Antemortem detection of Parkinson's disease pathology in peripheral biopsies using artificial intelligence.

The diagnosis of Parkinson's disease (PD) is challenging at all stages due to variable symptomatology, comorbidities, and mimicking conditions. Postmortem assessment remains the gold standard for a definitive diagnosis. While it is well recognized that PD manifests pathologically in the central nervous system with aggregation of α-synuclein as Lewy bodies and neurites, similar Lewy-type synucleinopathy (LTS) is additionally found in the peripheral nervous system that may be useful as an antemortem biomarker. We have previously found that detection of LTS in submandibular gland (SMG) biopsies is sensitive and specific for advanced PD; however, the sensitivity is suboptimal especially for early-stage disease. Further, visual microscopic assessment of biopsies by a neuropathologist to identify LTS is impractical for large-scale adoption. Here, we trained and validated a convolutional neural network (CNN) for detection of LTS on 283 digital whole slide images (WSI) from 95 unique SMG biopsies. A total of 8,450 LTS and 35,066 background objects were annotated following an inter-rater reliability study with Fleiss Kappa = 0.72. We used transfer learning to train a CNN model to classify image patches (151 × 151 pixels at 20× magnification) with and without the presence of LTS objects. The trained CNN model showed the following performance on image patches: sensitivity: 0.99, specificity: 0.99, precision: 0.81, accuracy: 0.99, and F-1 score: 0.89. We further tested the trained network on 1230 naïve WSI from the same cohort of research subjects comprising 42 PD patients and 14 controls. Logistic regression models trained on features engineered from the CNN predictions on the WSI resulted in sensitivity: 0.71, specificity: 0.65, precision: 0.86, accuracy: 0.69, and F-1 score: 0.76 in predicting clinical PD status, and 0.64 accuracy in predicting PD stage, outperforming expert neuropathologist LTS density scoring in terms of sensitivity but not specificity. These findings demonstrate the practical utility of a CNN detector in screening for LTS, which can translate into a computational tool to facilitate the antemortem tissue-based diagnosis of PD in clinical settings.

Disease modification and biomarker development in Parkinson disease: Revision or reconstruction?

A fundamental question in advancing Parkinson disease (PD) research is whether it represents one disorder or many. Does each genetic PD inform a common pathobiology or represent a unique entity? Do the similarities between genetic and idiopathic forms of PD outweigh the differences? If aggregates of α-synuclein in Lewy bodies and Lewy neurites are present in most (α-synucleinopathies), are they also etiopathogenically significant in each (α-synuclein pathogenesis)? Does it matter that postmortem studies in PD have demonstrated that mixed protein-aggregate pathology is the rule and pure α-synucleinopathy the exception? Should we continue to pursue convergent biomarkers that are representative of the diverse whole of PD or subtype-specific, divergent biomarkers, present in some but absent in most? Have clinical trials that failed to demonstrate efficacy of putative disease-modifying interventions been true failures (shortcomings of the hypotheses, which should be rejected) or false failures (shortcomings of the trials; hypotheses should be preserved)? Each of these questions reflects a nosologic struggle between the lumper's clinicopathologic model that embraces heterogeneity of one disease and the splitter's focus on a pathobiology-specific set of diseases. Most important, even if PD is not a single disorder, can advances in biomarkers and disease modification be revised to concentrate on pathologic commonalities in large, clinically defined populations? Or should our efforts be reconstructed to focus on smaller subgroups of patients, distinguished by well-defined molecular characteristics, regardless of their phenotypic classification? Will our clinical trial constructs be revised to target larger and earlier, possibly even prodromal, cohorts? Or should our trials efforts be reconstructed to target smaller but molecularly defined presymptomatic or postsymptomatic cohorts? At the Krembil Knowledge Gaps in Parkinson's Disease Symposium, the tentative answers to these questions were discussed, informed by the failures and successes of the fields of breast cancer and cystic fibrosis.

Longitudinal Change of Clinical and Biological Measures in Early Parkinson's Disease: Parkinson's Progression Markers Initiative Cohort.

OBJECTIVE: The objective of this study was to assess longitudinal change in clinical and dopamine transporter imaging outcomes in early, untreated PD. METHODS: We describe 5-year longitudinal change of the MDS-UPDRS and other clinical measures using results from the Parkinson's Progression Markers Initiative, a longitudinal cohort study of early Parkinson's disease (PD) participants untreated at baseline. We also provide data on the longitudinal change in dopamine transporter 123-I Ioflupane striatal binding and correlation between the 2 measures. RESULTS: A total of 423 PD participants were recruited, and 358 remain in the study at year 5. Baseline MDS-UPDRS total score was 32.4 (standard deviation 13.1), and the average annual change (assessed medications OFF for the treated participants) was 7.45 (11.6), 3.11 (11.7), 4(11.9), 4.7 (11.1), and 1.74(11.9) for years 1, 2, 3, 4, and 5, respectively (P < .0001 for the change over time), with a steeper change in year 1. Dopaminergic therapy had a significant effect on the change of MDS-UPDRS. There was a significant longitudinal change in dopamine transporter binding in all striatal regions (P < .001). There was a significant but weak correlation between MDS-UPDRS and dopamine transporter binding at baseline and years 1, 2, and 4, but no correlation between the rate of change of the 2 variables. CONCLUSIONS: We present 5-year longitudinal data on the change of the MDS-UPDRS and other clinical and dopamine transporter imaging outcome measures in early PD. These data can be used for sample size estimates for interventional studies in the de novo PD population. © 2018 The Authors. Movement Disorders published by Wiley Periodicals, Inc. on behalf of International Parkinson and Movement Disorder Society.