Objective: To assess whether data obtained from an objective movement recording system such as the Parkinson’s KinetiGraph (PKG) could predict the motor criteria used by movement disorder clinicians to assess suitability of Parkinson’s disease (PD)  patients for deep brain stimulation (DBS).  

Background: While suitability of PD patients for advanced therapies is readily recognized in specialized centers, they may be overlooked by clinicians who have less experience with these therapies.  The recent developments of objective measurement of PD signs by continuously worn sensors, raise the possibility of a threshold score that could trigger a referral to a specialized centre for further assessment.

Methods: Thirty-six PD patients referred to a German and a French Movement Disorders centre for consideration for DBS were classified as either “suitable on motor ground” or “not ready” for DBS and were required to wear a PKG logger continuously for one week.  Bradykinesia, dyskinesia and motor fluctuations scores from the PKG were analysed as predictors of these two categories. A machine learning approach was applied, with various weightings to all variables available in the standard PKG and a single score that best predicted the clinical sorting was derived.  This was retested on data obtained from PKGs worn by 28 Australian PD patients before and six months after insertion of DBS.

Results: The standard scores from the PKG weakly predicted suitability for DBS at the two European centres (the best agreement was 67%, achieved with the fluctuation score).  However, machine learning provided a score that predicted the clinical sorting with a 95% selectivity and 87.5% specificity. Furthermore, 93% of the patients in the Australian cohort were correctly assigned as suitable for DBS.

Conclusions: These findings suggest that data from the PKG may predict decisions made by clinicians in sorting patients into “suitable on motor ground” or “not ready” for DBS.  The device could help to trigger timely patients’ referral to a specialised centers based on motor grounds.  However, further validation of the developed score with larger independent cohorts is needed before the reliability of this score can be proven for this clinical use.

« Back to 2017 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/an-evaluation-of-the-parkinsons-kinetigraph-pkg-as-a-tool-to-support-deep-brain-stimulation-eligibility-assessment-in-patients-with-parkinsons-disease/