An Integrative Nomogram for Identifying Early-Stage Parkinson's Disease Using Non-motor Symptoms and White Matter-Based Radiomics Biomarkers From Whole-Brain MRI

Zhenyu Shu, Peipei Pang, Xiao Wu, Sijia Cui, Yuyun Xu, Minming ZhangFrontiers in Aging Neuroscience 2020 · DOI

Abstract

Purpose: To develop and validate an integrative nomogram based on white matter (WM) radiomics biomarkers and nonmotor symptoms for the identification of early-stage Parkinson's disease (PD). Methods: The brain magnetic resonance imaging (MRI) and clinical characteristics of 336 subjects, including 168 patients with PD, were collected from the Parkinson's Progress Markers Initiative (PPMI) database. All subjects were randomly divided into training and test sets. According to the baseline MRI scans of patients in the training set, the WM was segmented to extract the radiomic features of each patient and develop radiomics biomarkers, which were then combined with nonmotor symptoms to build an integrative nomogram using machine learning. Finally, the diagnostic accuracy and reliability of the nomogram were evaluated using a receiver operating characteristic curve and test data, respectively. In addition, we investigated 58 patients with atypical PD who had imaging scans without evidence of dopaminergic deficit (SWEDD) to verify whether the nomogram was able to distinguish patients with typical PD from patients with SWEDD. A decision curve analysis was also performed to validate the clinical practicality of the nomogram. Results: The area under the curve values of the integrative nomogram for the training, testing and verification sets were 0.937, 0.922, and 0.836, respectively; the specificity values were 83.8, 88.2, and 91.38%, respectively; and the sensitivity values were 84.6, 82.4, and 70.69%, respectively. A significant difference in the number of patients with PD was observed between the high-risk group and the low-risk group based on the nomogram (P < 0.05). Conclusion: This integrative nomogram is a new potential method to identify patients with early-stage PD.

Introduction

Diagnosing Parkinson’s Disease (PD) in its early stages is a significant clinical challenge. Because the disease's onset is subtle, many patients are diagnosed only after reaching middle or late stages, when neuroprotective interventions are less effective. Current diagnostic protocols rely heavily on medical history and motor symptoms. However, early-stage PD often presents with "atypical" non-motor symptoms—such as sleep disorders or loss of smell—that overlap with other conditions, leading to diagnostic error rates as high as 25% among non-specialists.

While advanced imaging techniques like PET scans or specialized MRI of the substantia nigra can aid diagnosis, they are often expensive and not part of standard clinical workflows. This study explores the use of standard T1-weighted MRI scans—widely available and cost-effective—to detect microstructural changes in the brain's white matter (WM) that may serve as early indicators of PD.

Methodology and Data

The researchers developed a multi-step machine learning pipeline to create a predictive "nomogram." The process involved several key stages:

  1. Radiomics Extraction: Standard T1 MRI scans from the Parkinson’s Progress Markers Initiative (PPMI) were used. The whole-brain white matter was segmented, and 378 quantitative "radiomic" features (texture, shape, and intensity) were extracted.
  2. Feature Selection and Model Building: To prevent overfitting, the researchers used the mRMR (minimum redundancy maximum relevance) algorithm and LASSO (least absolute shrinkage and selection operator) to identify the most predictive features. A Gradient Boosting Decision Tree (GBDT) then generated a "radiomic biomarker" score.
  3. Integrative Nomogram: Finally, stepwise logistic regression combined the radiomic score with clinical non-motor symptoms—specifically impaired olfaction and cognitive decline—to produce the final diagnostic tool.
Workflow for building the radiomic model, including segmentation, feature extraction, selection, and model building.

The study utilized a primary cohort of 336 subjects (168 early-stage PD patients and 168 healthy controls) split into training (n=234) and test (n=102) sets. Additionally, a verification set of 58 patients with SWEDD (Scans Without Evidence of Dopaminergic Deficit) was used to test the model's ability to distinguish typical PD from clinical "look-alikes."

Results and Key Insights

The integrative nomogram demonstrated exceptional diagnostic performance:

  • Training Set: AUC of 0.937.
  • Test Set: AUC of 0.922 (82.4% sensitivity, 88.2% specificity).
  • Verification (SWEDD) Set: AUC of 0.836.

A critical finding was the role of non-motor symptoms. Among various subgroups, impaired olfaction (IO) showed the most significant correlation with the radiomic scores, suggesting a potential link between white matter integrity and the brain's olfactory processing pathways in early PD.

Comparison of Rad scores across different clinical subgroups, highlighting the high significance of impaired olfaction (IO) in distinguishing patients.

The model's success highlights that white matter microstructural alterations often precede visible gray matter atrophy. By fusing these imaging biomarkers with clinical observations like olfactory loss, the nomogram provides a more accurate diagnostic tool than either method alone. Furthermore, its high performance on the SWEDD dataset suggests it identifies genuine PD pathology rather than general neurological distress.

Discussion and Limitations

This study presents a "clinic-ready" approach by utilizing standard MRI sequences. However, some limitations remain. The sample size, while robust for this type of study, should be expanded through multi-center validation. Future research should also account for the potential structural impact of long-term medications and investigate whether the nomogram can track disease progression longitudinally. While its sensitivity is slightly lower than some invasive cerebrospinal fluid biomarkers, the nomogram offers a non-invasive and highly accessible alternative for early PD identification.

Summary updated

Abstract

<b>Purpose:</b> To develop and validate an integrative nomogram based on white matter (WM) radiomics biomarkers and nonmotor symptoms for the identification of early-stage Parkinson's disease (PD). <b>Methods:</b> The brain magnetic resonance imaging (MRI) and clinical characteristics of 336 subjects, including 168 patients with PD, were collected from the Parkinson's Progress Markers Initiative (PPMI) database. All subjects were randomly divided into training and test sets. According to the baseline MRI scans of patients in the training set, the WM was segmented to extract the radiomic features of each patient and develop radiomics biomarkers, which were then combined with nonmotor symptoms to build an integrative nomogram using machine learning. Finally, the diagnostic accuracy and reliability of the nomogram were evaluated using a receiver operating characteristic curve and test data, respectively. In addition, we investigated 58 patients with atypical PD who had imaging scans without evidence of dopaminergic deficit (SWEDD) to verify whether the nomogram was able to distinguish patients with typical PD from patients with SWEDD. A decision curve analysis was also performed to validate the clinical practicality of the nomogram. <b>Results:</b> The area under the curve values of the integrative nomogram for the training, testing and verification sets were 0.937, 0.922, and 0.836, respectively; the specificity values were 83.8, 88.2, and 91.38%, respectively; and the sensitivity values were 84.6, 82.4, and 70.69%, respectively. A significant difference in the number of patients with PD was observed between the high-risk group and the low-risk group based on the nomogram (<i>P</i> < 0.05). <b>Conclusion:</b> This integrative nomogram is a new potential method to identify patients with early-stage PD.
BibTeX
@inproceedings{shu2020integrative,
  author = {Zhenyu Shu and Peipei Pang and Xiao Wu and Sijia Cui and Yuyun Xu and Minming Zhang},
  title = {{An Integrative Nomogram for Identifying Early-Stage Parkinson's Disease Using Non-motor Symptoms and White Matter-Based Radiomics Biomarkers From Whole-Brain MRI}},
  booktitle = {Frontiers in Aging Neuroscience},
  year = {2020},
  doi = {10.3389/fnagi.2020.548616},
}
Figure