Because of the similarity between PD and MSA-P symptoms in early disease stages, the differential diagnosis between these two diseases can be very challenging. Indeed, despite recent efforts, no validated objective marker is currently available to guide the clinician in this differential diagnosis. The need of such markers is hence very high in the neurology community, particularly given the severity of the prognosis of MSA-P.
The innovative goal of our project is two-fold, i.e. to develop a non-invasive objective digital marker to assist in the early differential diagnosis between PD and MSA-P, and the diagnosis of PD compared to healthy controls (HC). Moreover, our ambition is to develop a very low cost and portable digital tool. In case of success, this would allow a large scale and world-wide use; every neurologist would have the possibility to use the digital marker in his clinic or private office.
The originality of our project stands in the methodology we adopt to achieve its innovative and ambitious goal. Indeed, since dysarthria is a common and early symptom in both diseases and of different origin, our approach is to use dysarthria, through a digital processing of voice recordings of patients, as a vehicle to distinguish between PD and MSA-P, and between PD and HC, in early disease stages. This topic is an emerging research field with several teams around the world working on the discrimination between PD and HC based on voice recordings. Hitherto, only one very recent study has used voice recordings to address the more challenging distinction between PD and other neurodegenerative parkinsonian disorders such as MSA. Beyond previous attempts, the originality of our methodology stands (as will be explained in section II) in the clinical protocol setting for data collection and in the signal processing techniques that we use to achieve the desired objectives.
In summary, our objective is to perform two pilot studies in parallel:
A proof-of-concept study on the differential diagnosis between PD and MSA-P which has a high potential of success and an immediate utility and applicability.
A study on the distinction between PD and HC in order to put solid scientific foundations for an early diagnosis of PD. Indeed, analyzing PD speech in a differential diagnosis setting will yield a deeper and finer understanding of PD speech characteristics than comparing only PD and HC speech.
To do so, we have to address and solve several scientific challenges:
1. We have to determine and automatically extract the discriminative speech features with acceptable classification accuracy (higher than 80%).
2. We have to develop a classifier which is robust to intra/inter speaker variations and also to recording conditions.
3. The final digital softwares have to be easy to use by the clinicians and have to be flexible enough to incorporate new knowledge and/or new data.
Our project has also mid and long-term ambitious objectives. Indeed, based on the results obtained in this project, we plan to launch a large multi-center study involving de novo patients with parkinsonian syndromes (i.e. patients who see for the first time in their life a neurologist because of their parkinsonism). This trial may start as early as around the middle of Voice4PD-MSA life time, if the results of the clinical proof-of-concept study are conclusive or promising. General neurologists will participate in this study by recording patients with a parkinsonian syndrome in their office at first visit. Disease progression will then be followed and the patients’ voice/speech will be recorded every 12 months for at least four years.
From this perspective, Voice4PD-MSA can be seen as the necessary first step towards achieving these ultimate mid and long-term crucial and ambitious objectives.