Imaging student biomarker for combined diagnosis of Parkinson's disease and diagnosis model

Abstract

The invention relates to the field of medical diagnosis, in particular to an imaging biomarker for combined diagnosis of Parkinson's disease and a diagnosis model. According to the invention, through two-sample MR, the causal association between the 3370 brain image index and the Parkinson's disease as well as clinical manifestation and severity of the Parkinson's disease and the direction of the causal association are systematically studied. Then, positive results in MR results are tested in an Asian Parkinson's disease queue, a diagnosis model is constructed, and verification is carried out in a verification set. The result of the invention supports the causal relationship between brain iconography change and Parkinson's disease, and reveals the pathology of Parkinson's disease and a new perspective of early intervention of Parkinson's disease.

Description

Technical Field

[0001] The present invention relates to the field of medical diagnosis, and in particular to imaging biomarkers and diagnostic models for combined diagnosis of Parkinson's disease. Background Art

[0002] Parkinson's disease (PD) is the second most common neurodegenerative disease in the world, bringing heavy health and economic burdens. The pathogenesis of PD is unclear, and there is still no effective causal treatment. In addition, due to its insidious onset and long clinical latency, the best window for neuroprotection is often missed by the time motor symptoms appear. Therefore, it is particularly important to identify potential biomarkers for early diagnosis and to perform early diagnosis and screening of PD.

[0003] Brain imaging is an effective non-invasive tool that can effectively detect changes in brain structure and function. Observational studies suggest that there is a strong correlation between brain imaging phenotypes (IDPs) and PD. However, observational studies have some unavoidable defects. First, observational studies cannot completely eliminate the effects of confounding factors. This will cause the research results to be affected by some factors that can affect both brain structure and function and the clinical manifestations of PD, such as medication. In addition, observational studies will introduce reverse causality, and it is impossible to distinguish whether brain imaging changes cause PD or whether the PD disease state causes brain imaging changes. Studying the causal relationship between brain imaging changes and PD is particularly important for early diagnosis and screening of PD.

[0004] Mendelian randomization (MR), similar to natural randomized controlled studies, involves random assignment during the generation of fertilized eggs. Therefore, potential causal relationships between exposure and outcomes can be explored. By using single nucleotide polymorphisms (SNPs) as instrumental variables, MR can control confounding factors and eliminate reverse causality. In 2023, Zhu et al. studied the relationship between whole-brain Cecilia lesions and PD through MR, but their study only focused on three white matter phenotypes: white matter hyperintensities, various heterogeneity scores, and mean dispersion. There is currently a lack of systematic studies on the causal relationship between multimodal brain imaging and PD, the degree of PD research, and the clinical manifestations of PD. Therefore, it is necessary to comprehensively explore the causal relationship between multimodal brain imaging and PD, including various motor and non-motor symptoms and disease severity. Summary of the invention

[0005] The purpose of the present invention is to provide imaging biomarkers and diagnostic models for the combined diagnosis of Parkinson's disease. The present invention first systematically studied the causal relationship and direction between 3370 IDPs (n=39,691) and PD (n=482,730) and PD manifestations and severity (n=4,093) through two-sample MR. Then, the present invention tested the positive results in the MR results in an Asian PD cohort, constructed a diagnostic model, and verified it in the validation set.

[0006] In a first aspect of the present invention, an imaging biomarker for the combined diagnosis of Parkinson's disease is provided, wherein the imaging biomarker for the combined diagnosis of Parkinson's disease consists of the median T2 star value of the left putamen, the median T2 star value of the right caudate nucleus, the left thalamus volume, the right VPL volume, the left Pt volume, the left uncus FA value, and the left temporal pole plane thickness.

[0007] A second aspect of the present invention provides a detection product for imaging biomarkers for joint diagnosis of Parkinson's disease, including a product for detecting imaging biomarker data for joint diagnosis of Parkinson's disease as described above through SWI, T1, SWI multimodal magnetic resonance imaging (MRI).

[0008] The third aspect of the present invention provides an application of the imaging biomarker for combined diagnosis of Parkinson's disease as described above, and a detection product of the imaging biomarker for combined diagnosis of Parkinson's disease in establishing a Parkinson's disease diagnosis system or a diagnosis product.

[0009] A fourth aspect of the present invention provides a Parkinson's disease diagnosis model, wherein the calculation formula of the Parkinson's disease diagnosis model is as follows:

[0010] Prediction score = -0.047518*median T2 star value of left putamen + 0.016002*median T2 star value of right caudate + 0.000141*left thalamus volume + 0.00196*right VPL volume + 0.22711*left PT volume -14.868015*left uncus FA value -3.926307*left temporal pole plane thickness + 15.273922.

[0011] When the predicted score is greater than -1.347, the possibility of PD is high; when the predicted score is less than -1.347, the possibility of PD is low.

[0012] According to a fifth aspect of the present invention, a diagnostic system for Parkinson's disease is provided, the diagnostic system comprising an analysis and judgment module; the analysis and judgment module contains the diagnostic model as described above, and is used to output a diagnosis result of whether the sample suffers from Parkinson's disease.

[0013] Furthermore, the diagnostic system also includes a detection module, which detects imaging biomarker data for the joint diagnosis of Parkinson's disease in the sample to be tested and transmits the data to the analysis and judgment module, wherein the imaging biomarker for the joint diagnosis of Parkinson's disease consists of the median T2 star value of the left putamen, the median T2 star value of the right caudate nucleus, the left thalamus volume, the right VPL volume, the left Pt volume, the left uncus FA value, and the left temporal pole plane thickness.

[0014] A sixth aspect of the present invention provides a use of the above-mentioned diagnostic system in the preparation of a diagnostic product for Parkinson's disease.

[0015] According to a seventh aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, wherein the computer program is a computer program for implementing the functions of the Parkinson's disease diagnostic model as described above or the diagnostic system for Parkinson's disease as described above.

[0016] In an eighth aspect of the present invention, a product for diagnosing Parkinson's disease is provided, wherein the product comprises:

[0017] A detection product for the numerical value of imaging biomarkers for the combined diagnosis of Parkinson's disease; the imaging biomarkers for the combined diagnosis of Parkinson's disease are composed of the median T2 star value of the left putamen, the median T2 star value of the right caudate nucleus, the left thalamus volume, the right VPL volume, the left Pt volume, the left uncus FA value, and the left temporal pole plane thickness;

[0018] and a diagnostic system for Parkinson's disease as described above, or a computer-readable storage medium as described above.

[0019] The advantages of the present invention are:

[0020] 1. The present invention first systematically studied the causal relationship and direction between 3370 IDPs (n=39,691) and PD (n=482,730) and PD manifestation and severity (n=4,093) through two-sample MR. Then, the present invention tested the positive results in the MR results in the Asian PD cohort, constructed a diagnostic model, and verified it in the validation set.

[0021] 2. This invention is the first to evaluate the causal relationship between multimodal imaging (including brain structure, connectivity and function) and PD and multiple motor and non-motor manifestations and their progression. It has advantages over traditional observational studies. The use of MR minimizes the possibility of confounding factors and reverse causality because genetic variants are randomly determined at conception and are therefore not modified by environmental factors or diseases. These advantages of MR research allow us to explore the underlying mechanisms underlying the phenotype and identify the causal relationship between MRI changes and PD.

[0022] 3. The results of the present invention support the causal relationship between brain imaging changes and PD, revealing a new perspective on PD pathology and early intervention of PD.

[0023] 4. The results of the present invention provide a model for assisting PD diagnosis through simple and non-invasive magnetic resonance imaging. This technology has important clinical application prospects and is of great significance for the diagnosis of early PD. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 :The causal inference workflow between IDPs and PD. The MR study and observational study used to explore the causal relationship between IDPs and PD are shown. The process includes selecting specific SNPs from GWAS datasets, performing MR analysis, sensitivity analysis, and observational cohort studies.

[0025] Figure 2 :The causal relationship between IDPs and PD and PD manifestations was investigated using the inverse variance weighted method. Forest plots show the significant causal effects of IDPs on PD (a), PD severity (b), and PD manifestations (c), with pfdr < 0.05. (d) Forest plot shows the significant causal effect of PD on IDPs.

[0026] Figure 3: Observational associations between brain IDPs and PD. (a, e) Schematic diagrams and bar charts showing the median T2 weeks of the caudate nucleus, putamen, and globus pallidus in HCs, iRBD, and PD patients. (b, f) Schematic diagrams and bar charts showing the volume results of the left thalamus and left Pt and right VPT (purple). (c, g) Schematic diagrams and bar charts showing the FA results in the left uncinate fasciculus. (d, h) Schematic diagrams and bar charts showing the thickness of the left temporal pole plate. The present invention uses a fixed-effect model, in which linear or non-linear age and gender effects are retained. According to the data distribution and homoscedasticity, ANOVA or Kruskal-Wallis tests are used, with logarithmic transformation or reciprocal transformation if necessary. For multiple comparisons, Dunnett Contrast test is performed after ANOVA, and FDR is performed after Kruskal-Wallis test. The discriminatory ability of the values is evaluated by receiver operating characteristic curve (ROC) analysis. The mean and SEM are set according to the reference precision and dispersion, and the error bars are set to SEM. Significance codes: 'ns', not significant; '^': ANOVA or Kruskal-Wallis test, 0.05 < p < 0.1; '#': ANOVA or Kruskal-Wallis test, 0.01 < p < 0.05; '##': ANOVA or Kruskal-Wallis test, 0.001 < p < 0.01; '': ANOVA or Kruskal-Wallis test, p < 0.001; '.': Dunnett Contrast test, FDR corrected, or t-test, 0.05 < p < 0.1; '..': Dunnett Contrast test, FDR corrected, or t-test, 0.01 < p < 0.05; '...': Dunnett Contrast test, FDR corrected, or t-test, 0.001 < p < 0.01; '....': Dunnett Contrast test, FDR corrected, or t-test, p < 0.001. **Abbreviations**: AUC: area under the ROC curve.

[0027] Figure 4 : Analyses of diagnostic specificity and sensitivity were performed in the discovery set and validation set using ROC curve analysis. Detailed implementation manners

[0028] The following detailed description of the specific implementation manners provided by the present invention is given in conjunction with the embodiments.

[0029] Example 1:

[0030] 1. Experimental method

[0031] First, the present invention prepared GWAS summary data of brain imaging indicators and PD, and studied the bidirectional causal relationship between brain imaging and PD through two-sample bidirectional MR analysis. The present invention verified these positive indicators on Asian populations, constructed a diagnostic model, and then verified it in the validation set.

[0032] (1) GWAS datasets for brain imaging

[0033] The summary data set of brain imaging comes from UK Biobank. Stephen M. Smith conducted a GWAS study on 39,691 brain imaging samples and found 17,103,079 SNPs. The present invention screened the following brain imaging features for inclusion in the study of the present invention: 1) T1 source: gray matter volume, white matter volume, cerebrospinal fluid volume, whole brain volume, hypothalamus and its subregion volume, caudate nucleus, putamen, globus pallidus, hippocampus, amygdala, nucleus accumbens, cortical area, cortical thickness. 2) SWI source: FA, MO, MD, L1, L2, L3, ICVF, OD, ISOVF calculated using two methods of TBSS nuclear probability tracking. 3) Task state and resting state MRI. Resting state MRI includes ICA25 nodes and edges, as well as ICA100 nodes and edges. Finally, 3,370 influencing features were included in the study.

[0034] (2) GWAS dataset of PD

[0035] The summary dataset of PD comes from the data released by IPDGC in 2019, including 33,674 patients and 449,056 healthy controls.

[0036] (3) Instrumental variable selection and SNP synergy

[0037] In MR studies, the present invention carefully selects instrumental variables that meet the following three core assumptions: 1) are strongly associated with exposure, 2) are independent of possible confounding factors, and 3) can only affect the outcome through exposure.

[0038] (4) Bidirectional two-sample MR analysis

[0039] MR analysis consists of two parts: forward analysis refers to the causal relationship between brain imaging and PD, and reverse MR focuses on whether the results of forward MR have reverse causal relationships. The present invention uses inverse variance weighting (IVW) as the primary analysis, and uses MR-Egger, median weighting and mode weighting for supplementary evaluation. When only one SNP is selected, Wald ratio is used for analysis.

[0040] (5) Sensitivity analysis

[0041] In order to detect the robustness of MR analysis, the present invention performs a variety of sensitivity analyses, including Cochran's Q, leave-one-out method, MR Egger intercept analysis, and MR-PRESSO analysis. Afterwards, the present invention eliminates possible reverse effects by Steiger test. SNPs that fail the sensitivity analysis will first be excluded, and then the MR analysis will be re-performed until they pass the sensitivity analysis.

[0042] (6) Ethical review and informed consent

[0043] The MR study was a secondary study, and ethical review and informed consent had been conducted in the original primary study. In the observational study, the research ethics were reviewed by the Ethics Review Committee of Ruijin Hospital Affiliated to Shanghai Jiao Tong University School of Medicine. All participants were fully informed of the purpose, process, potential benefits and risks of the study. Each subject voluntarily signed an informed consent form before enrollment.

[0044] (7) Participants included in observational studies

[0045] The participants of this study were enrolled in the healthy control group by the Department of Neurology of Ruijin Hospital Affiliated to Shanghai Jiaotong University School of Medicine, and subjects without neurological diseases and major systemic diseases voluntarily signed up to participate. PD patients were diagnosed by at least two movement disorder doctors according to the current diagnostic criteria of MDS. All subjects met the following inclusion and exclusion criteria: 1) aged between 45 and 80 years, 2) no history of severe dementia or mental illness, 3) no history of intracranial surgery and craniocerebral trauma, 4) no family history of PD and other neurological diseases, 5) no systemic inflammatory diseases and serious underlying diseases. All subjects signed informed consent, and the study was reviewed by the Ethics Committee of Ruijin Hospital.

[0046] (8) Brain imaging data processing

[0047] All imaging data were processed according to the UKB pipeline (https: / / git.fmrib.ox.ac.uk / falmagro / UK_biobank_pipeline_v_1).

[0048] (9) Statistics

[0049] The main MR analysis used IVW, MR-EGGER, median weighted and mode weighted methods to evaluate the causal association between brain imaging and PD. A positive result needs to meet the following requirements: 1) IVW method pfdr < 0.05, and 2) MR-EGGER, median weighted and mode weighted methods all have the same direction as the IVW method results. Afterwards, the present invention used Cochran's Q test, leave-one-out method, MR-Egger intercept analysis, MR-PRESSO and other methods for sensitivity analysis, and Steiger test to exclude potential reverse causality. The mean and 95% execution interval were used to measure the central tendency and dispersion in the MR analysis. The present invention uses OR for categorical variable outcomes, β for continuous variable outcomes, and p-values ​​to evaluate the strength of evidence. For all results obtained in the MR analysis, the present invention uses the false positive rate FDR method for multiple comparison correction.

[0050] For observational studies, a logistic regression fixed-effect model was used to establish a model of brain imaging feature values ​​between different groups, and age and gender were put into the equation using linear or nonlinear effects according to the situation. Afterwards, the present invention used ANOVA or Kruskal-Wallis test to compare whether there were differences among the three groups based on data distribution and variance homogeneity, and logarithmic and reciprocal transformations were performed if necessary. For multiple comparison correction, Dunnett test was used after ANOVA, and FDR correction was performed after Kruskal-Wallis. The comparison between the two groups used a two-sample t test to compare whether there were differences between the two groups. The present invention evaluated the model discrimination efficiency by receiver operating characteristic curve ROC in the discovery set and validation set. In these tests, the present invention used QQ graphs and Shapiro-Wilk normality tests to test data normality, Levene's test to test variance homogeneity, and residual linearity to detect residuals by residual vs fitted graphs. The present invention uses R2 to detect the degree of fit of the model, and verifies the model through residual vs Fitted graph, normal QQ-residual graph, Scale-Location graph, and residual vs Leverage graph. The α value of all tests is set to 0.05, and all tests are two-tailed. For continuous variables, the present invention uses the mean and standard error of the mean SEM or standard deviation SD to describe the central tendency and dispersion, and the p value is used to evaluate the strength of evidence. For categorical variables, the present invention compares component differences by chi-square test, the p value evaluates the strength of evidence, and is described by frequency (%).

[0051] 2. Experimental results

[0052] (1) Positive MR: Causal effect of IDPs on PD

[0053] like Figure 1As shown, GWAS of IDPs were extracted from the UK Biobank imaging cohort (n=39,691), GWAS of PD were extracted from the International Parkinson's Disease Genomics Consortium (n=482,730), and GWAS of PD manifestations were obtained from 25 GWAS meta-analyses (n=4,093). Overall, the present invention identified 13 statistically significant causal effects of 13 different IDPs on PD ( Figure 2 ).

[0054] like Figure 2 As shown in a, increased T2 values ​​in the right striatum are associated with an increased risk of PD [OR 1.23, 95% CI 1.08-1.40, FDR-corrected p value (pfdr) 0.0057]. Increased T2 values ​​were also observed in the bilateral lentiform nuclei (left: OR 1.25, 95% CI 1.09-1.43, pfdr 0.0056; right: OR 1.25, 95% CI 1.10-1.43, pfdr 0.0056). T1-weighted imaging reveals structural changes in PD patients. Anterior MR evaluation showed that an increase in the volume of the left thalamus (OR 1.50, 95% CI 1.14-1.96, pfdr 0.016) may significantly contribute to the onset of PD. Since the thalamic subregions are functionally independent, the present invention further explores whether certain thalamic subregions have a significant causal association with the onset of PD. The present invention further analyzed the volume of thalamic subregions and found that there was a direct causal relationship between PD and the lateral paraventricular nucleus (OR 2.25, 95% CI 1.38-3.67, PFDR 0.015) and the ventrolateral dorsal nucleus (OR 1.78, 95% CI 1.33-2.38, PFDR 0.0041), which may be related to the onset of PD. In DWI analysis, the mean diffusion tensor pattern (MO) of the posterior part of the left internal capsule showed a reduced risk of PD (OR 0.68, 95% CI 0.54-0.86, PFDR 0.024). Likewise, weighted mean MO within the left auditory radiation (OR 0.33, 95% CI 0.17-0.63, PFDR 0.012) and MO within the right uncinate fasciculus (OR 0.44, 95% CI 0.25-0.75, PFDR 0.017) may be causally associated with a reduced risk of PD.

[0055] (2) Positive MR: Potential causal effect of IDP on PD severity and clinical symptoms

[0056] Regarding the severity of PD, an increase in the T2 star value in the right globus pallidus was causally associated with a decrease in the Hoehn and Yahr scale score (cont_HY), suggesting that an increase in the T2 star in the right globus pallidus could protect PD patients from increased disease severity (β-0.16, 95% CI -0.26--0.05, pfdr 0.0091, Figure 2 b).

[0057] The present invention then analyzed the potential causal relationship between IDPs and PD symptoms, including motor symptoms (HY, dyskinesia, symptom fluctuations, UPDRS) and non-motor symptoms (constipation, cognitive impairment, depression, hyposmia, insomnia, RBD, restless legs, SEADL, excessive daytime sleepiness, MMSE, MOCA). In DWI analysis, increased mean cell content fraction (ICVF) of the splenium of the corpus callosum was associated with the progression of depression in patients with Parkinson's disease (OR 6.67, 95% CI 1.93-23.13, pfdr 0.041, Figure 2 c). No significant results were found in other PD symptoms. Heatmaps show causal relationships with at least one significant MR signal.

[0058] (3) Positive MR: Potential causal effects of PD and its symptoms on IDPs

[0059] The present invention identified 7 additional IDPs with statistically significant causal relationships with PD ( Figure 2 ). The present invention found that PD leads to an increase in the temporal pole plane area (β-0.05, 95% CI -0.08--0.02, pfdr 0.040), which belongs to the upper part of the interhemispheric island of the brain. In addition, PD reduces the weighted mean fractional anisotropy (FA) in the left uncinate fasciculus (β-0.04, 95% CI -0.07--0.02, pfdr 0.017, Figure 2 d).

[0060] (4) Reverse MR

[0061] The relationship between IDPs and PD and its manifestations was detected in the forward MR, and there was no significant reverse causal relationship.

[0062] (5) Sensitivity analysis

[0063] In order to verify the results of the MR study, the present invention conducted a sensitivity analysis. In order to determine whether a single SNP can determine the results of MR, the present invention performed a Leave-One-Out test and did not find that any single SNP had an impact on the causal estimate. The present invention used the Steiger test to avoid reverse causality and found no significant evidence. In addition, the present invention used IVW Q and MR-EGGER Q tests for heterogeneity investigation, as well as Egger intercept and MR-PRESSO overall tests for multi-effect quality assessment. No heterogeneity or multi-effect quality was detected.

[0064] (6) Observational studies and MR analysis show consistent effects

[0065] Table 1 shows the demographic and clinical characteristics of individuals diagnosed with PD (n=85), iRBD (n=47), and HC (n=81). iRBD is a prodromal stage of α-synucleinopathies that converts to PD with a high conversion frequency. No significant differences were observed between healthy controls (HCs), iRBD, and PD patients in terms of age and gender.

[0066] Table 1 Demographic information of subjects

[0067]

[0068]

[0069] The number of people finally included in the discovery set and validation set were: 81 healthy controls and 85 PD patients in the discovery set, and 20 healthy controls and 20 PD patients in the validation set.

[0070] MR studies have revealed that increased striatal T2 star values ​​and enlarged thalamus as well as Pt and VPL are causally related to PD, so the present invention compared T2 star values ​​in HC, iRBD and PD patients. Figure 3 As shown in the figure, when the median T2 star of the striatum was compared, the present invention found that iRBD patients had a significant increase on both sides of the caudate nucleus, while PD patients had a decrease (ANCOVA: left side: p<2*10^-16, right side: p<2*10^-16; HC vs iRBD: left side: p=0.00119, right side: p=0.0222; HC vs PD: left side: p=0.17, right side: p=0.03). Similarly, the same trend was observed on both sides of the globus pallidus and putamen ( Figure 3 a, e). Regarding thalamic volume, the present invention found that there were significantly enlarged left thalamus, left PT and right VPL in iRBD and PD groups ( Figure 3b, f). In addition, the MR study of the present invention also showed that the FA value of the left uncinate fasciculus of PD patients was reduced and the polar plane (planum polare) was thinned. In the observation study, the present invention observed that the FA value of the left uncinate fasciculus of PD patients was reduced (p<2*10^-16, Figure 3 d, g), and polar plane thinning (p<2*10^-16, Figure 3 c, h).

[0071] The present invention established the relationship between the imaging indexes with significant differences and PD through logistic regression, and the regression equation is:

[0072] Prediction score = -0.047518*median T2 star value of left putamen + 0.016002*median T2 star value of right caudate + 0.000141*left thalamus volume + 0.00196*right VPL volume + 0.22711*left PT volume -14.868015*left uncus FA value -3.926307*left temporal pole plane thickness + 15.273922.

[0073] When the prediction score is greater than -1.347, the possibility of PD is high, and when it is less than -1.347, the possibility of PD is low.

[0074] To determine whether these imaging markers can serve as potential biomarkers for early PD (HY score ≤ 2), ROC curve analysis was performed to analyze the diagnostic specificity and sensitivity ( Figure 4 ). The ROC curves of imaging markers and PD showed that the AUC of the median T2 star in the left putamen was 0.63 (95% CI 0.5389-0.7207), the AUC of the left thalamus volume was 0.63 (95% CI 0.5418-0.7255), etc. All these significant markers were combined through the regression equation to significantly increase the AUC to 0.80 (95% CI 0.73-0.88).

[0075] The area under the ROC curve of the combined equation was 0.7000 (95% CI 0.54-0.86) in the validation set. Figure 4 .

[0076] Embodiment 2:

[0077] The clinical diagnostic steps of the Parkinson's disease diagnostic model of the present invention are as follows:

[0078] 1) Obtain 13 imaging indicators through brain imaging;

[0079] 2) The predicted score was calculated by the regression equation of the prediction model; predicted score = -0.047518*median T2 star value of the left putamen + 0.016002*median T2 star value of the right caudate nucleus + 0.000141*left thalamus volume + 0.00196*right VPL volume + 0.22711*left Pt volume -14.868015*left uncus FA value -3.926307*left temporal pole plane thickness + 15.273922.

[0080] 3) When the prediction score is greater than -1.347, the possibility of PD is high, and when it is less than -1.347, the possibility of PD is low.

[0081] The preferred embodiments of the present invention have been specifically described above, but the present invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. An imaging biomarker for combined diagnosis of Parkinson's disease, characterized in that: The imaging biomarkers for the combined diagnosis of Parkinson's disease consist of the median T2 star value of the left putamen, the median T2 star value of the right caudate nucleus, the left thalamus volume, the right VPL volume, the left Pt volume, the left uncus FA value, and the left temporal pole plane thickness.

2. A detection product for combined diagnosis of imaging biomarkers of Parkinson's disease, characterized in that: A product comprising imaging biomarker data for joint diagnosis of Parkinson's disease as described in claim 1 detected by SWI, T1, and SWI multimodal magnetic resonance imaging.

3. Use of the imaging biomarker for combined diagnosis of Parkinson's disease as claimed in claim 1 or the detection product of the imaging biomarker for combined diagnosis of Parkinson's disease as claimed in claim 2 in establishing a Parkinson's disease diagnostic system or diagnostic product.

4. A Parkinson's disease diagnostic model, characterized in that: The calculation formula of the Parkinson's disease diagnosis model is as follows: Prediction score = -0.047518*median T2 star value of left putamen + 0.016002*median T2 star value of right caudate + 0.000141*left thalamus volume + 0.00196*right VPL volume + 0.22711*left Pt volume -14.868015*left uncus FA value -3.926307*left temporal pole plane thickness + 15.273922; When the prediction score is greater than -1.347, the possibility of being diagnosed with Parkinson's disease is high; when the prediction score is less than -1.347, the possibility of being diagnosed with Parkinson's disease is low.

5. A diagnostic system for Parkinson's disease, characterized in that: The diagnostic system comprises an analysis and judgment module; the analysis and judgment module contains the diagnostic model as claimed in claim 4, which is used to output a diagnosis result of whether the sample suffers from Parkinson's disease.

6. The diagnostic system for Parkinson's disease according to claim 5, characterized in that: The diagnostic system further comprises a detection module, which detects imaging biomarker data for joint diagnosis of Parkinson's disease in the sample to be tested and transmits the data to the analysis and judgment module.

7. Use of the diagnostic system according to claim 5 or 6 in preparing a diagnostic product for Parkinson's disease.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is a computer program that realizes the functions of the Parkinson's disease diagnosis model as claimed in claim 4 or the Parkinson's disease diagnosis system as claimed in claim 5.

9. A Parkinson's disease diagnosis product, characterized in that: The diagnostic products include: The detection product for the value of imaging biomarkers for combined diagnosis of Parkinson's disease as claimed in claim 1; and a diagnostic system for Parkinson's disease as claimed in claim 5 or 6, or a computer-readable storage medium as claimed in claim 8.

Images