Method for establishing clinical interpretable PD diagnosis model based on multi-sequence magnetic resonance brain characteristics

Abstract

The invention discloses a method for establishing a clinical interpretable PD diagnosis model based on multi-sequence magnetic resonance brain characteristics, and aims to promote more accurate diagnosis of PD by extracting multi-level and multi-dimensional brain image information from multi-modal brain magnetic resonance image data and constructing a diagnosis model by means of machine learning. The method comprises the steps of brain image data preprocessing, feature standardization and screening, model construction, independent verification set and cross-disease set verification, a group of brain function and metabolism features taking a PD abnormal perfusion mode and nigra iron distribution as the principal thing are obtained, and the constructed model has high diagnosis efficiency in a training set and the independent verification set. Secondly, clinical interpretability of the model is promoted through curve calibration, DCA and CIC analysis and correlation analysis of predicted individual scores obtained based on the model and clinical symptoms. The diagnostic efficiency of the constructed PD diagnostic model in different disease sets shows that the model has good diagnostic efficiency on PSP and MSA diseases.

Description

Technical Field

[0001] The present invention belongs to the field of brain imaging, and in particular relates to a method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features. Background Art

[0002] Parkinson's disease (PD) is the second most common neurodegenerative disease. Its diagnosis currently relies on the patient's clinical symptoms and their response to conventional anti-Parkinson drugs, which is subjective and empirical. In recent years, the auxiliary diagnostic role of imaging in the revision of PD clinical diagnostic criteria has received increasing attention. However, the diagnosis of PD still lacks effective clinical reference standards, and the medical care of PD patients remains an international focus. Therefore, it is crucial to develop objective imaging markers for PD diagnosis and disease assessment.

[0003] PD has a complex pattern of brain damage, involving extensive changes in brain structure, function, and metabolism. Noninvasive magnetic resonance imaging (MRI) technology can obtain multidimensional information of brain tissue in a single scan, providing an effective means to evaluate PD neurodegeneration. Previous researchers have used the technical advantages of MRI to explore the brain degenerative characteristics of PD and attempted to build objective biomarkers based on these findings. A PD diagnostic model constructed using MRI structural data from an international multicenter cohort database reported the beneficial value of MRI in PD diagnosis, but the model's diagnostic efficacy was moderate and lacked independent data verification. More studies have used small samples and single-modality imaging data to explore the diagnostic value of PD, and as the sample size increases, the model's diagnostic efficacy tends to decrease, which hinders the clinical translation of related models.

[0004] Therefore, based on PD brain imaging technology, capturing the characteristics of PD neurodegeneration by integrating large-scale sample sizes with brain structure and functional indicators may help us understand the disease heterogeneity of PD, thereby promoting the further development of disease diagnosis and clinical evaluation markers. Summary of the invention

[0005] The purpose of the present invention is to provide a method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features in response to the current lack of objective PD imaging biomarkers. The present invention transforms multimodal MRI (quantitative susceptibility imaging, high-resolution T1-weighted imaging, and arterial spin labeling imaging) into brain imaging features that reflect PD degeneration, and uses machine learning methods to construct a diagnostic model with the selected features, thereby promoting the improvement of PD diagnostic efficiency and verifying its generalization ability in multiple disease sets.

[0006] The objective of the present invention is achieved through the following technical solutions: The present invention provides a method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features, comprising the following steps:

[0007] (1) T1-weighted images, enhanced sensitivity weighted angiography (ESWAN) images, and arterial spin labeling (ASL) images were obtained from PD patients, normal controls (NC), progressive supranuclear palsy (PSP) patients, multiple system atrophy (MSA) patients, and essential tremor (ET) patients. PD patients and normal controls were divided into training sets and independent validation sets in proportion, and PSP patients, MSA patients, and ET patients were used as cross-disease datasets.

[0008] (2) Segment the cortical gray matter of the T1-weighted image obtained in step (1) to obtain 68 cortical regions and 1 TIV; post-process the ESWAN image obtained in step (1) using the STAR-QSM algorithm to obtain a quantitative susceptibility map (QSM), and segment the QSM map using the DeepQSMSeg algorithm to obtain 10 subcortical nuclei including the bilateral substantia nigra, red nucleus, putamen, caudate nucleus, and globus pallidus; perform principal component analysis-proportional subsection model analysis on the ASL image obtained in step (1) to obtain the abnormal perfusion pattern of PD;

[0009] (3) extracting brain image features for each cortical region, subcortical nucleus, and PD abnormal perfusion pattern segmented in step (2); and standardizing the brain image features using a general linear model;

[0010] (4) The least absolute shrinkage and selection operator regression (LASSO) method was used to select multiple brain imaging features with the highest contribution as input, and a PD diagnosis model was constructed using binary logistic regression and trained based on the training set;

[0011] (5) According to step (4), the individual prediction scores obtained based on the PD diagnostic model were correlated with the corresponding clinical symptoms. The top six brain imaging features with the highest contribution were screened using SHAP (SHapley Additive exPlanations) analysis. Inter-group differences were compared in the training set, independent validation set, PSP, MSA, and ET patient data sets, and FDR (False Discovery Rate) correction was performed.

[0012] Furthermore, in step (1), a fast spoiler gradient sequence is used to obtain a T1-weighted image; a gradient echo sequence is used to obtain an ESWAN image; and a pulsed continuous arterial spin labeling technique is used to obtain an ASL image.

[0013] Furthermore, in step (2), the cortical gray matter of the T1-weighted image obtained in step (1) is segmented to obtain 68 cortical regions, specifically: the subcortical nuclei of the T1-weighted image are segmented using FreeSurfer software, and the recon-al preprocessing process is used to perform automated head motion correction, skull stripping, spatial normalization, spatial registration, and cortical segmentation, and then based on the Desikan-Killiany atlas, 68 cortical regions and total intracranial volume (TIV) are segmented;

[0014] The ESWAN image obtained in step (1) is post-processed using the STAR-QSM algorithm to obtain a quantitative susceptibility map, specifically: the original phase map is unwrapped using Laplace-based phase unwrapping, and the spatial low-frequency components of all brain tissues are retained; the background phase is removed using spherical mean filtering, thereby specifically obtaining brain tissue phase data; the susceptibility boundary is estimated and magnetization artifacts are removed; the tissue susceptibility map is obtained, and the QSM image is calculated based on this;

[0015] The DeepQSMSeg algorithm tool sets the true value to the historical data of manual and semi-automatic segmentation; after using the DeepQSMSeg algorithm to segment the QSM image to obtain the subcortical nuclei, it also includes: performing manual correction to correct the misalignment of the surrounding tissue voxels caused by the alignment deviation.

[0016] Furthermore, in step (2), the ASL image obtained in step (1) is subjected to principal component analysis-proportional subsection model analysis to obtain the abnormal PD perfusion pattern, specifically:

[0017] (2.2.1) The ASL image and the cerebral blood flow (CBF) image generated based on ASL were preprocessed by SPM 12, including: head motion correction, spatial registration and normalization of the ASL image and the CBF image using DARTEL, CBF image standardization, and spatial smoothing; the spatial smoothing used a Gaussian kernel with a full width at half maximum (FWHM) of 6 mm;

[0018] (2.2.2) ASL_PDRP was constructed on the CBF images preprocessed in step (2.2.1) using the principal component analysis-scaled sub-profile model analysis (PCA-SSM) method, including: constructing individual sub-profiles, principal component analysis (PCA), scaled sub-profile model analysis (SSM), and constructing PD-related cerebral perfusion patterns;

[0019] (2.2.3) The topological profile score (TPR) was used to calculate the individual ASL_PDRP expression score in the independent validation set and PSP, MSA, and ET patient datasets.

[0020] Further, in step (2.2.2), based on step (2.2.1), the constructed individual sub-profile performs voxel-level analysis on the CBF maps of all participants in the training set and represents them as high-dimensional vectors; the principal component analysis performs matrix centering and covariance matrix decomposition, and outputs the eigenvector and score corresponding to each principal component; the principal component with the highest score is selected to form a proportional sub-profile model, and the expression value ASL_PDRP of each participant on the proportional sub-profile model is calculated;

[0021] In step (2.2.3), based on the proportional sub-profile model obtained based on the training set in step (2.2.2), the individual standardized CBF map based on the independent validation set and the cross-disease dataset obtained by the TPR algorithm is multiplied and accumulated with the weights of the corresponding voxels in the proportional sub-profile model to obtain the individual's overall score on the proportional sub-profile model as the ASL_PDRP obtained based on TPR.

[0022] Furthermore, in step (3), the brain image features extracted for each cortical region, subcortical nucleus and PD abnormal perfusion pattern segmented in step (2) are specifically:

[0023] (3.1) For the QSM image, the average magnetic susceptibility values ​​of the 10 subcortical nuclei were calculated as 10 brain imaging features;

[0024] (3.2) For T1-weighted images, 68 average values ​​of cortical thickness and 1 TIV were calculated based on the Desikan-Killiany atlas as 69 brain imaging features;

[0025] (3.3) For ASL images, ASL_PDRP or ASL_PDRP based on TPR is used as a brain image feature;

[0026] The general linear model is used to standardize the brain image features, specifically: according to the brain image features of the normal control group in the training set and the corresponding age and gender, a general linear model (GLM) is constructed:

[0027] Y~β+β age X age +β gender X gender

[0028] Among them, Y is the brain image feature, X age is age, X gender For gender, β, β age , β genderare the parameters of the general linear model; the constructed general linear model is applied to all participants in step (1) to obtain estimated brain image features; it is applied to the brain image features extracted in step (3) to obtain brain image features that eliminate the effects of age and gender; and the brain image features after eliminating the effects are z-valued to [-1,1].

[0029] Furthermore, the method for using the PD diagnostic model is as follows: the T1-weighted image, ESWAN image and ASL image to be diagnosed are processed in steps (2) to (3) in sequence, and the influence of age and gender is eliminated based on the GLM; according to step (4), multiple brain image features with the highest contribution are screened using LASSO regression analysis, and the trained PD diagnostic model is input, and then expanded to be applied to the diagnosis of different types of neurodegenerative diseases.

[0030] Furthermore, in step (4), based on the current brain image features and the corresponding classification labels, the features with higher contribution are selected. Specifically, LASSO analysis is performed and cross-validation is performed using the R language package glmnet to select the optimal regularization parameter, and the 20 most important brain image features are screened. The selected important brain image features are subjected to binary logistic regression analysis using the R language package glm to construct a PD diagnostic model.

[0031] Furthermore, in step (5), based on the trained PD diagnostic model, the predict function of the R language package was used to calculate the individual prediction score, and further Pearson correlation analysis was performed with the UPDRS II and UPDRS III scores; SHAP analysis was used to screen the top 6 brain imaging features with the greatest contribution, and ANOVA analysis of variance was used to compare the differences between groups in different data sets, and FDR correction was performed.

[0032] Furthermore, the method also includes verifying the diagnostic ability of the trained PD diagnostic model across disease sets, specifically: using the predict function of the R language package to apply the trained PD diagnostic model to the MSA, PSP, and ET patient data sets, using the Delong test to evaluate the diagnostic efficacy, and performing FDR correction.

[0033] The beneficial effects of the present invention are:

[0034] (1) A set of 20 important brain imaging features, mainly abnormal perfusion patterns of PD and iron distribution in the substantia nigra, were obtained. The PD diagnostic model constructed based on the above imaging features had an area under the diagnostic curve (AUC value) of 0.839 (95% CI: 0.794-0.879) in the training set and an AUC value of 0.797 (95% CI: 0.682-0.886) in the independent validation set, indicating that the constructed PD diagnostic model has good stability and diagnostic ability in different PD populations and has potential clinical significance.

[0035] (2) The calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC) of the PD diagnostic model further showed that the model was relatively repeatable between the training set and the independent validation set. The predicted individual scores based on the model were significantly correlated with the severity of clinical symptoms (UPDRS II: R = 0.12, P = 0.01; UPDRS III: R = 0.13, P = 0.026), indicating that the model has the potential to serve as a clinically interpretable imaging marker for PD and has certain clinical practical value.

[0036] (3) The diagnostic efficacy of the PD diagnostic model constructed based on the above brain imaging features in different disease sets showed that the model had good diagnostic efficacy for PSP and MSA patients, but poor diagnostic efficacy in ET patients (PSP cohort vs ET cohort: P < 0.001; MSA cohort vs ET cohort: P = 0.002). The diagnostic AUC values ​​of the model in the PSP and MSA cohorts were 0.927 (95% CI: 0.844-0.986) and 0.902 (95% CI: 0.800-0.978), respectively, and the diagnostic AUC value in the ET cohort was 0.651 (95% CI: 0.543-0.755). The association between the imaging diagnostic markers constructed in the PD population and the above diseases is still unknown. This result further indicates that although the diagnostic model was constructed by PD patients, PSP and MSA patients may also benefit from it, and its model generalization ability deserves further exploration. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1The preprocessing and analysis process of QSM and ASL images. (A.) The preprocessing and analysis process of QSM images, where a is the individual space QSM image, b is the QSM template in the standard space, c is the mask of the subcortical nuclei after segmentation by the DeepQSMSeg algorithm, and d is the average magnetic susceptibility value of the subcortical nuclei extracted after manual modification. (B.) The preprocessing and analysis process of ASL images, where a is the ASL image and the CBF image calculated based on the ASL image, b is the gray matter and white matter template generated by T1-weighted image segmentation, c is the spatial normalization of the ASL image and CBF image, d is the smoothed image, e is the voxel vector matrix of the CBF image of all participants in the training set, f is PC1 as ASL_PDRP, g is the topographic profile of ASL_PDRP, and h is the individual ASL_PDRP expression in other disease sets calculated using TPR.

[0038] Figure 2 This is a flowchart for constructing a PD diagnostic model based on brain imaging features. (A.) is the inclusion of 4 clinical cohorts and the collection of clinical data. (B.) is the construction process of the PD diagnostic model. A total of 80 image features were extracted using T1-weighted images, ASL and QSM images; a general linear model was constructed based on 122 normal controls in the training cohort to control the effects of age and gender on the extracted image features; the optimal number of image features and the best image features were selected by data-driven LASSO regression; all feature matrices of subjects (80×369) and subject feature selection matrices (20×369) were obtained from the training set; and the diagnostic model was constructed using binary logistic regression analysis. (C.) is the application and validation of the diagnostic model in different disease cohorts. The diagnostic model was applied to the MSA, PSP and ET cohorts respectively; the ROC curves were compared using the Delong test, and the AUC values ​​(95% CI) were obtained by 1000 bootstrappings. (D.) Evaluation of the clinical practicality of the diagnostic model. The diagnostic efficacy of the model was verified in an independent validation set. Calibration curve, DCA and CIC analyses were performed on the training set and the independent validation set. SHAP analysis was used to determine the brain imaging features that contributed more to the diagnostic model, and then single feature difference analysis was performed. Correlation analysis was performed between individual prediction scores and clinical scale scores.

[0039] Figure 3 ROC curves of the feature selection process and diagnostic model for LASSO regression in the training set and independent validation set. (A.) Log Lambda (λ) and the error model of LASSO regression, using ten-fold cross validation, cross validation using the R language package "glmnet", select λ min=20 as the optimal value for determining the number of image features; (B.) Log(λ) and LASSO regression coefficients, and the optimal λ or penalty coefficient is determined from LASSO regression. (C.) ROC curves of the diagnostic model in the training set and independent validation set, respectively.

[0040] Figure 4 The clinical evaluation of the PD diagnostic model. (A.) is the calibration curve analysis of the diagnostic model in the training set; (B.) is the calibration curve analysis of the diagnostic model in the independent validation set. (C.) is the clinical decision curve analysis of the diagnostic model in the training set; (D.) is the clinical decision curve analysis of the diagnostic model in the independent validation set. (E.) is the clinical impact curve analysis of the diagnostic model in the training set; (F.) is the clinical impact curve analysis of the diagnostic model in the independent validation set.

[0041] Figure 5 This is an exploration of the clinical practicality of the PD diagnostic model. (A.) Pearson correlation analysis between the individual prediction scores of the included PD patients and the clinical UPDRS I score; (B.) Pearson correlation analysis between the individual prediction scores of the included PD patients and the clinical UPDRS II score; (C.) Pearson correlation analysis between the individual prediction scores of the included PD patients and the clinical UPDRS III score; (D.) Pearson correlation analysis between the individual prediction scores of the included PD patients and the total score of the Purdue Peg Board; (E.) Pearson correlation analysis between the individual prediction scores of the included PD patients and the assembly score; (F.) SHAP mean absolute value bar chart helps to screen imaging features with high contribution.

[0042] Figure 6The difference analysis of the standardized values ​​of the first 6 brain imaging features from the SHAP analysis after removing the influence of age and gender. (A.) The difference of the standardized values ​​of the first 6 brain imaging features in the training set, including: ASL_PDRP: P<0.001; R_SN: P<0.001; L_SN: P<0.001; (B.) The difference of the standardized values ​​of the first 6 brain imaging features in the independent validation set: ASL_PDRP: P=0.015; R_SN: P=0.008; L_SN: P=0.008). (C.) The differences in the standardized values ​​of the first six brain imaging features in the PSP-NC cohort, including: ASL_PDRP: P<0.001; R_SN: P<0.001; L_SN: P<0.001; R_pallium: P=0.001; R_caudate: P=0.030; rh_supermargin: P=0.030; (D.) The differences in the standardized values ​​of the first six brain imaging features in the PSP-PD cohort, including: R_SN: P<0.001; L_SN: P<0.001; R_pallium: P<0.001; R_caudate: P=0.018). (E.) is the difference in the standardized values ​​of the first 6 brain imaging features in the MSA-NC cohort, among which: ASL_PDRP: P = 0.007; L_SN: P = 0.007; (F.) is the difference in the standardized values ​​of the first 6 brain imaging features in the MSA-PD cohort. (G.) is the difference in the standardized values ​​of the first 6 brain imaging features in the ET-NC cohort, and (H.) is the difference in the standardized values ​​of the first 6 brain imaging features in the ET-PD cohort, among which: ASL_PDRP: P = 0.017). The above comparisons were all corrected for FDR. *P < 0.05, **P < 0.01, ***P < 0.001. DETAILED DESCRIPTION

[0043] The present invention aims to extract multi-level and multi-dimensional brain imaging information from multi-sequence brain magnetic resonance brain features, and construct a clinically interpretable PD diagnostic model with the help of machine learning. Multimodal MRI images include QSM, ASL and T1-weighted images, from which brain imaging features that can fully reflect the characteristics of PD disease can be extracted. The research framework defined in the present invention includes brain imaging data preprocessing, feature standardization and screening, model construction, independent validation set and cross-disease set validation. In addition, the performance of the model of the present invention in the cross-disease validation set shows that the model has good diagnostic efficacy for PSP and MSA, but limited diagnostic efficacy in ET.

[0044] The research implemented in the present invention was approved by the Medical Ethics Committee of the Second Affiliated Hospital of Zhejiang University School of Medicine, and all PD patients, PSP patients, MSA patients and NC subjects signed informed consent.

[0045] The clinical and imaging data of this example were collected from August 2014 to December 2023. The diagnosis of PD was made by a senior neurologist according to the British PD Association Brain Bank Standards and the Movement Disorder Association Clinical Diagnosis Criteria. Initially, a total of 356 PD patients and 183 NC patients were recruited, all of whom underwent multimodal MRI scans and clinical evaluations, of which 77 subjects were excluded due to head movement, significant brain atrophy, multiple microbleeds, history of head trauma surgery, cerebrovascular disease, and other neurological / psychiatric diseases. The final cohort included 307 PD patients and 155 NC patients. For drug-treated PD patients, MRI scans and clinical evaluations were performed during the anti-Parkinson drug washout period (at least 12 hours of drug withdrawal). The diagnosis of MSA and PSP was based on consensus diagnostic criteria, and the diagnosis of ET was based on the Movement Disorder Association statement; the same exclusion criteria were applied to the cross-disease data set, and finally 18 MSA, 15 PSP and 80 ET subjects were included.

[0046] (1) Acquisition of MRI sequences. All subjects were scanned using a 3.0T MRI scanner (GE Discovery 750) equipped with an 8-channel head coil. During the MRI scan, the head was fixed with a foam pad and earplugs were provided to reduce noise; the subjects remained awake with their eyes closed. If the image quality was affected by motion artifacts or other factors, the affected sequence was rescanned to ensure the best quality images for preprocessing. The scanning parameters of each MRI sequence are recorded as follows:

[0047] (1.1) High-resolution 3D T1-weighted images were acquired using a fast spoiler gradient sequence: repetition time = 7.336 ms; echo time = 3.036 ms; inversion time = 450 ms; flip angle = 11°; field of view = 260 × 260 mm 2 ; Matrix = 256×256; Layer thickness = 1.2 mm; Number of layers = 196.

[0048] (1.2) ESWAN images were acquired using a gradient echo sequence: repetition time = 33.7 ms; first echo time / interval / eighth echo time = 4.556 ms / 3.648 ms / 30.092 ms; flip angle = 20°; field of view = 240 × 240 mm 2 ; Matrix = 416×384; Layer thickness = 2 mm; Layer spacing = 0 mm; Number of layers = 64.

[0049] (1.3) ASL images were obtained using pulsed continuous arterial spin labeling technology: repetition time = 4632 ms; echo time = 10.536 ms; flip angle = 111°; field of view = 240×240 mm 2 ; Matrix = 128 × 128; Layer thickness = 4 mm; Layer spacing = 0 mm; Number of layers = 36.

[0050] (2) Preprocess the obtained T1-weighted images, ESWAN images and ASL images.

[0051] (2.1) The present invention uses the STAR-QSM algorithm in the STI Suite V3.0 software package to reconstruct the ESWAN data obtained in step (1) to obtain a QSM map (using the STISuite V3.0 software package developed by the University of California, Berkeley,

[0052] https: / / people.eecs.berkeley.edu / ~chunlei.liu / software.html). In order to obtain the original magnetic susceptibility of each nucleus in the individual space, the DeepQSMSeg algorithm based on deep learning was used to segment the 10 subcortical nuclei in the individual space QSM map. The steps are as follows: Figure 1 A, specifically:

[0053] (2.1.1) The steps of reconstructing QSM images mainly include: ① unwrapping the original phase map using Laplace-based phase unwrapping and retaining the spatial low-frequency components of all brain tissues; ② removing the background phase using spherical mean filtering to specifically obtain brain tissue phase data; ③ estimating the magnetic susceptibility boundary and removing magnetization artifacts; ④ obtaining the tissue susceptibility map to calculate the QSM image.

[0054] (2.1.2) An end-to-end deep learning-based tool (DeepQSMSeg algorithm) was used to segment subcortical nuclei. The DeepQSMSeg algorithm was developed by setting the true value to the historical data of previous manual and semi-automatic segmentation. It is mainly used to segment subcortical nuclei such as the substantia nigra and red nucleus. After automated segmentation, all nucleus images obtained by segmentation were manually corrected by an experienced neuroradiologist to correct the misalignment of surrounding tissue voxels caused by registration deviation.

[0055] (2.1.3) After the nuclear segmentation was completed, the average magnetic susceptibility of the 10 subcortical nuclei, including the bilateral substantia nigra, red nucleus, putamen, caudate nucleus and globus pallidus, was extracted.

[0056] (2.2) The cortical gray matter of the original individual T1-weighted image obtained in step (1) was segmented using the FreeSurfer open source software. The “recon-al” preprocessing process was used to segment the cortex of the individual spatial T1-weighted image, including automatic head motion correction, skull stripping, spatial normalization, spatial registration, and cortical segmentation. Based on the Desikan-Killiany atlas, 68 cortical regions (34 cortical regions in each hemisphere) and TIV ( Figure 2 B).

[0057] (2.3) The ASL image obtained in step (1) is preprocessed using SPM 12, and ASL_PDRP is constructed as follows: Figure 1 B, specifically:

[0058] (2.3.1) The ASL images and the CBF images generated based on ASL were preprocessed by SPM 12. The Segment standard segmentation algorithm of SPM 12 performed tissue segmentation on the T1-weighted images of each subject, co-registered them to the MNI standard space, and generated the DARTEL gray matter and white matter templates based on the study population.

[0059] (2.3.2) The DARTEL toolkit was used to perform spatial registration and normalization of the ASL and CBF images in the MNI standard space.

[0060] (2.3.3) The normalized CBF map was smoothed using a Gaussian kernel with a FWHM of 6 mm.

[0061] (2.3.4) PCA-SSM method was used to construct ASL_PDRP for CBF images: To construct individual sub-profiles, voxel-level analysis of the CBF images of all participants in the training set was required, and the images were expressed as high-dimensional vectors to construct voxel matrices. The PCA method was used to perform matrix centering and covariance matrix decomposition on the voxel matrix, specifically: log transformation and double decentering were performed, and the obtained participant residual profile (SRP) image represented the deviation between the participant and the group mean. Finally, the covariance matrix was decomposed to output the eigenvectors (i.e., spatial patterns, representing the unique spatial distribution network of functionally related brain regions expressed to different degrees by each subject) and scores (subject scores, representing the degree of expression of the eigenvectors in the image data of a specific individual) corresponding to each principal component (PC). The linear combination of the top scores or individual PCs constituted a proportional sub-profile model (i.e., composite pattern), and the expression value (ASL_PDRP) of each subject on the proportional sub-profile model was calculated.

[0062] (2.3.5) The individual standardized CBF map based on the independent validation set and the cross-disease dataset obtained by the TPR algorithm was multiplied and accumulated with the weight of the voxel in the proportional sub-profile model (i.e., the inner vector product of the participant's SRP vector and the final pattern vector) to obtain the overall score of the individual on the proportional sub-profile model as the ASL_PDRP obtained based on TPR.

[0063] Among them, the individual standardized CBF map is a standardized image obtained by applying the above-mentioned pattern mask in an independent validation set and a cross-disease dataset using the TPR algorithm, and also performing log transformation and double decentered averaging.

[0064] (3) Brain image feature screening and extraction.

[0065] (3.1) For the QSM image, the present invention calculated the average magnetic susceptibility values ​​of 10 subcortical nuclei as 10 brain image features, and initially formed a 369×10 feature matrix.

[0066] (3.2) For T1-weighted image segmentation, 68 cortical region thicknesses and 1 TIV feature were obtained in the bilateral cerebral hemispheres, which initially formed a 369×69 feature matrix.

[0067] (3.3) For ASL images, ASL_PDRP or ASL_PDRP based on TPR is used alone as a brain image feature to initially form a 369×1 feature matrix.

[0068] (4) Brain image feature standardization: general linear model ( Figure 2 B).

[0069] Since age and gender are common confounding factors, the present invention uses the data of 122 NCs in the training set to construct a GLM to estimate the effects of age and gender of the normal population on each brain imaging feature:

[0070] Y~β+β age X age +β gender X gender

[0071] Among them, Y is the brain image feature, X age is age (continuous variable), X gender is gender (categorical variable), β, β age , β gender are the parameters of the GLM.

[0072] For each brain feature Y i (122×1)(i=1~80), we can calculate a set of β, β age , β gender , and the corresponding general linear model (80) was obtained. The constructed GLM was applied to the brain imaging features of all participants (including PD, PSP, MSA, and ET), and the standard brain imaging feature matrices (80×369; 80×93; 80×51; 80×48; 80×113) were obtained after eliminating the effects of age and gender. The brain imaging features after eliminating the effects were z-valued to [-1,1] using SPSS25.0.

[0073] (5) Feature selection and model construction: LASSO analysis was performed and ten-fold cross validation was performed using the R language package “glmnet” to select the optimal regularization parameter, perform brain image feature selection, and build a PD diagnostic model ( Figure 2 B), as follows:

[0074] (5.1) Using the standardized brain image features preprocessed in step (4), LASSO analysis was performed and 10-fold cross validation was used to analyze the difference between the model prediction results and the real data, and then the 80 brain image features processed in step (4) were sorted and screened by deviation (Deviance) Figure 3 A); and select the 20 most important (i.e., the smallest deviation) brain image features as the training set features input into the subsequent classification model ( Figure 3 B).

[0075] (5.2) Logistic regression analysis was used to establish a diagnostic model. Figure 2 D. The present invention divides all PD and NC involved in this project into a training set (for training the classification model) and an independent validation set (for internal validation) according to the inclusion time point (around November 2021), with a split ratio of 4:1. In this embodiment: 247 PD patients and 122 NC included before November 2021 are used as the training set; 60 PD patients and 33 NC included after November 2021 are used as the independent validation set. The training set is used for 1000 iterations of sampling to train the PD diagnostic model, so that the model can be stably applied to other independent disease data sets ( Figure 2 B). The input of the model is the 20 brain image features selected in step (5.1) of the corresponding dataset.

[0076] (6) Clinical efficacy analysis and cross-disease dataset validation

[0077] (6.1) In order to enhance the clinical interpretability of the PD diagnostic model, we performed calibration curve, DCA, and CIC analysis in the training set and independent validation set, respectively; and used Pearson correlation analysis to explore the relationship between the predicted individual scores obtained based on the model and the severity of clinical symptoms (UPDRS I-III) ( Figure 2 D).

[0078] (6.2) In order to further verify the generalization ability of the model obtained in the present invention in independent validation sets and cross-disease datasets. The present invention first used 60 PD patients and 33 normal controls recruited from November 2021 to December 2023 to independently validate the constructed diagnostic model. In addition, the present invention also included 18 MSA, 15 PSP and 80 ET patients to perform cross-disease validation on the constructed PD diagnostic model ( Figure 2C) All procedures were performed independently except for using the previously estimated GLM to eliminate the effects of age and sex.

[0079] (7) To further explain the model, the present invention conducted a single feature analysis and used SHAP analysis to screen the top six brain imaging features with the greatest contribution ( Figure 2 D), and ANOVA analysis of variance was used to compare the differences between groups in standardized brain imaging features in different data sets such as training set, independent validation set, MSA cohort, PSP cohort and ET cohort, and FDR correction was performed ( Figure 6 ).

[0080] The method for using the PD diagnostic model of the present invention is: the T1-weighted image, ESWAN image and ASL image to be diagnosed are processed in steps (2) to (3) in sequence, and then the influence of age and gender is eliminated based on the GLM according to step (4), and then according to step (5), LASSO regression analysis is used to screen brain image features with higher contribution, and the features are input into the trained PD diagnostic model, and the model is expanded to be applied to the diagnosis of different types of neurodegenerative diseases.

[0081] In summary, the present invention proposes a method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features, which provides an objective basis for the clinical diagnosis of PD, helps improve its diagnostic accuracy, and has good clinical transformation value. The present invention establishes a model construction process integrating training-independent verification-cross-disease verification, and calculates the model diagnostic performance index AUC value and its confidence interval through 1000 iterations. The model construction process includes brain imaging data preprocessing, feature standardization and screening, model construction, independent verification set and cross-disease set verification. The research characteristics and innovation of the present invention are that a set of brain function and metabolic characteristics based on ASL_PDRP and substantia nigra iron distribution are obtained, which not only reaffirms the diagnostic effect of ASL_PDRP on PD in previous studies, but also clarifies the important position of substantia nigra iron metabolism abnormalities in PD disease, and verifies the effectiveness of ASL_PDRP combined with brain iron metabolism and brain structure-related features as imaging biomarkers for PD. Secondly, the clinical interpretability of the model is promoted by analysis such as calibration curves, DCA, CIC, and correlation analysis between predicted individual scores and clinical symptoms obtained based on the model. Finally, the diagnostic efficacy of the PD diagnostic model constructed based on the above-mentioned brain imaging features in different disease sets showed that the diagnostic efficacy of the model was different in different diseases (PSP, MSA, ET), with good diagnostic efficacy for PSP and MSA, but limited diagnostic efficacy in ET, indicating that the model has potential cross-disease generalization capabilities.

Claims

1. A method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features, characterized in that: The following steps are involved: (1) T1-weighted images, ESWAN images, and ASL images were obtained from PD patients, normal controls, PSP patients, MSA patients, and ET patients. The PD patients and normal controls were divided into training sets and independent validation sets in proportion, and the PSP patients, MSA patients, and ET patients were used as cross-disease datasets. (2) Segment the cortical gray matter of the T1-weighted image obtained in step (1) to obtain 68 cortical regions and 1 TIV; post-process the ESWAN image obtained in step (1) using the STAR-QSM algorithm to obtain a QSM map, and segment the QSM map using the DeepQSMSeg algorithm to obtain 10 subcortical nuclei including the bilateral substantia nigra, red nucleus, putamen, caudate nucleus, and globus pallidus; perform principal component analysis-proportional subsection model analysis on the ASL image obtained in step (1) to obtain the abnormal perfusion pattern of PD; (3) extracting brain image features for each cortical region, subcortical nucleus, and PD abnormal perfusion pattern segmented in step (2); The brain imaging features were standardized using a general linear model; (4) The LASSO method was used to select multiple brain imaging features with the highest contribution as input, and binary logistic regression was used to construct a PD diagnosis model based on the training set; (5) According to step (4), the individual prediction scores obtained based on the PD diagnostic model were correlated with the corresponding clinical symptoms. SHAP analysis was used to screen the top 6 brain imaging features in terms of contribution, and inter-group differences were compared in the training set, independent validation set, PSP, MSA, and ET patient data sets, and FDR correction was performed.

2. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 1, characterized in that: In step (1), a fast spoiler gradient sequence is used to obtain a T1-weighted image; a gradient echo sequence is used to obtain an ESWAN image; and a pulsed continuous arterial spin labeling technique is used to obtain an ASL image.

3. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 1, characterized in that: In step (2), the cortical gray matter of the T1-weighted image obtained in step (1) is segmented to obtain 68 cortical regions, specifically: the subcortical nuclei of the T1-weighted image are segmented using FreeSurfer software, and the recon-al preprocessing process is used to perform automated head motion correction, skull stripping, spatial normalization, spatial registration, and cortical segmentation, and then based on the Desikan-Killiany atlas, 68 cortical regions and TIV are segmented; The ESWAN image obtained in step (1) is post-processed using the STAR-QSM algorithm to obtain a quantitative magnetic susceptibility map, specifically: the original phase map is unwrapped using Laplace-based phase unwrapping, and the spatial low-frequency components of all brain tissues are retained; the background phase is removed using spherical mean filtering, thereby specifically obtaining brain tissue phase data; the magnetic susceptibility boundary is estimated to remove magnetization artifacts; Obtain tissue magnetic susceptibility maps to calculate QSM images; The DeepQSMSeg algorithm tool sets the true value to the historical data of manual and semi-automatic segmentation; After the DeepQSMSeg algorithm is used to segment the QSM image to obtain the subcortical nuclei, it also includes: manual correction to correct the misalignment of the surrounding tissue voxels caused by the registration deviation.

4. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 1, characterized in that: In step (2), the ASL image obtained in step (1) is subjected to principal component analysis-proportional subsection model analysis to obtain the abnormal PD perfusion pattern, specifically: (2.2.1) The ASL image and the CBF image generated based on ASL were preprocessed by SPM 12, including: head motion correction, spatial registration and normalization of the ASL image and the CBF image using DARTEL, CBF image standardization, and spatial smoothing; the spatial smoothing used a Gaussian kernel with a half-width of 6 mm; (2.2.2) ASL_PDRP was constructed on the CBF images preprocessed in step (2.2.1) using the principal component analysis-proportional sub-profile model analysis method, including: constructing individual sub-profiles, principal component analysis, proportional sub-profile model analysis, and constructing PD-related cerebral perfusion patterns; (2.2.3) Topological profile scores were used to calculate individual ASL_PDRP expression scores in an independent validation set and in PSP, MSA, and ET patient datasets.

5. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 4, characterized in that: In step (2.2.2), based on step (2.2.1), the individual sub-profile is constructed to perform voxel-level analysis on the CBF maps of all participants in the training set and express them as high-dimensional vectors; the principal component analysis performs matrix centering and covariance matrix decomposition, and outputs the eigenvector and score corresponding to each principal component; the principal component with the highest score is selected to form a proportional sub-profile model, and the expression value ASL_PDRP of each participant on the proportional sub-profile model is calculated; In step (2.2.3), based on the proportional sub-profile model obtained based on the training set in step (2.2.2), the individual standardized CBF map based on the independent validation set and the cross-disease dataset obtained by the TPR algorithm is multiplied and accumulated with the weights of the corresponding voxels in the proportional sub-profile model to obtain the individual's overall score on the proportional sub-profile model as the ASL_PDRP obtained based on TPR.

6. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 5, characterized in that: In step (3), the brain image features extracted for each cortical region, subcortical nucleus and PD abnormal perfusion pattern segmented in step (2) are specifically: (3.1) For the QSM image, the average magnetic susceptibility values ​​of the 10 subcortical nuclei were calculated as 10 brain imaging features; (3.2) For T1-weighted images, 68 average values ​​of cortical thickness and 1 TIV were calculated based on the Desikan-Killiany atlas as 69 brain imaging features; (3.3) For ASL images, ASL_PDRP or ASL_PDRP based on TPR is used as a brain image feature; The general linear model is used to standardize the brain image features, specifically: according to the brain image features and corresponding age and gender of the normal control group in the training set, a general linear model is constructed: Y~β+β age X age +b gender X gender Among them, Y is the brain image feature, X age is age, X gender For gender, β, β age , β gender are the parameters of the general linear model; the constructed general linear model is applied to all participants in step (1) to obtain the estimated brain imaging features; The brain image features extracted in step (3) are applied to obtain brain image features that eliminate the effects of age and gender; the brain image features after eliminating the effects are then z-valued to [-1, 1].

7. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 1, characterized in that: The method for using the PD diagnostic model is as follows: the T1-weighted image, ESWAN image and ASL image to be diagnosed are processed in steps (2) to (3) in sequence, and the influence of age and gender is eliminated based on the GLM; according to step (4), multiple brain image features with the highest contribution are screened using LASSO regression analysis, and the trained PD diagnostic model is input, and then expanded to be applied to the diagnosis of different types of neurodegenerative diseases.

8. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 1, characterized in that: In step (4), based on the current brain image features and the corresponding classification labels, the features with higher contribution are selected. Specifically, LASSO analysis is performed and cross-validation is performed using the R language package glmnet to select the optimal regularization parameter, and the 20 most important brain image features are screened. The selected important brain image features are subjected to binary logistic regression analysis using the R language package glm to construct a PD diagnostic model.

9. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 1, characterized in that: In step (5), based on the trained PD diagnostic model, the predict function of the R language package was used to calculate the individual prediction score, and further Pearson correlation analysis was performed with the UPDRS II and UPDRS III scores; SHAP analysis was used to screen the top 6 brain imaging features with the greatest contribution, and ANOVA analysis of variance was used to compare the differences between groups in different data sets, and FDR correction was performed.

10. The method for establishing a clinically interpretable PD diagnostic model based on multi-sequence magnetic resonance brain features as claimed in claim 1, characterized in that: The method also includes verifying the diagnostic ability of the trained PD diagnostic model across disease sets, specifically: using the predict function of the R language package to apply the trained PD diagnostic model to the MSA, PSP, and ET patient data sets, using the Delong test to evaluate the diagnostic efficacy, and performing FDR correction.

Images