The Personal Parkinson Project, De novo
Parkinson’s disease (PD) is the second most prevalent degenerative brain disease. The recent failures of disease modifying interventions were explained by the negative outcomes of clinical trials where the tested compounds were in fact effective, but the current clinical scales were not sensitive enough to detect genuine effects on the underlying disease process. There was therefore an urgent need for development of objective outcome measures that could be obtained longitudinally and unobtrusively while patients perform in their own natural home environment.
In the Parkinson field, the MDS-UPDRS scale remained the gold standard to document the outcomes in clinical trials for Parkinson’s disease (PD). However, there were increasing concerns about the validity and reliability of such clinical rating scales, for a variety of reasons. Digital biomarkers may be able to overcome the limitations of the MDS-UPDRS, as they continuously collects real-time data, during the patient’s day to day activities. In this study we were interested in developing algorithms to track progression of bradykinetic gait, tremor, and autonomic dysfunction by using a Study watch.
The PPP de novo cohort consisted of 103 patients with de novo (i.e., newly diagnosed and previously untreated) Parkinson’s disease. The observation in the earliest course of the disease was highly relevant for the development of disease modifying interventions, which were likely to have the biggest potential in the earliest phases of the disease, when the loss of substantia nigra cells is minimal. The concept of digital biomarkers was developed using the Study Watch, and was tested with the data collected in the PPP de novo cohort.
Despite the COVID-19 pandemic, we deployed a large PPP de novo cohort for the development of digital biomarkers. Patient engagement was an essential strategy for a successful project. Based on representative labeled data, we showed that it is feasible to monitor tremor, arm swing during gait and heart rate based on wrist IMU and PPG data, collected through passive monitoring. While the PPP de novo data collection was underway, we have validated these digital biomarkers for measuring disease progression using the watch sensor data from the original Personalized Parkinson Project cohort (including PD patients diagnosed for <5 years, followed up for 2 years). The ongoing evaluation using the PPP de novo cohort will shed light on the digital biomarkers' sensitivity to disease progression in unmedicated patients. This is a crucial step towards their uptake in clinical trials investigating disease-modifying treatments.
The developed algorithms were made available in the ParaDigMa (Parkinson’s Disease Digital Markers) python toolbox
