CN103930030A - Computer-aided bone scan assessment with automated lesion detection and quantitative assessment of bone disease burden changes - Google Patents
Computer-aided bone scan assessment with automated lesion detection and quantitative assessment of bone disease burden changes Download PDFInfo
- Publication number
- CN103930030A CN103930030A CN201280051415.0A CN201280051415A CN103930030A CN 103930030 A CN103930030 A CN 103930030A CN 201280051415 A CN201280051415 A CN 201280051415A CN 103930030 A CN103930030 A CN 103930030A
- Authority
- CN
- China
- Prior art keywords
- image
- patient
- pathological changes
- bone
- pixel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
- G06T7/0014—Biomedical image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10116—X-ray image
- G06T2207/10128—Scintigraphy
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Abstract
A computer-aided bone scan assessment system and method provide automated lesion detection and quantitative assessment of bone disease burden changes.
Description
Prioity claim
The application requires the U.S. Provisional Patent Application No.61/548 submitting on October 18th, 2011,498 and name be called the area of computer aided bone scanning evaluation that COMPUTER-AIDED BONE SCAN ASSESSMENT WITH AUTOMATED LESION DETECTION AND QUANTITATIVE ASSESSMENT OF BONE DISEASE BURDEN CHANGES(has the quantization assessment of automatization's lesion detection and osteopathia load variations) rights and interests.The application requires the U.S. Provisional Patent Application No.61/714 submitting on October 16th, 2012,318 and name be called COMPUTER-AIDED BONE SCAN ASSESSMENT(area of computer aided bone scanning evaluation) rights and interests.
Merge by reference
For all objects by reference by the U.S. Provisional Patent Application No.61/548 submitting on October 18th, 2011,498 and the U.S. Provisional Patent Application No.61/714 that submits on October 16th, 2012,318 integral body are integrated with the application.
Technical field
The present invention relates to medical imaging field.The present invention relates more specifically to bone scanning, osseous lesion and osteopathia evaluation.
Background technology
Bone tumor may originate from bone, or they may originate from other site and spread (transfer) to skeleton.For example, the secondary tumors in bone results from the carcinoma of prostate of transfer often.Image display from bone scanning goes out the pathological changes relevant to constitutional osteocarcinoma or metastatic carcinoma, and the explanation of the image from bone scanning is widely used in the diagnosis and treatment of disease.
Reported a few area of computer aided lesion detecting system for bone scanning.These technology comprise semi-automatic Image Segmentation program, and this semi-automatic Image Segmentation program spends the long time while using in clinical setting (those the semi-automatic Image Segmentation programs of people such as the people such as Erdi and Yin) often.The semi-automatic method that the people such as Erdi describe needs user that seed points is inserted in each transport zone on image, consider that the patient who suffers from bone transfer often has a plurality of disease sites, so this is the program of non-trivial.Referring to Erdi YE, Humm JL, Imbriaco M, Yeung H, the quantification bone transfer analysis of the article Quantitative bone metastases analysis based on image segmentation(that Larson SM delivers on 1406 pages of 38 phases 1401 – in 1997 of J Nucl Med based on Image Segmentation).Also referring to Yin TK, the article A computer-aided diagnosis for locating abnormalities in bone scintigraphy by a fuzzy system with a three-step minimization approach(that Chiu NT delivers on 23 phases 639 – 654 in 2004 of IEEE Trans Med Imaging in bone scintigraphy, utilize three step minimization methods by fuzzy system for locating abnormal computer-aided diagnosis).
Recently, full-automatic method by people's exploitations such as Sadik makes the osseous lesion detection of being undertaken by Image Segmentation combine with the scanning evaluation of being undertaken by artificial neural network, there is the probability of bone transfer according to patient, they are classified, thus the binary scanning grading that obtains likely " bone transfer " or may " shift without bone ".Referring to Sadik M, Jakobsson D, Olofsson F, Ohlsson M, Suurkula M, mono-kind of article A new computer-based decision-support system for the interpretation of bone scans(that Edenbrandt L. delivers on 423 pages of 27 phases 417 – in 2006 of Nucl Med Commun new for explaining the computer based DSS of bone scanning).
Although this system presents the good correlation with the assessment of the bone metastatic potential of being determined by doctor, this system does not provide the quantisation metric that continuous sweep is contrasted (metric), and the instrument of evaluation therapeutic outcome is not provided yet.
Importantly, also the result of report is not related to the true measurement of patient's benefit (for example minimizing of skeleton dependent event or the prolongation of life), the research of the basic measurement aspect of formation supervision department license.
On the contrary, developed the system strengthening for image, so that from the image standardization of continuous sweep, so that doctor makes an explanation, but this system also does not attempt to identify pathological changes.Referring to Jeong CB, Kim KG, Kim TS, the article Comparison of image enhancement methods for the effective diagnosis in successive whole-body bone scans(that Kim SK delivers on 436 pages of 24 phases 424 – in 2011 of J Digit Imaging is for the contrast of the image Enhancement Method of the efficient diagnosis in Whole body bone scan).
Before, by the bone flicker of the transfer osteopathia to loading in carcinoma of prostate, carry out quantization assessment, wherein comprised the exploitation of tolerance, for example the percentage ratio (%PABS) of the positive area in bone scanning index (BSI) and bone scanning.Referring to Imbriaco M, Larson SM, Yeung HW, Mawlawi OR, Erdi Y, the article A new parameter for measuring metastatic bone involvement by prostate cancer:the Bone Scan Index(that the people such as Venkatraman ES deliver on 1772 pages of 4 phases 1765 – in 1998 of Clin Cancer Res is for measuring the new argument of the transfer bone being involved by carcinoma of prostate: bone scanning index).Also referring to Noguchi M, Kikuchi H, Ishibashi M, the percentage ratio of the positive area that the article Percentage of the positive area of bone metastasis is an independent predictor of disease death in advanced prostate cancer(bone that Noda S. delivers on 201 pages of 88 phases 195 – in 2003 of Br J Cancer shifts is the independentpredictor to advanced prostate cancer disease death).
BSI and %PABS have all experienced initial evaluation as the Prognostic Factors of suffering from the patient of carcinoma of prostate, but spend the oversize time for calculating the method for these tolerance, because need to carry out a large amount of manual annotation to bone scanning.As evaluation therapeutic response feasible tolerance, to the evaluation of %PABS and BSI just in development.Referring to Yahara J, Noguchi M, the article Quantitative evaluation of bone metastases in patients with advanced prostate cancer during systemic treatment(that Noda S. delivers on 384 pages of 92 phases 379 – in 2003 of the BJU Int quantitatively evaluating that the bone during systematic treating shifts to the patient who suffers from advanced prostate cancer).Also referring to Morris MJ, Jia X, Larson SM, Kelly A, Mezheritzky I, after Post-treatment serial bone scan index (BSI) the as an outcome measure predicting survival(treatment that the people such as Stephenson RD deliver on the Genitourinary Cancers Symposium of 2008, serial bone scanning index (BSI) is as the measurement result of prediction survival).
Although computer aided detection (CAD) system has been applied in bone scanning analysis before, they lack the feature in embodiment of the present invention.For example, these known systems are typically only carried out lesion detection to patient's single sweep operation, and continuous sweep are not contrasted.
Summary of the invention
A kind of system and method with the bone scanning evaluation of automatization's lesion detection and the quantization assessment of osteopathia load variations that provides.
In various embodiments, with bone scanning accurate cutting and the quantification of in single sweep operation, pathological changes being carried out that carried out monitor treatment effect utilization, and the contrast of the pathological changes measured value between continuous sweep.Embodiment provides accurately a kind of and has reproducibly cut and quantize the automated system of osseous lesion to assist doctor to contrast between patient itself and patient.
The inventor has carried out analysis verification to combining the bone scanning computer assisted therapy assessment system of the quantization assessment of automatization's pathological changes cutting (comprising image standardization) and disease burden.Successful grouping to the patient of untreated patient and treatment, is used to the ability that evaluation system is evaluated therapeutic effect.
Checking shows, this system can reduce the diversity of the bone scanning analysis of manual annotations, so that the measured value that self-consistentency ground obtains is objective, can reproduce and quantize, thus for individual bulk measurement is clinical with other and laboratory result data between the associated basis that provided of expection.
Checking shows, this system can be carried out the lesion segmentation of automatization's bone scanning accurately (detection of pathological changes pixel) and the measures of quantization of pathological changes load can be provided, and the measures of quantization of this pathological changes load then can the variation at the morbid state in treatment and untreated patient for evaluation.
In each embodiment, the present invention can use computer program (software) to realize.This program can be carried out by video capturing device, read work station, server and/or other suitable device.Processing on server can contribute to mutual with centralized image file system, and contributes to bone scanning analysis report to be stored in centralized data base.This system all right (for example, via the Internet) is by remote access.
Embodiments of the present invention can be described with reference to the schematic flow sheet of equation, algorithm and/or method.These methods can realize separately or realize as an assembly of system by instruction set.Like this, the operation of each equation, algorithm, flow process and/or their combination in many ways (for example hardware, firmware and/or software) realize.Computer program instructions can be loaded on computers, so that computer program instructions provides a kind of mode realizing by the function of equation, algorithm and/or flow process appointment.
Accompanying drawing explanation
Describe the present invention with reference to the accompanying drawings.Be incorporated herein and form these accompanying drawings of a part of description exemplified with embodiments of the present invention, and be further used for explaining principle of the present invention together with explanation, and make those skilled in the relevant art can implement and use the present invention.
Fig. 1 shows according to area of computer aided of the present invention and quantizes bone scanning assessment system.
Fig. 2 shows according to the general introduction of area of computer aided bone scanning evaluation process of the present invention.
Fig. 3 shows the more detailed general introduction of the evaluation process of Fig. 2.
Fig. 4 shows according to the method for Fig. 2 based on calculate the method for normal bone standardized value RIMEDIAN with reference to image.
Fig. 5 shows the method that comprises the intensity threshold that uses expert and identification indication pathological changes.
Fig. 6 shows according to the second method of using the intensity threshold of expert and identification indication pathological changes comprising of the process of Fig. 2.
Fig. 7 shows according to the true positives pixel of the identification indication pathological changes of the process of Fig. 2.
Fig. 8 shows according to the evaluation anatomy of the process of Fig. 2 and specifically measures.
Fig. 9 shows according to the simple tolerance of all anatomical area of evaluation of the process of Fig. 2.
Figure 10 shows according to the operation scenario of the exemplary assessment system of the process of Fig. 2.
The specific embodiment
The content description providing hereinafter the example of some embodiments of the present invention.Design, accompanying drawing and description are the non-limiting examples of its disclosed embodiment.For example, other embodiment of disclosed device and/or method can comprise or can not comprise feature described herein.In addition, disclosed advantage and benefit can be only applicable to specific implementations of the present invention, and are not used in the disclosed invention of restriction.
As used in this article, term " coupling " comprises direct connection and is indirectly connected.In addition,, when mentioning first device and the second device coupling, between this first device and this second device, can there is the middle device that includes source apparatus.
Fig. 1 shows according to area of computer aided of the present invention and quantizes bone scanning assessment system 100.The image data that processing unit 108 receives from image capturing equipment 102.From these images and other image, through processing or untreated data via for storing the reference cell 104 of the data through selecting, can be obtained by processor.In each embodiment, the input that the input of processor comprises user and set 106, for example by expert to the quality of image of the evaluation of any patient status, expectation and resulting those inputs of equipment performance and setting.In some embodiments, the input of processing and/or user is carried out at least in part in image capturing equipment.
Bone scanning imaging device 102 is included in normally used any suitable bone imaging device in nuclear medicine.For example, in each embodiment, imaging device utilizes radiosiotope and radiosensitive camera, those radiosiotope and the radiosensitive camera that for example in scintillography system and device, use.Specifically, scintigraphy is a kind of associating to use radiosiotope and γ camera with indication, to accumulate the technology of radioisotopic tissue.Here, γ camera pixel intensity illustrates the regional area (event of indication osseous lesion) of the radiosiotope that accumulates in bone and hypermetabolism activity.
Reference device 104 comprises for storing any applicable instrument of data and/or data association message.Typical reference cell comprises digital data storage unit, and this digital data storage unit comprises semiconductor memory, move media memorizer (for example hard disk drive), optical memory and known similar device and equipment in calculating field.
User input equipment 106 comprises for the input of user being delivered to any suitable instrument of processor 108.In each embodiment, use for example, in keyboard, mouse, touch screen and relevant input processing instrument (personal computer) any.
In some embodiments, provide regulon or regulated station 114 1 for processor output data 109 are strengthened and quality control.Although can for example use mode identification technology Shi Gai function automatization, in each embodiment, for example, by expert (the nuclear medicine radiologist with nuclear medicine Professional knowledge), strengthen and/or correct bone scanning images/data.As picked out for example, pathological changes indication by the caused mistake of situation (joint disease and fracture) of preexist, and correspondingly by its solution.
Bone scanning is processed 108 and is comprised treatment facility, Method and Process.Applicable equipment is included in any applicable messaging device known in calculating field.Specifically, in each embodiment, any or multiple digital processing device comprising in microprocessor, personal computer, work station, large-scale parallel calculating facility and supercomputer provides applicable processing capacity.
Visualization device 110 is included in any applicable equipment known in calculating field, comprises display and printer.Display comprises CRT, LED, plasma scope, fluorescence display and el display device.Printer comprises information is fixed on to the device in tangible medium, for example laser printer and the device with similar purposes.
Graph visualization aid 112 is for making physical arrangement visual, and is particularly useful for making to indicate the bone scanning data visualization of pathological changes.The special case that visual aid comprises synthetic image, this synthetic image adopts the form that presents the image set that base image or basic covered by upper layer image.Translucent and/or the transparent quality of upper layer image make to observe simultaneously basic at least part of and upper layer image at least partly.
As shown in the drawing, processor outfan 109 by information conveyance one or more in visualization device 110, graph visualization aid 112 and regulon 114.In each embodiment, visualization device and aid make one of unregulated processor output 111 and output of the processor through regulating 115 or both are all visual.
Accompanying drawing has below been described in further detail in area of computer aided and has been quantized method and the program in bone scanning assessment system 100, carried out, is included in method and the program in processing unit 108, carried out.
Fig. 2 shows according to the general introduction 200 of area of computer aided bone scanning assessment procedure of the present invention.Initialization step 202 enables standardized test image step 204 and the evaluation to testing image, and the evaluation of testing image is comprised to identification pathological changes and generating quantification tolerance 206.As shown in the figure, follow-up testing image evaluation does not typically need repetition initialization step.
In initialization step 202, obtain one or more reference images.Typically, from for constitutional osteocarcinoma or metastatic bone cancer, indication has a plurality of patients of positive indication and obtains a plurality of reference images.From various patient populations, select one group of reference images to contribute to improve following probability: benchmark norm (reference norm) is the osseous lesion in contrast test image subsequently by indication, rather than fail to indicate the osseous lesion in contrast test image subsequently.
Initialization step 202 comprises the normalization factor of determining indication normal bone intensity.The intensity here refers to visual light intensity, for example the intensity of pixel in the image obtaining by γ camera.Initialize and also comprise: the intensity threshold of determining indication osseous lesion by reference images.In each embodiment, normalization factor and intensity threshold that storage is determined by initialization step, in order to being used in the future.In some embodiments, these values are stored in reference cell 104.
After having completed initialization step, then carry out normalization step 204 and evaluation procedure 206.In normalization step, by image capturing equipment 102 or miscellaneous equipment, obtain testing image (as represented the data of image), and according to this image of the methodological standardization further describing below.
The testing image for evaluating has been prepared in standardization 204, and this evaluation comprises pathological changes identification and tolerance generates 206.In each embodiment, in order to improve the repeatability of lesion segmentation and quantification, standardization has reduced the impact of the intensity change producing due to the dosage level at build, radioactive indicator and/or the temporal difference between tracer administration and scanning are obtained.After strength criterion, the pixel intensity of normal bone is consistent between time point, thereby can in serial patient's image, obtain reproducible lesion segmentation and quantization assessment.
Tolerance from evaluation procedure 206 provides the measures of quantization to pathological changes load.As shown in deciding step 208, can carry out standardization and evaluation to extra testing image, and not need repetition initialization step 202.After handling all testing images, arrive end step 210.
In one embodiment, be that a concrete patient makes a plurality of testing images and these testing images are processed.Whether each testing image all provides the measures of quantization to this patient's pathological changes load, and the testing image of making at different time like this provides patient health monitoring, wherein comprise disease and whether treatment is responded, stablized or no development.
Fig. 3 shows the more detailed general introduction 300 of the embodiment of area of computer aided bone scanning evaluation of the present invention.As previously mentioned, after initialization step 202, carry out normalization step 204 and evaluation procedure 206.
Initialization step 202 comprises: obtain reference images, use reference images to determine the standardized value (" RIMEDIAN ") 400 of normal bone, and use reference images to determine specific the intensity threshold (" IT of anatomical area
r") 500,600.As mentioned above, typically from a plurality of osteocarcinoma patients, obtain more than 302 reference images.
Normalization step 204 comprises: obtain testing image 304, determine testing image normal bone intensity level (" TI75 ") 314, and testing image pixel intensity is carried out to standardization 324.
As shown in equation 1, according to testing image normal bone intensity TI75 and reference images standardized value RIMEDIAN normalized factor NF.
Equation 1, normalization factor: NF=(RIMEDIAN/TI75)
Normalization factor NF is for carrying out standardization to the pixel intensity of testing image.As shown in equation 2, testing image pixel is carried out to standardization, wherein TIPI
ithe testing image pixel intensity of specific pixel, TIPIN
ithe standardized test image pixel intensity of this pixel.
Equation 2:TIPIN
i=TIPI
ix NF
After this standardization, the normal bone intensity in reference images is consistent with the normal bone intensity in testing image.
Standardization 204 has been prepared testing image for evaluating 206.Evaluation procedure comprises: indicate pathological changes 700, evaluate the specific tolerance 800 of anatomical area, and evaluate for gathering of all anatomical area measuring 900, will be described further each step below.
These tolerance provide the measures of quantization to pathological changes load.As shown in deciding step 338, can carry out standardization and evaluation to extra testing image, and not need repetition initialization step 202.After handling all testing images, arrive end step 348.
Fig. 4 shows the method 400 of calculating normal bone standardized value RIMEDIAN according to reference images.Identification anatomical area 411, carries out anatomy to benchmark bone image set and cuts apart 413, identifies normal bone intensity level 415 in each image, and determines the benchmark bone strength value 417 that represents all reference images.
In step 411, identification anatomical area.These regions represent skeleton region conventionally.In embodiment, anatomy is cut apart by image collection of illustrative plates and anatomical landmarks are below carried out recently identifying anatomical area: spinal column, rib, head, extremity and pelvis.
In step 413, auto Segmentation benchmark bone image set.Cutting apart corresponding to above-identified anatomical area here.
In step 415, in each reference images, identify normal bone intensity level.In each embodiment, in the specific region of each reference images, select to have the intensity of statistics value to represent normal bone intensity.The statistics assignment of using can be evaluated based on trial-erroneous procedures based on experience, or determine with another way well known by persons skilled in the art.
In the exemplary cases based on inventor's experience, normal bone intensity is by 75% value RI75 of the intensity block diagram in particular anatomical region
x(quantity in 1≤x≤region) indication.In embodiment, by the 75% value RI75 selecting from epiphysis district intensity block diagram
xdetermine the normal bone intensity in reference images.
In step 417, determine the benchmark bone strength value that represents all reference images.In each embodiment, this representativeness bone strength value is and 75% value RI75 above-mentioned
xthe intermediate value RIMEDIAN of set correspondence.
Fig. 5 and Fig. 6 show for determine the method 500,600 of the specific intensity threshold of anatomy from reference images.Fig. 5 shows the method 500 that comprises the intensity threshold that uses expert and identification indication pathological changes.First step 511 provides expert, for example, understanding nucleus medical image and especially suffering from the expert in patient's the field of bone scanning image of osseous lesion.In this step, the pathological changes in Expert Location reference images.In second step 513, expert's indication is evaluated.This evaluates the intensity threshold of determining indication pathological changes.
Fig. 6 shows the second method 600 that comprises the intensity threshold that uses expert and identification indication pathological changes.Step comprises annotation 611, classification 613 and definite intensity threshold IT
r.
Annotation 611 has utilized expert, routine expert as mentioned above.Here, expert annotates each reference images, to indicate pathological changes.The 613 couples of experts that classify indicate and classify, so that pixel is associated with pathological changes.In embodiment, use binary classification device system, thereby the expert of indication pathological changes is indicated and is categorized as true positives pixel, and other bone pixel is categorized as to true negative pixel.
For each anatomical area, intensity threshold determines that 615 determine single intensity threshold IT
r, contribute to reappear the classification for all patients of benchmark group.
For example, for each anatomical area, find that intensity threshold all makes the number of true positives maximize (average sensitivity of increase) for all patients in group, makes false-positive number minimize (the average specificity of increase) simultaneously.
In each embodiment, via the specific intensity threshold of anatomical area, standardization image is carried out to lesion segmentation, with by specific threshold being applied to the element filtering that is then connected on standardization image, detect the pathological changes in each anatomical area.
In addition, in each embodiment, use receiver operating characteristic curve (ROC or ROC curve) to evaluate the performance of above-mentioned binary classification device system.False positive part (FPR=false positive ratio) by the TPF in each threshold setting place labelling positive (TPR=true positives ratio) contrast feminine gender, creates curve/criterion.TPR is also known as sensitivity, and FPR 1 deducts specificity or true negative ratio.
Changing discrimination threshold (is intensity threshold (IT here
r)) to determine IT
rvalue, this IT
rvalue contributes to optimize average sensitivity and average specificity.For example, the ROC curve that true positives pathological changes pixel is labeled as to the function of false positive pathological changes pixel will typically have the best IT of indication
rthe peculiar variation of the slope of value.
Fig. 7 shows the true positives pixel 700 of identification indication pathological changes.As mentioned above, identification anatomical area 711.Make pixel intensity in each testing image anatomical area and corresponding intensity threshold IT
rmatch, carry out test 713 below.
Equation 5, the indication of pathological changes: TIPIN
r,i>IT
r
Obviously, this equation is by the standardized test image pixel intensity in particular anatomical region and the specific intensity threshold IT of anatomy that is derived from reference images
rcompare.When testing image pixel intensity is greater than corresponding intensity threshold, this pixel is the true positives pixel that indication exists pathological changes.
Fig. 8 shows and evaluates the specific tolerance 800 of anatomy.Specifically, for each anatomical area, the number Z to true positives pixel
rcount 813.In addition, in subsequent step 815, the intensity summation SUMI to all true positives pixels
r.For each anatomical area, all repeat these steps 817.
Fig. 9 shows to evaluate and gathers tolerance 900.Gather tolerance and comprise all anatomical area.
As shown in equation 6, evaluate and gather osseous lesion area 911, wherein P
arepresent the area of 1 pixel.
Equation 6, gathers bone scanning pathological changes area (BSLA)
As shown in equation 7, evaluate and gather osseous lesion intensity 913.
Equation 7, gathers bone scanning pathological changes intensity (SBLI):
Therefore, gather osseous lesion area and represent the size of the active region in bone scanning and the quantification of number, and bone scanning pathological changes intensity represents the level of bone formation activity.
In each embodiment, evaluation osseous lesion counting 915.In embodiment, utilize pathological changes area identifier pathological changes to be identified as to the separated region that comprises at least 5 contiguous pixels, each pixel is all higher than definite intensity threshold.In each embodiment; the size of pathological changes identifier and contiguous pixels group is not only considered the feature even as big as meriting attention, but also to consider whether there is following probability: the group of selected size will be subject to the impact of most common failure (for example defective scanner camera pixel) simultaneously.
Equation 8, gathers callus pathological changes counting (BSLC):
Number higher than the separated region of at least 5 contiguous pixels of definite intensity threshold.
For the reaction of quantization assessment patient to treatment, can in the process of reaction evaluation, calculate the variation from the pathological changes load tolerance between given patient's serial bone scanning.Can change to evaluate therapeutic advance and/or reaction with the percentage ratio of pathological changes load tolerance, and there is separation in describing other percentage ratio variation of each response class.For example, during bone scanning image pathological changes area change 30% or more, can think development, and can think reaction while reducing 30% or more.
In operation, can change step in above-mentioned evaluation to be suitable for image, to be derived from the data of image and the availability of reference images treatment step.For example, after testing image is carried out to standardization, can determine the specific intensity threshold of anatomical area from reference images.In another embodiment, can process a plurality of reference images collection, and corresponding RIMEDIAN and TIr value are used for to one or more testing images.Similarly, those of ordinary skills should recognize from the disclosure that the order different from the order of above-described image processing step is suitable in some cases, for example, wherein have the situation of search optimal criteria image set.The operation that therefore, can change said system is to be suitable for specific needs and restriction.
Figure 10 is exemplified with exemplary assessment system operation scenario 1000.Conventionally, use the date processing input image from reference images, to produce the quantization assessment to osteopathia load.To a certain extent, there is the patient's image in treatment interval, reaction, the disease of development or stable disease in the variation indication of the quantification labelling of disease to treatment.
As shown in the figure, can obtain the image 1020 of patient's baseline and the 6th week.Evaluation process provides the quantification disease marker of baseline and the 6th week image in the identical mode of cardinal principle.
At first, carry out anatomy and cut apart input image 1002.In Image Segmentation process, image is divided into anatomical area.In this example, selected anatomical area is spinal column, rib, head, extremity and pelvis.The anatomy of input image is cut apart to provide and is similar to the illustrative image of cutting apart of cutting apart image 1022.
The input image of cutting apart is carried out to image intensity standardization 1004.This standardization produces the image that is similar to illustration standardization baseline and the 6th week image 1024.
After image standardization 1004, carry out lesion segmentation or identification 1006.Here, the strength criterion pixel in each region at input image and the intensity threshold that is derived from the corresponding region of reference images as explained above are contrasted.The pathological changes in image is inputted in the region certain strength threshold value indication that is derived from reference images.
Optional user inspection and editor (to obtain the lesion segmentation 1008 of user approval and adjusting 114 shown in Figure 1) make to carry out manual adjustment to evaluation.For example, can solve at this moment the false positive causing due to joint disease and knochenbruch.The lesion segmentation that tool is with or without optional user inspection and editor produces the image that is similar to illustrative baseline and the 6th week pathological changes indication image 1026.
At lesion segmentation 1006 and user inspection and editor 1008(if any) afterwards, carry out the calculating 1010 of pathological changes load.In this step, determine the pathological changes counting of measured value, pathological changes intensity and specific region and/or the All Ranges of pathological changes area.In embodiment, the pathological changes area of All Ranges is sued for peace, and the pathological changes intensity of All Ranges is sued for peace.In each embodiment, gathering together with value of pathological changes counting (for example the pathological changes of All Ranges counting and) and pathological changes area and pathological changes intensity provides and quantized the instrument that pathological changes is loaded.
In each embodiment, patient reaction evaluates 1012 and comprises patient reaction's classification report, and this patient reaction's classification report illustrates for example reaction or development or stable.In embodiment, chart 1028 provides baseline and within the 6th week, has measured the quantification contrast in bone scanning pathological changes counting, bone scanning pathological changes area and bone scanning pathological changes intensity.
In some embodiments, image contrast has utilized base image and translucent coverage diagram.In one embodiment, there is the area image of painted or non-staining automatic segmentation, as the translucent coverage diagram on bone scanning image (original or standardized).So especially, can be as regulating and/or edit as required operator's supplementary means of image 114.
In one embodiment, for example, together with the translucent coverage diagram of base image (Baseline Images 1026 with lesion segmentation) and the 6th week image similar image of another treatment interval (or from), use.Here, can use color or painted strengthening the visual difference between " before and afterwards " situation.Apparently, these image contrasts provide the information that is regarded as qualitatively " intuitively " information.To a certain extent, note being kept at the details in raw video, they have also embodied the measures of quantization of osteopathia load variations.
Although described each embodiment of the present invention in the above, should understand them and only in mode for example and not limitation, be suggested.Without departing from the spirit and scope of the present invention in the situation that, can carry out in form and details various variations for those skilled in the art significantly.Like this, width of the present invention and scope should not be subject to the restriction of above-mentioned illustrative embodiments, but should limit according to following claims and equivalent thereof.
Claims (15)
1. an automatic mode, loads for the treatment of patient's bone scanning image and quantification osseous lesion, said method comprising the steps of:
The patient's bone scanning being formed by pixel image is provided;
The anatomy that described image is carried out based on atlas is cut apart, to identify the anatomical area collection being included on described image;
The intensity of described image is carried out to standardization, make the intensity of the normal bone in described image consistent with the intensity of normal bone in one or more benchmark bone scanning images;
By the intensity of the pixel in each region of described image and the region certain strength threshold value that derives from described one or more benchmark bone scanning images are contrasted, detect the osseous lesion in described each region of described image; And,
By the feature of set of pixels corresponding to the osseous lesion with detecting, quantize osseous lesion load.
2. automatic mode according to claim 1, further comprising the steps:
For described patient's bone scanning image, according to group pathological changes area, pathological changes intensity and pathological changes, count and determine that at least one quantizes osseous lesion load labelling;
For the bone scanning image of processing before, determine corresponding quantification osseous lesion load labelling; And
Based on described quantification osseous lesion load labelling to recently determining patient's reaction.
3. automatic mode according to claim 2, wherein, described quantification osseous lesion load labelling is for All Ranges accumulative total.
4. automatic mode according to claim 1, further comprising the steps:
For described patient's bone scanning image, according to group pathological changes area, pathological changes intensity and pathological changes counting, determine that at least two quantize osseous lesion load labelling;
For the bone scanning image of processing before, determine at least two corresponding quantification osseous lesion load labellings; And
Based on described quantification osseous lesion load labelling to recently determining patient's reaction.
5. automatic mode according to claim 1, further comprising the steps:
For described patient's bone scanning image, according to pathological changes area, pathological changes intensity and pathological changes, count and determine a plurality of quantification osseous lesion load labellings;
For the bone scanning image of processing before, determine corresponding a plurality of quantification osseous lesion load labellings; And,
Based on described quantification osseous lesion load labelling to recently determining patient's reaction.
6. automatic mode according to claim 1, further comprising the steps:
Selection by standardization and processed patient's Baseline Images of usining lesions showed as base image;
Selection by standardization and processed patient of usining lesions showed after a while the image of time as translucent covering image; And
The image of stack is presented to unprofessional person, as explaining that the means of what kind of variation have occurred the pathological changes load of given patient.
7. automatic mode according to claim 1, further comprising the steps:
Select patient's raw video or patient's standardization image as base image;
The translucent coverage diagram of the pathological changes that selection detects is as covering image; And,
The image that presents stack, as the degree and the visual means that distribute that make the pathological changes load of given patient.
8. automatic mode according to claim 7, wherein, described covering image is and described base image image of the same period.
9. automatic mode according to claim 2, further comprising the steps:
From a plurality of reference scan, determine normal bone standardized value, described normal bone standardized value is not that region is specific;
From described patient's bone scanning image, determine normal bone intensity level; And
By the normal bone standardized value of described reference scan and the normal bone intensity of described patient's bone scanning image, implement described normalization step.
10. automatic mode according to claim 9, further comprising the steps:
A plurality of reference scan are annotated to indicate pathological changes;
Annotation is categorized as to true positives pixel or true negative pixel; And
For each anatomical area, determine that one contributes to reappear the intensity threshold of described classification.
11. 1 kinds of automatic modes, for the treatment of patient's bone scanning image and quantification osseous lesion load, said method comprising the steps of:
Use the anatomical structure of scanner scanning bone, to produce the patient's bone scanning image being formed by pixel, described pixel has the intensity of indication bone metabolism rate;
Use anatomy collection of illustrative plates, described patient's imaging anatomy is divided into a plurality of regions;
Use, from the normal bone labelling in benchmark bone scanning image set with from the normal bone labelling in described patient's image, is carried out standardization to the intensity of patient's image pixel;
The intensity and the region certain strength threshold value that derives from reference images of using patient's image pixel, detect the pathological changes pixel in described patient's image; And
If there is pathological changes pixel in described patient's image, according to the feature of described pathological changes pixel, quantize osseous lesion load.
12. automatic modes according to claim 11, further comprising the steps:
For described patient's image, according to the number of the pathological changes pixel in All Ranges, be multiplied by elemental area and determine bone scanning pathological changes area;
For the bone scanning image of processing before, determine corresponding quantification osseous lesion load labelling; And
Based on described quantification osseous lesion load labelling to recently determining patient's reaction.
13. automatic modes according to claim 12, further comprising the steps:
For described patient's image, according to the summation of the intensity of the pathological changes pixel in All Ranges, determine the intensity amounting to;
For described patient's image, according to the summation of the pathological changes pixel in All Ranges, determine the pathological changes pixel amounting to;
According to the intensity of described total, divided by the pathological changes pixel of described total, determine described bone scanning pathological changes intensity;
For the bone scanning image of processing before, determine corresponding quantification osseous lesion load labelling; And,
Based on described quantification osseous lesion load labelling to recently determining patient's reaction.
14. automatic modes according to claim 11, further comprising the steps:
Selection can not be subject to the number k of the contiguous pixels of most common failure impact simultaneously, and described most common failure is for example defective scanner camera pixel;
For described patient's image, consider All Ranges, determine the number j of the group with k above pathological changes pixel; And
Bone scanning pathological changes counting is set to equal j.
15. 1 kinds of devices for the treatment of patient's bone scanning image and quantification osseous lesion load, described device comprises:
Radioactive tracer dosage form scanner, processor and digital data memory;
Be stored in the anatomy collection of illustrative plates in memorizer;
Normal bone labelling is derived and is stored in memorizer from reference scan, and the region certain strength threshold set of indication pathological changes is derived and is stored in memorizer from described reference scan;
Described scanner is operable as and obtains the patient's bone scanning image being formed by pixel;
Described processor is operable as and with described anatomy collection of illustrative plates, described patient's image is carried out to anatomy and cut apart;
Described processor is operable as with described normal bone labelling the intensity of the pixel in described patient's image is carried out to standardization;
Described processor is operable as with described intensity threshold and detects pathological changes pixel; And,
Described processor is operable as according to pathological changes pixel characteristic and calculates and quantize pathological changes load tolerance.
Applications Claiming Priority (5)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US201161548498P | 2011-10-18 | 2011-10-18 | |
| US61/548,498 | 2011-10-18 | ||
| US201261714318P | 2012-10-16 | 2012-10-16 | |
| US61/714,318 | 2012-10-16 | ||
| PCT/US2012/060394 WO2013059177A1 (en) | 2011-10-18 | 2012-10-16 | Computer-aided bone scan assessment with automated lesion detection and quantitative assessment of bone disease burden changes |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103930030A true CN103930030A (en) | 2014-07-16 |
| CN103930030B CN103930030B (en) | 2017-06-16 |
Family
ID=48141274
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201280051415.0A Expired - Fee Related CN103930030B (en) | 2011-10-18 | 2012-10-16 | The area of computer aided bone scanning evaluation of the quantization assessment with automation lesion detection and bone disease load variations |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN103930030B (en) |
| WO (1) | WO2013059177A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110428886A (en) * | 2018-05-01 | 2019-11-08 | 谷歌有限责任公司 | It is diagnosed using the autoradiogram of mobile device |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107274389B (en) * | 2017-05-25 | 2021-06-01 | 中国科学院苏州生物医学工程技术研究所 | Femur and acetabulum anatomical parameter obtaining method based on CT three-dimensional sequence image |
| TWI709147B (en) * | 2019-10-16 | 2020-11-01 | 中國醫藥大學附設醫院 | System of deep learning neural network in prostate cancer bone metastasis identification based on whole body bone scan images |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6246745B1 (en) * | 1999-10-29 | 2001-06-12 | Compumed, Inc. | Method and apparatus for determining bone mineral density |
| JP2002531198A (en) * | 1998-11-30 | 2002-09-24 | ホロジック， インコーポレイテッド | DICOM compliant file communication including quantitative and image data |
| JP2004508118A (en) * | 2000-09-11 | 2004-03-18 | シーメンス アクチエンゲゼルシヤフト | Method, apparatus and software for separating individual objects of a segmented anatomical structure from a 3D dataset of a medical examination method |
| CN1689020A (en) * | 2002-09-16 | 2005-10-26 | 成像治疗仪股份有限公司 | Imaging markers in musculoskeletal disease |
| CN1735373A (en) * | 2003-01-07 | 2006-02-15 | 成像治疗仪股份有限公司 | Methods of predicting musculoskeletal disease |
| US20100310141A1 (en) * | 2005-11-11 | 2010-12-09 | Wilson Kevin E | Estimating risk of future bone fracture utilizing three-dimensional bone density model |
-
2012
- 2012-10-16 WO PCT/US2012/060394 patent/WO2013059177A1/en active Application Filing
- 2012-10-16 CN CN201280051415.0A patent/CN103930030B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002531198A (en) * | 1998-11-30 | 2002-09-24 | ホロジック， インコーポレイテッド | DICOM compliant file communication including quantitative and image data |
| US6246745B1 (en) * | 1999-10-29 | 2001-06-12 | Compumed, Inc. | Method and apparatus for determining bone mineral density |
| JP2004508118A (en) * | 2000-09-11 | 2004-03-18 | シーメンス アクチエンゲゼルシヤフト | Method, apparatus and software for separating individual objects of a segmented anatomical structure from a 3D dataset of a medical examination method |
| CN1689020A (en) * | 2002-09-16 | 2005-10-26 | 成像治疗仪股份有限公司 | Imaging markers in musculoskeletal disease |
| CN1735373A (en) * | 2003-01-07 | 2006-02-15 | 成像治疗仪股份有限公司 | Methods of predicting musculoskeletal disease |
| US20100310141A1 (en) * | 2005-11-11 | 2010-12-09 | Wilson Kevin E | Estimating risk of future bone fracture utilizing three-dimensional bone density model |
Non-Patent Citations (2)
| Title |
|---|
| MATTIAS OHLSSON,ET AL.,: "Automated Desion Support for Bone Scintigraphy", 《COMPUTED -BASED MEDICAL SYSTEMS,CBMS,2009,22ND IEEE INTERNATIONAL SYMPOSIUM ON》 * |
| YUSUF E.ERDI,ET AL.,: "A System for the Quantitative Analysis of Bone Metastases by Image Segmentation", 《NUCLEAR SCIENCE SYMPOSIUM,CONFERENCE RECORD,1999 IEEE ANAHEIM》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110428886A (en) * | 2018-05-01 | 2019-11-08 | 谷歌有限责任公司 | It is diagnosed using the autoradiogram of mobile device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103930030B (en) | 2017-06-16 |
| WO2013059177A1 (en) | 2013-04-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9002081B2 (en) | Computer-aided bone scan assessment with automated lesion detection and quantitative assessment of bone disease burden changes | |
| Petrick et al. | Evaluation of computer‐aided detection and diagnosis systems a | |
| Li et al. | Unsupervised machine learning of radiomic features for predicting treatment response and overall survival of early stage non-small cell lung cancer patients treated with stereotactic body radiation therapy | |
| Nakajima et al. | Bone scan index: a new biomarker of bone metastasis in patients with prostate cancer | |
| Keek et al. | A review on radiomics and the future of theranostics for patient selection in precision medicine | |
| US20220319008A1 (en) | Automated tumor identification and segmentation with medical images | |
| CN101231678B (en) | Medical image-processing apparatus and medical image processing method | |
| US7639848B2 (en) | Lesion marking and characterization quality assurance method and system | |
| Orlhac et al. | Radiomics in PET imaging: a practical guide for newcomers | |
| Wang et al. | Computer-aided classification of mammographic masses using visually sensitive image features | |
| WO2015023732A1 (en) | Systems, methods and devices for analyzing quantitative information obtained from radiological images | |
| Brown et al. | Computer-aided quantitative bone scan assessment of prostate cancer treatment response | |
| Chakraborty | Recent advances in observer performance methodology: jackknife free-response ROC (JAFROC) | |
| Tokuda et al. | Investigation of computer-aided diagnosis system for bone scans: a retrospective analysis in 406 patients | |
| US8401255B2 (en) | Computer-assisted reconciliation of multiple image reads | |
| Aslantas et al. | CADBOSS: A computer-aided diagnosis system for whole-body bone scintigraphy scans | |
| Huo et al. | Quality assurance and training procedures for computer‐aided detection and diagnosis systems in clinical use a | |
| Liu et al. | Deep learning for the automatic diagnosis and analysis of bone metastasis on bone scintigrams | |
| Acosta et al. | The need for medical artificial intelligence that incorporates prior images | |
| Villanueva-Meyer et al. | Machine learning in neurooncology imaging: from study request to diagnosis and treatment | |
| Hofmann et al. | Machine learning approaches along the radiology value chain–Rethinking value propositions | |
| US20120041785A1 (en) | Medical image processing device, medical image grouping method, and program | |
| CN103930030A (en) | Computer-aided bone scan assessment with automated lesion detection and quantitative assessment of bone disease burden changes | |
| Dahlblom et al. | Artificial intelligence detection of missed cancers at digital mammography that were detected at digital breast tomosynthesis | |
| Nichols et al. | Texture analysis for automated evaluation of Jaszczak phantom SPECT system tests |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170616 Termination date: 20211016 |