UCB and Sharecare conducted a three-month prospective real-life study in the United States involving 82 patients with moderate to severe autoimmune myasthenia gravis.
The participants sent selfie videos to the investigators, which were analysed automatically by a margin reflex distance (MDR1) measurement tool developed using an artificial neural network:
- these measurements proved to be effective, with a strong correlation between their results and those of manual measurements carried out on the images,
- even though the participants’ smartphones (60 different models) had transmitted images of very different quality (definition, light, camera angle, etc.).
- The patients also sent information on their symptoms on a daily basis via their smartphones:
- exacerbations of the disease for 55% of them, on average for a total of 6.3 days over the course of the study,
- a significant increase in the MG-ADL score, based on patients’ self-reports, during exacerbations (7 on average versus 0.3 outside exacerbations),
- a correlation between this score and their feeling of exacerbation of the disease (“yes”, “no” or “maybe”),
- the existence of four sub-groups of patients reporting exacerbations, identified using machine learning methods and each characterised by a particular profile of symptoms and their severity.
In myasthenia, perhaps even more than in other diseases, the possibility of remote, real-life assessment is invaluable because of the wide fluctuation in symptoms and the possibility of a myasthenic crisis, which can be life-threatening.