🤩 New AI Model Diagnoses Parkinson's Disease with 96% Accuracy
Nearly 12 million people worldwide suffer from Parkinson's disease. While the incidence of Parkinson's rises significantly after the age of 60, an estimated 5-10% are diagnosed between the ages of 40 and 50. The disease remains incurable, but early detection could help slow its progression.
At least one in four patients is misdiagnosed, as Parkinson's disease can be easily confused with other conditions, especially in the early stages. This delay in accurate diagnosis means lost opportunities for treatment, making it much more challenging to help the patients.
Researchers at the University of Florida have developed AI-powered software that will help doctors detect the main types of Parkinson's disease and related conditions based on the analysis of the patient's MRI brain scans.
The model was tested on several hundred patients aged 40 to 80 with clinically confirmed Parkinson's disease. The algorithm achieved a diagnostic accuracy of over 96%, significantly outperforming the 55-78% accuracy typically observed during the first five years of conventional assessments.
The developers are looking forward to the software's widespread deployment and hope to obtain approval from the U.S. Food and Drug Administration soon.
🤩 Why It's Useful
According to David Vaillancourt, chair and a professor in the University of Florida Department of Applied Physiology and Kinesiology, machines will not substitute humans in diagnostics. Still, the new tool could help even the most experienced doctors increase diagnostic efficacy between different disorders.
The algorithm could reduce the examination and diagnosis times for thousands of patients. The software is compatible with MRI machines from all major manufacturers, making it even more accessible.
More on the topic:
