Artificial Intelligence could be used to Diagnose Dystonia

Mass Eye and Ear researchers have discovered a unique diagnostic tool that can detect dystonia from MRI scans. It is the first technology of its kind to provide an objective diagnosis of the disorder. Dystonia is a potentially disabling neurological condition which causes involuntary muscle contractions, driving to abnormal movements and postures. It is often mistreated and sometimes takes people up to 10 years to get a correct diagnosis.

A new study by PNAS researches shows that they have developed an AI-based deep learning platform on September 28, called DystoniaNet to compare brain MRIs of 612 people. These numbers include 392 patients with three separate forms of isolated focal dystonia and 220 healthy individuals. DystoniaNet diagnosed dystonia with 98.8% accuracy. During the experiment, researchers detected a new microstructural neural network biological marker of dystonia. With the next measures, such as testing and validation, they believe the platform can be easily incorporated into clinical decision-making.

Kristina Simonian is a senior study author, MD, PhD, Dr med, Director of Laryngology Research at Mass Eye and Ear, Associate Neuroscientist at Massachusetts General Hospital, and Associate Professor of Otolaryngology (Head and neck surgery at Harvard Medical School). She says, "There is currently no biomarker of dystonia and no 'gold standard test for its diagnosis. Because of this, a lot of patients have to undergo unnecessary procedures and see different specialists until other diseases are ruled out, and the diagnosis of dystonia is set up." She adds, "There is a critical requirement to develop, validate, and incorporate objective testing tools for the diagnosis of this neurological condition, and our outcomes show that DystoniaNet may fill this gap."

A disorder notoriously hard to diagnose

Nearly 35 out of every 100,000 people have isolated or primary dystonia. It is prevalent, likely to be underestimated because of the current challenges in diagnosing this disorder. Dystonia can be a result of a neurological event like Parkinson's disease or a stroke in some cases. However, most of the isolated dystonia cases have an unknown cause and affect a single muscle group in the body. These focal dystonias can lead to disability and complications with the physical and emotional quality of life.

The study categorised focal dystonia in three, and they are Laryngeal dystonia, Cervical dystonia, and Blepharospasm. Laryngeal dystonia is also known as spasmodic dystonia, characterised by involuntary movements of the vocal cords that can cause difficulties with speech. Cervical dystonia causes the neck muscles to spasm and the neck to tilt unusually. Blepharospasm is a focal dystonia of the eyelid that causes involuntary twitching and forceful eyelid closure.

Dr Simonyan explained, "A dystonia diagnosis is traditionally made based on clinical observations." Past studies have explored the agreement on dystonia between clinicians based on purely clinical assessments are as low as 34%. These have also reported that around 50% of the cases go misdiagnosed or underdiagnosed at a first patient visit.

DystoniaNet could help making Medical Decisions

DystoniaNet uses a particular type of AI algorithm called deep learning to analyse data from individual MRI and identify subtler differences in brain structure. The platform can identify clusters of abnormal structures in several regions of a human brain that are known to control processing and give commands. A naked eye cannot catch these small changes in MRI. And the patterns are only evident through the platform's ability to take 3D brain images and zoom in to their microstructural details.

The first study author at Mass Eye and Ear, and PhD, Davide Valeriani elucidated, "Our study suggests that the implementation of the DystoniaNet platform for dystonia diagnosis would be transformative for the clinical management of this disorder." He adds, "Importantly, our platform was designed to be efficient and interpretable for clinicians, by providing the patient's diagnosis, the confidence of the AI in that diagnosis, and information about which brain structures are not normal."

Being a patent-pending platform, DystoniaNet interprets an MRI scan for microstructural biomarker in 0.36 seconds. The platform has also been trained using Amazon Web Services computational cloud platform. The researchers believe this technology can easily be pushed into the clinical setting by being integrated into an electronic medical record or directly in the MRI scanner software. A physician can use the tool for diagnosis and suggest a course of treatment without any delay if DystoniaNet finds a high possibility of dystonia in the MRI. Although dystonia can be cured, some treatments can help reduce the incident of dystonia-related spasms.