Osteoporosis screening method based on dispersion model

By using continuous time random walk and fractional calculus model, fitting and computing more complex diffusion parameters is solved, and the problem of limited accuracy in evaluating multi-factor influence is achieved, achieving more accurate osteoporosis screening and identification of diseases.

CN119007985AInactive Publication Date: 2024-11-22FIRST AFFILIATED HOSPITAL OF DALIAN MEDICAL UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411159352.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-22
Publication Date
2024-11-22
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The traditional single-index diffusion model is limited in its accuracy when evaluating the impact of multiple factors on the diffusion process of osteoporosis, making it difficult to provide accurate screening results.

Method used

A more complex continuous time random walk model (CTRW) and fractional-order calculus model (FROC) were used to fit and calculate diffusion parameters, and six parameter values ​​for each subject were obtained, including αCTRW, βCTRW, DCTRW, μFROC, βFROC and DFROC were compared with statistical tests and classified screening for osteoporosis.

Benefits of technology

By accurately predicting the impact of multiple factors on the diffusion process, the screening accuracy and identification efficiency of osteoporosis are improved, and the accuracy and diagnostic efficiency of screening are significantly improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119007985A_ABST
    Figure CN119007985A_ABST
Patent Text Reader

Abstract

The invention discloses an osteoporosis screening method based on a dispersion model, and mainly relates to the field of functional magnetic resonance imaging. Comprising the following steps: S1, acquiring a lumbar vertebra magnetic resonance scanning diffusion weighted image of a subject; s2, sketching a region of interest in the weighted image, and obtaining a signal intensity mean value; s3, a dispersion index is calculated, and a parameter value of the single-index dispersion weighted imaging model and six parameter values of the continuous time random walk model and the fractional calculus model are obtained; s4, analyzing differences among the same dispersion parameter values of the three groups of subjects, and if the differences have significance, determining that the dispersion parameter values have relatively strong screening efficiency; and S5, performing classified screening on the types of the osteoporosis patients by using the dispersion parameters with relatively strong screening efficiency. The osteoporosis screening method has the beneficial effects that the influence of various factors on the diffusion process is accurately predicted through the more complex continuous time random walk model or the fractional order calculus model, so that the accuracy of osteoporosis screening is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of functional magnetic resonance imaging, in particular to an osteoporosis screening method based on a diffusion model. Background Art

[0002] With the increasing global aging population, osteoporosis has become an increasingly serious public health problem. Every year, about 200 million people are affected by osteoporosis, resulting in nearly 89 million fractures. The global prevalence of osteoporosis in women is about 30%, and about one-third of women over 50 years old may suffer osteoporotic fractures in their lifetime. Therefore, early screening for osteoporosis is crucial.

[0003] Magnetic resonance imaging (MRI) is a non-invasive examination method that has unique advantages in the screening of osteoporosis due to its advantages of no ionizing radiation and higher tissue contrast. At the same time, diffusion weighted imaging (DWI), as a functional MRI technology, can perform quantitative analysis of lesions.

[0004] Among them, the traditional single exponential DWI model is a simplified mathematical model that assumes that the diffusion process of water molecules is a single exponential decay. The model describes and predicts the diffusion behavior of water molecules in tissues by calculating the apparent diffusion coefficient (ADC). Although the single exponential model can provide reasonable predictions in some cases, its accuracy may be limited when it is necessary to evaluate the combined effects of multiple factors on the diffusion process because it simplifies the existence and influence of microcirculation in tissues. Summary of the invention

[0005] The purpose of the present invention is to provide an osteoporosis screening method based on a diffusion model, which accurately predicts the influence of multiple factors on the diffusion process through a more complex continuous time random walk model (CTRW) or fractional order calculus model (FROC), thereby more accurately screening and identifying osteoporosis.

[0006] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0007] A method for screening osteoporosis based on a diffusion model comprises the following steps:

[0008] S1. Obtain diffusion-weighted images of 16 b-values ​​of lumbar magnetic resonance scans of each subject in three groups: osteoporosis group, low bone mass group and normal control group, and automatically generate diffusion parameter maps based on these diffusion-weighted images, wherein the lowest b-value among the 16 b-values ​​is 0;

[0009] S2. Delineating a region of interest (ROI) in the diffusion-weighted image with b=0 in step S1, and superimposing the delineated ROI on the diffusion-weighted images of the other 15 b values, and obtaining the mean signal intensity of the ROI of the diffusion-weighted images corresponding to the 16 b values;

[0010] S3. According to the mean signal intensity of the ROI of the diffusion-weighted image corresponding to the 16 b-values ​​obtained in step S2, the diffusion parameters are fitted and calculated to obtain the parameter values ​​of the traditional single-exponential diffusion-weighted imaging (DWI) model for each subject and the six parameter values ​​of the continuous-time random walk (CTRW) model and the fractional order calculus (FROC) model;

[0011] S4. By using statistical tests, the differences between the values ​​of the same diffusion parameter of the three groups of subjects are compared. If the differences are significant, the diffusion parameter value has a strong screening efficacy;

[0012] S5. Classify and screen the types of osteoporosis patients based on the diffusion parameters with strong screening efficacy determined in step S4.

[0013] Furthermore, the six parameter values ​​in step S3 are α CTRW , β CTRW , D CTRW , μ FROC , β FROC and D FROC The quantitative parameters of the traditional monoexponential DWI model, namely, the apparent diffusion coefficient (ADC) value, are calculated based on the mean signal intensity of 16 b-values ​​in the diffusion-weighted image ROI. The ADC value of the parameter index of the traditional monoexponential DWI model is calculated based on the monoexponential equation fitting:

[0014] S / S0=e -b*aDC ,

[0015] Among them, S0 and S correspond to the signal intensity of water molecule movement under the conditions of no diffusion sensitization and diffusion sensitization, respectively, and b is the diffusion sensitization factor;

[0016] For the three diffusion parameters (α CTRW , β CTRW and D CTRW ) is calculated by the following fitting formula:

[0017] S / S0=E α [(-b·D) β ],

[0018] Among them, E αis the Mittag-Leffler function of order α, D is the anomalous diffusion coefficient, and α and β are diffusion indices reflecting the temporal and spatial diffusion heterogeneity, respectively;

[0019] For the three diffusion parameters (μ FROC , β FROC and D FROC ) is calculated by the following fitting formula:

[0020]

[0021] Among them G dδ and δ are the amplitude and pulse width of the diffusion gradient, respectively, Δ is the gradient lobe separation time, β and D describe the spatial diffusion heterogeneity of water molecules, respectively, and μ characterizes the spatial diffusion parameter.

[0022] Compared with the prior art, the beneficial effects of the present invention are:

[0023] By acquiring the diffusion-weighted image of the subject, a diffusion parameter map is automatically generated based on the diffusion-weighted image, and the region of interest (ROI) is outlined by the diffusion-weighted image to obtain the mean signal intensity within the ROI of the diffusion parameter map. After calculation, the ADC value of the traditional single-exponential DWI model and the six parameter values ​​of the continuous-time random walk (CTRW) model or the fractional order calculus (FROC) model are obtained for each subject, and the influence of multiple factors on the diffusion process is accurately predicted, so as to classify and screen osteoporosis more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Attached Figure 1 It is the receiver operating characteristic ROC curve of single parameter and multi-parameter joint diagnosis of the apparent diffusion coefficient ADC, continuous time random walk model CTRW and fractional order calculus model FROC diffusion parameter of the present invention. DETAILED DESCRIPTION

[0025] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the application equally.

[0026] Embodiment: A method for screening osteoporosis based on a diffusion model comprises the following steps:

[0027] S1. Obtain diffusion-weighted images of 16 b-values ​​of lumbar magnetic resonance scans of each subject in three groups: osteoporosis group, low bone mass group and normal control group, and automatically generate diffusion parameter maps based on these diffusion-weighted images, wherein the lowest b-value among the 16 b-values ​​is 0;

[0028] Firstly, based on the data of three groups of subjects, namely osteoporosis group, low bone mass group and normal control group, the apparent diffusion coefficient (ADC) and abnormal diffusion coefficient D CTRW and D FROC , time parameter α CTRW and the spatial parameter β CTRW , β FROC and μ FROC Specifically, three groups of subjects were selected, namely, osteoporosis group, low bone mass group and normal control group. For each subject, 3.0T magnetic resonance imaging and 32-channel spinal coil were used to perform magnetic resonance scanning on the patient's lumbar spine.

[0029] The scanning sequence and parameters were as follows: sagittal diffusion-weighted imaging (DWI) sequence used single-shot echo-planar imaging pulse sequence (SS-EPI).

[0030] Scan parameters were: repetition time / echo time 4000 / 77.7 ms, number of layers 11, layer thickness 4 mm, layer spacing 0.8 mm, FOV 300×300 mm 2 , matrix 123×176, voxel 2.44×1.70×4.00, phase encoding direction A>P, number of excitations 1-5, flip angle 90°, bandwidth 1790, b value (number of repetitions): 01, 101, 201, 401, 801, 1101, 1401, 1701, 2001, 3001, 4001, 5001, 6001, 8004, 10004, 12005s / mm 2 , a total of 16 b values ​​were collected, and the collection time was 5 minutes and 21 seconds.

[0031] After the subject undergoes an MRI scan, a diffusion-weighted image with 16 b values ​​will be automatically generated. The weighted image with b=0 is used to outline the region of interest (ROI). A total of 16×11 images and the ADC parameter image automatically generated by the MRI scanner are stored together in a mobile hard disk.

[0032] S2. Delineating a ROI in the diffusion-weighted image with b=0 in step S1, and superimposing the delineated ROI on the diffusion-weighted images with other 15 b values, and obtaining the mean signal intensity of the ROI of the diffusion-weighted images corresponding to the 16 b values;

[0033] Using ITK-SNAP (open source software; https: / / www.itk-snap.org) Open the magnetic resonance scan data stored in the mobile hard disk, including the diffusion-weighted images with 16 b values ​​and the automatically generated ADC parameter map. ROI delineation is performed on this set of 16×11 diffusion-weighted images.

[0034] Among them, in the ITK-SNAP software, the ROI was outlined within the outer boundary of the intervertebral disc from the first to the fifth lumbar vertebrae, and then the outlined ROI was superimposed on the diffusion-weighted images of the other 15 b values, and the signal intensity mean of the ROI of the 16 sets of diffusion-weighted images was automatically obtained in the ITK-SNAP software.

[0035] S3. According to the mean signal intensity of the ROI of the diffusion-weighted image corresponding to the 16 b values ​​obtained in step S2, the diffusion parameters are fitted and calculated to obtain the parameter values ​​of the traditional single exponential DWI model for each subject and the six parameter values ​​of the continuous time random walk (CTRW) model and the fractional order calculus (FROC) model; wherein the six parameter values ​​are α CTRW , β CTRW , D CTRW , μ FROC , β FROC and D FROC ;

[0036] The quantitative parameter index of the traditional single exponential DWI model is the ADC value, which is calculated based on the single exponential equation fitting:

[0037] S / S0=e -b*ADC ,

[0038] Among them, S0 and S correspond to the signal intensity of water molecule movement under the conditions of no diffusion sensitization and diffusion sensitization, respectively, and b is the diffusion sensitization factor;

[0039] For the three diffusion parameters (α CTRW , β CTRW and D CTRW ) is calculated by the following fitting formula:

[0040] S / S0=E α [(-b·D) β ],

[0041] Among them, E α is the Mittag-Leffler function of order α, D is the anomalous diffusion coefficient, and α and β are diffusion indices reflecting the temporal and spatial diffusion heterogeneity, respectively;

[0042] For the three diffusion parameters (μ FROC , β FROC and D FROC) is calculated by the following fitting formula:

[0043]

[0044] Among them G dδ and δ are the amplitude and pulse width of the diffusion gradient, respectively, Δ is the gradient lobe separation time, β and D describe the spatial diffusion heterogeneity of water molecules, respectively, and μ characterizes the spatial diffusion parameter.

[0045] S4. By using statistical tests, the differences between the values ​​of the same diffusion parameter of the three groups of subjects are compared. If the differences are significant, the diffusion parameter value has a strong screening efficacy;

[0046] It is understandable that each subject in the osteoporosis group, low bone mass group and normal control group has 7 diffusion parameter values ​​corresponding to 16 b values, namely ADC, α CTRW , β CTRW , D CTRW , μ FROC , β FROC and D FROC In this step, the diffusion parameter values ​​of the three groups of subjects, namely the osteoporosis group, the low bone mass group and the normal control group, are statistically analyzed to screen out the diffusion parameter values ​​with diagnostic efficacy.

[0047] Specifically, the same diffusion parameter value of the osteoporosis group, low bone mass group and normal control group is taken as a statistical unit. For example, there are 20 subjects in the osteoporosis group, 20 subjects in the low bone mass group and 20 subjects in the normal group. Then the ADC of 60 subjects is taken as a group. Similarly, the α CTRW , β CTRW , D CTRW , μ FROC , β FROC and D FROC As a group respectively.

[0048] Statistical analysis was performed on each group of diffusion parameter values. When there was a significant difference between the diffusion parameter values ​​of the three groups of patients, further post hoc multiple comparisons were performed. If the P value of the difference between the two groups of diffusion parameter values ​​was less than 0.05, it indicated that there was a significant difference between the diffusion parameter values ​​of the two groups of subjects.

[0049] The screening efficacy of seven diffusion parameter values ​​was evaluated based on statistical test methods, and the parameters of the continuous time random walk (CTRW) model and the fractional order calculus (FROC) model were combined to evaluate the screening efficacy, so as to obtain the diffusion parameter value and diffusion parameter model combination with the highest screening efficacy.

[0050] Get the maximum area under the curve AUC, and calculate the seven diffusion parameter values ​​ADC, α CTRW , β CTRW , D CTRW , μ FROC , β FROC and D FROC Compare and obtain the diffusion parameter with the highest screening efficiency. Among them, the area under the curve (AUC) is an important indicator to measure the sensitivity and specificity of a screening method, also known as the area under the receiver operating characteristic (ROC) curve. The effect of a screening method is evaluated by drawing the ROC curve and calculating the area under it. The value range of AUC is usually 0.5-1, 1 means complete accuracy, 0.5 means random guessing, and the closer to 1, the higher the screening accuracy of the method.

[0051] It also contains two parameters:

[0052] 1) Evaluate the performance of joint screening of parameters of the continuous time random walk (CTRW) model α CTRW +β CTRW , α CTRW +D CTRW , β CTRW +D CTRW , α CTRW +β CTRW +D CTRW .

[0053] 2) Evaluate the performance of the joint screening of parameters of the fractional order calculus (FROC) model FROC +β FROC , μ FROC +D FROC , β FROC +D FROC , μ FROC +β FROC +D FROC .

[0054] The specific method is to calculate the area under the curve (AUC) of the diffusion parameter values ​​of the osteoporosis group and the normal control group. For example, there are three models, including 1) the traditional DWI model, 2) the continuous time random walk (CTRW) model, and 3) the diffusion parameter of the fractional order calculus (FROC) model, and 15 AUCs can be drawn.

[0055] Generally, the larger the AUC is, the higher the screening efficiency is. Therefore, among the 15 AUCs, the largest AUC is selected as the preliminarily determined AUC with the largest screening efficiency and re-verified.

[0056] It should be noted that, in the process of calculating the area under the curve AUC of each diffusion parameter with screening efficacy, the critical value, sensitivity and specificity of each diffusion parameter with screening efficacy are automatically generated, and the critical value is used to screen osteoporosis.

[0057] S5. Classify and screen the types of osteoporosis patients based on the diffusion parameters with strong screening efficacy determined in step S4.

[0058] In step S4, the diffusion parameters with the highest screening efficiency are screened out. Then, for patients whose osteoporosis is unknown, the diffusion parameters with the highest screening efficiency are collected from these patients for classification screening to improve the accuracy of osteoporosis identification.

[0059] Experimental data:

[0060] The osteoporosis screening method proposed in the present invention was tested, and a total of 131 patients who underwent dual-energy X-ray absorptiometry (DXA) examinations in the past two years were selected. According to the WHO guidelines, 37 patients were diagnosed with osteoporosis, 61 patients had low bone mass, and 33 patients had normal bone mass. Figure 1 is the ROC curve of different diffusion parameter values ​​of the subjects, Table 1 is the difference in diffusion parameter values ​​of the three groups of subjects: osteoporosis group, low bone mass group and normal control group, and Table 2 is the ROC analysis results of different diffusion parameters of the subjects.

[0061] Table 1. Differences in diffusion parameter values ​​among subjects in the normal group, low bone mass group, and osteoporosis group.

[0062]

[0063] DWI, diffusion-weighted imaging; ADC, apparent diffusion coefficient; CTRW, continuous-time random walk model; FROC, fractional-order calculus model.

[0064] Table 2. ROC analysis results of different diffusion parameter values ​​in the tested patients.

[0065]

[0066]

[0067] DWI, diffusion-weighted imaging; ADC, apparent diffusion coefficient; CTRW, continuous-time random walk model; FROC, fractional-order calculus model.

[0068] In summary, the present invention uses a more complex continuous time random walk (CTRW) model or fractional order calculus (FROC) model to accurately measure the impact of multiple factors on the diffusion process, thereby improving the accuracy of osteoporosis screening and identification.

Claims

1. A method for screening osteoporosis based on a diffusion model, characterized in that: The following steps are involved: S1. Obtain diffusion-weighted images of 16 b-values ​​of lumbar magnetic resonance scans of each subject in three groups: osteoporosis group, low bone mass group and normal control group, and automatically generate diffusion parameter maps based on these diffusion-weighted images, wherein the lowest b-value among the 16 b-values ​​is 0; S2. Delineating a region of interest (ROI) in the diffusion-weighted image with b=0 in step S1, and superimposing the delineated ROI on the diffusion-weighted images of the other 15 b values, and obtaining the mean signal intensity of the ROI of the diffusion-weighted images corresponding to the 16 b values; S3. According to the mean signal intensity of the ROI of the diffusion-weighted image corresponding to the 16 b-values ​​obtained in step S2, the diffusion parameters are fitted and calculated to obtain the parameter values ​​of the traditional single-exponential diffusion-weighted imaging (DWI) model for each subject and the six parameter values ​​of the continuous-time random walk (CTRW) model and the fractional order calculus (FROC) model; S4. By using statistical tests, the differences between the values ​​of the same diffusion parameter of the three groups of subjects are compared. If the differences are significant, the diffusion parameter value has a strong screening efficacy; S5. Classify and screen the types of osteoporosis patients based on the diffusion parameters with strong screening efficacy determined in step S4.

2. The osteoporosis screening method based on a diffusion model according to claim 1, characterized in that: The six parameter values ​​in step S3 are respectively α CTRW , β CTRW , D CTRW , μ FROC , β FROC and D FROC The quantitative parameters of the traditional monoexponential DWI model, namely, the apparent diffusion coefficient (ADC) value, are calculated based on the mean signal intensity of 16 b-values ​​in the diffusion-weighted image ROI. The ADC value of the parameter index of the traditional monoexponential DWI model is calculated based on the monoexponential equation fitting: S / S0=e -b*aDC , Among them, S0 and S correspond to the signal intensity of water molecule movement under the conditions of no diffusion sensitization and diffusion sensitization, respectively, and b is the diffusion sensitization factor; For the three diffusion parameters (α CTRW , β CTRW and D CTRW ) is calculated by the following fitting formula: S / S0=E α [(-b·D) β ], Among them, E α is the Mittag-Leffler function of order α, D is the anomalous diffusion coefficient, and α and β are diffusion indices reflecting the temporal and spatial diffusion heterogeneity, respectively; For the three diffusion parameters (μ FROC , β FROC and D FROC ) is calculated by the following fitting formula: in and δ are the amplitude and pulse width of the diffusion gradient, respectively, Δ is the gradient lobe separation time, β and D describe the spatial diffusion heterogeneity of water molecules, respectively, and μ characterizes the spatial diffusion parameter.