Phenotype-Specific Screening Model for Deep Brain Stimulation in Parkinson's Disease.

BACKGROUND

Deep brain stimulation (DBS) is an effective therapy for Parkinson's disease (PD), yet referral for formal DBS evaluation is absent or delayed for many eligible patients. Standard candidacy assessments rely on OFF-ON medication testing and neuropsychological evaluation, which are time-consuming and often unavailable in routine practice. Additionally, traditional UPDRS-based metrics may inadequately capture DBS candidacy across heterogeneous PD motor phenotypes.

OBJECTIVES

The objective of this study is to evaluate a phenotype-specific, longitudinal screening approach to support earlier identification of PD patients who may receive a DBS referral.

METHODS

Retrospective longitudinal data from the Parkinson's Progression Markers Initiative was used to train supervised logistic regression models to estimate the probability of undergoing DBS during follow-up. Separate models were developed for bradykinesia/rigidity-predominant, tremor-predominant phenotypes, and the total cohort. Analyses were restricted to visits within a clinically plausible DBS screening window (age <75 years, Hoehn & Yahr stage 2-3, MoCA ≥26). Model performance was assessed using discrimination, threshold-based operating characteristics, calibration, and interpretability via SHAP values.

RESULTS

Across 1,227 patients, phenotype-specific and combined models demonstrated moderate discrimination for eventual DBS implantation (AUC 0.80-0.82). In bradykinesia/rigidity-predominant PD, greater functional impairment and daytime sleepiness consistently increased likelihood of DBS implantation, while higher motor severity reduced referral likelihood. However, in tremor-predominant disease, better cognition and higher motor severity were associated with increased referral likelihood, while impulse-control burden strongly decreased implantation likelihood.

CONCLUSIONS

A phenotype-specific, longitudinal screening approach can stratify DBS referral likelihood in PD. This approach supports earlier and more appropriate DBS referrals while complementing gold-standard evaluation.