Synthetic intelligence platform diagnoses dystonia with excessive accuracy in 0.36 seconds

0
77
Artificial intelligence platform diagnoses dystonia with high accuracy in 0.36 seconds

A diagram of the DystoniaNet platform's process in diagnosing dystonia using MRI. Credit: Earth Eye and Ear

Mass Eye and Ear researchers have developed a unique diagnostic tool that can be used to detect dystonia using MRI scans. This is the first technology of its kind that allows an objective diagnosis of the disorder. Dystonia is a potentially disabling neurological condition that causes involuntary muscle contractions, resulting in abnormal movements and postures. It is often misdiagnosed and can take up to 10 years to reach a correct diagnosis.

In a new study published in PNAS on September 28, researchers developed an AI-based deep learning platform called DystoniaNet to compare brain MRIs from 612 people, including 392 patients with three different forms of isolated focal dystonia and 220 healthy people. The platform diagnosed dystonia with an accuracy of 98.8 percent. During the process, the researchers identified a new biological marker for dystonia in the microstructural neural network. They believe that with further testing and validation, DystoniaNet can be easily incorporated into clinical decision-making.

"Currently there is no biomarker for dystonia and no 'gold standard' test for diagnosis. As a result, many patients have to undergo unnecessary procedures and see various specialists until other diseases are ruled out and a diagnosis of dystonia is made is, "he told Kristina Simonyan, MD, Ph.D., Dr. med., Director of Laryngology Research at Mass Eye and Ear, Associate Neuroscientist at Massachusetts General Hospital, and Associate Professor of Ear, Nose and Throat Surgery at Harvard Medical School. "There is an urgent need to develop, validate and integrate objective test tools for diagnosing this neurological condition. Our results show that DystoniaNet can fill this gap."

A disorder that is known to be difficult to diagnose

Approximately 35 out of 100,000 people have isolated or primary dystonia – a prevalence that is likely to be underestimated given the current challenges in diagnosing the disorder. In some cases, dystonia can be due to a neurological event such as Parkinson's disease or a stroke. However, the majority of isolated dystonia cases have no known cause and affect a single group of muscles in the body. These so-called focal dystonias can lead to disabilities and problems with the physical and emotional quality of life.

The study included three of the most common types of focal dystonia: laryngeal dystonia, characterized by involuntary movements of the vocal cords that can lead to speech difficulties (also known as convulsive dysphonia); Cervical dystonia, in which the muscles of the neck contract and the neck tilt in unusual ways; Blepharospasm, a focal dystonia of the eyelid that causes involuntary twitching and forceful closing of the eyelids.

Traditionally, a diagnosis of dystonia is made based on clinical observations, said Dr. Simonyan. Previous studies have shown that the inter-clinician agreement on diagnosing dystonia based on purely clinical assessments is only 34 percent, and have reported that approximately 50 percent of cases are misdiagnosed or underdiagnosed on a first patient visit.

DystoniaNet could be integrated into medical decision-making

DystoniaNet uses deep learning, a specific type of AI algorithm, to analyze data from the individual MRI and identify more subtle differences in brain structure. The platform is capable of detecting clusters of abnormal structures in multiple regions of the brain that are known to drive processing and engine commands. These small changes cannot be seen on MRI with the naked eye, and the patterns are only discernible by the platform's ability to take 3D brain images and enlarge their microstructural details.

"Our study suggests that the implementation of the DystoniaNet platform for the diagnosis of dystonia would be transformative to the clinical management of this disorder," said first study author Davide Valeriani, Ph.D., a postdoctoral fellow in the dystonia and speech motor lab the Mass Eye and Ear and Harvard Medical School. "What's important is that our platform is efficient and interpretable for clinicians by providing the patient's diagnosis, the AI's confidence in that diagnosis, and information about which brain structures are abnormal."

DystoniaNet is a patent-pending proprietary platform developed by Dr. Simonyan and Dr. Valeriani was developed in collaboration with Mass General Brigham Innovation. The technology interprets an MRI scan for microstructural biomarkers in 0.36 seconds. DystoniaNet was trained using Amazon Web Services' computerized cloud platform. The researchers believe that this technology can be easily transferred to the clinical setting, for example by integrating it into an electronic patient record or directly into the MRI scanner software. If DystoniaNet shows a high likelihood of dystonia on an MRI, a doctor can use this information to safely confirm the diagnosis, take future action, and suggest treatment immediately. There is no cure for dystonia, but some treatments can help reduce the occurrence of cramps associated with dystonia.

Future studies will cover more types of dystonia and will include studies in multiple hospitals to further validate the DystoniaNet platform in a larger number of patients.

Sun exposure and latitude related to the development of dystonia symptoms

More information:
Davide Valeriani el al., "A Novel Microstructural Neural Network Biomarker for Dystonia Diagnosis Identified by a DystoniaNet Deep Learning Platform", PNAS (2020). www.pnas.org/cgi/doi/10.1073/pnas.2009165117

Provided by
Massachusetts Eye and Ear Infirmary

Quote::
The artificial intelligence platform diagnoses dystonia with high accuracy in 0.36 seconds (2020, September 28).
accessed on September 28, 2020
from https://medicalxpress.com/news/2020-09-artificial-intelligence-platform-dystonia-high.html

This document is subject to copyright. Apart from fair treatment for the purpose of private study or research, no
Part may be reproduced without written permission. The content is provided for informational purposes only.

LEAVE A REPLY

Please enter your comment!
Please enter your name here