Objective: To assess the feasibility of a smartphone-based application (“Dopaminometer”) for predicting abnormal vs. normal dopamine transporter (DaT) scans and dopaminergic ratios in individuals with prodromal (REM-sleep-behaviour disorder–RBD) and manifest Parkinson’s disease (PD).

Background: Abnormal DaT scans are increasingly used as entry criteria in PD trials. Because many trials aim to slow disease progression, RBD participants—poised to phenoconvert—represent an ideal prodromal cohort. However, DaT scans are costly and not always accessible. A non-invasive, cost-effective screening method to estimate dopaminergic function, a predictor of future conversion, is therefore needed. To date, no study has attempted to predict these pathophysiological metrics using smartphone-based active testing.

Method: Smartphone motor assessments for voice, gait, balance, reaction time, dexterity, rest and postural tremor were matched to DaT scans (±1 year). DaT scans were labelled normal or abnormal by a radiologist. Age-corrected striatal ratios were derived by comparing uptake in each region of interest to a reference region. A machine learning method, XGBoost classifier, distinguished abnormal vs. normal scans using both in-clinic and at-home data, while an XGBoost regressor predicted dopaminergic binding ratios in bilateral caudate and putamen. Models were evaluated via 5-fold cross-validation, stratified by subject, and compared to a benchmark model that used only the MDS-UPDRS-III.

Results: A total of 100 DaT scans (52 normal, 48 abnormal) were analyzed (5 healthy controls, 56 RBD, 39 PD). The optimal machine learning classifier trained using only the smartphone-based features achieved an AUC of 0.82 in differentiating normal and abnormal DaT scans, with a balanced accuracy of 0.74. Encouragingly, including the MDS-UPDRS-III in the model along with smartphone-based features increase the AUC to 0.84, and balanced accuracy to 0.84. The smartphone-based regressor showed errors comparable to MDS-UPDRS-III alone; combining smartphone features with MDS-UPDRS-III yielded the lowest overall error.

Conclusion: A remote, scalable “Dopaminometer” approach may serve as an accessible screening tool for prodromal and early-stage PD, potentially enabling earlier interventions and more efficient patient stratification in clinical trials.

Table 1.

Table 2.

Table 3.

Table 4.

Figure 1.

« Back to 2025 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/dopaminometer-a-smartphone-based-application-to-predict-striatal-dopaminergic-deficit-in-prodromal-and-manifest-parkinsons-disease/