordinate map in which the K _t value as a parameter. This shortened the processing time for creating a polar coordinate map and reduced the burden on the operator. In addition, this method has a higher spatial resolution than nuclear medicine examinations, is not affected by attenuation of scattered rays and gamma rays, which is a problem in nuclear medicine examinations, and has a faster and more accurate noninvasive myocardium. Hemodynamic tests are possible.

As described above, one embodiment of the method for quantifying myocardial blood flow according to the present invention has been described.However, the present invention is not limited to such an embodiment, and various modifications may be made without departing from the gist of the present invention. Of course, it can be implemented in an embodiment.

Claims

The scope of the claims

1. In a method of quantifying myocardial blood flow by Gated First-Pass MRI using a magnetic resonance contrast agent having a T1 shortening effect, a means for correcting body movement and a region of interest Means for setting (ROI) processing; means for creating input and output curves from signal intensities corresponding to temporal changes in the magnetic resonance contrast agent concentration; means for performing signal intensity saturation correction of the input curves; A method for quantifying myocardial blood flow, comprising: means for determining a slope K i of a linear part of a Patrac plot obtained from the corrected input curve and output curve.

2. When the signal intensity in blood is measured with the magnetic resonance contrast agent, a magnetic resonance contrast agent having a low concentration of the contrast agent is used, and when the signal intensity in myocardial tissue is measured, a normal concentration is used. 2. The method for quantifying myocardial blood flow according to claim 1, wherein a magnetic resonance contrast agent is used.

3. The method according to claim 1, wherein the means for performing body motion correction sets a junction between the left ventricle and the right ventricle, and moves the processed image so that the junction is a fixed point. Method for quantifying myocardial blood flow.

4. The means for performing the signal intensity saturation correction of the input curve corresponding to the time-dependent change of the magnetic resonance contrast agent concentration creates a quantitative line related to the relationship between the contrast agent concentration and the signal intensity (concentration signal curve) in advance. The method is characterized in that a regression line by the least-squares method is obtained on the curve using the data in the low-density region, and the regression line is defined as a relationship between the contrast agent concentration and the signal intensity in all the concentration regions (signal intensity saturation correction function). The method for quantifying myocardial blood flow according to any one of claims 1 to 3, wherein:

5. Obtain the above values for each segment of the myocardium, arrange the base of the heart around the apex, and concentrically form each angle divided from the side wall! The method for quantifying myocardial blood flow according to any one of claims 1 to 4, wherein the method further comprises displaying a ェ value.

6. The method for quantifying myocardial blood flow according to any one of claims 1 to 5, further comprising means for correcting nonuniform signal detection sensitivity in the left ventricular myocardium and the left ventricle lumen.