Mount Sinai spinout developed algorithm that accurately diagnoses early-stage Parkinson's, per new study

An artificial-intelligence-enabled tech has been found to accurately identify early-stage Parkinson’s disease in living patients, according to a new study. 

The research, in a collaboration with the Michael J. Fox Foundation for Parkinson’s Research, used technology developed by a Mount Sinai spinout focused on cancer diagnostics. The study leveraged PreciseDx’s algorithms to detect a protein in salivary glands, which is linked to the disease. It was able to detect Parkinson’s with 99% sensitivity and 99% specificity and was more accurate than human pathologists in predicting the disease. 

“These findings show the potential for technology to aid in diagnosis of Parkinson’s disease,” Jamie Eberling, Ph.D., senior vice president of research programs at the Michael J. Fox Foundation, said in a press release. “Objective diagnostic tools, especially early in disease, are critical to drive care decisions and to design trials toward better treatments and cures.”

Diagnosing Parkinson’s, like any neurodegenerative disease, is notoriously challenging. There is not much access to brain tissue for research, and symptoms are relatively generic, explained John Crary, M.D., a clinical diagnostic neuropathologist and professor at the Mount Sinai Icahn School of Medicine who worked on the study. With Parkinson’s, an autopsy has historically been the definitive diagnostic method, he said. However, biopsies can also reveal a protein that signifies Parkinson's, a promising development. 

“You can’t treat somebody once they’ve died, so you’re stuck,” Crary said. “There’s a lot of innovation in biomarkers for Parkinson's disease,” he went on, “but the idea that you could get a definitive tissue diagnosis from a peripheral biopsy is really enticing.”

In a past Michael J. Fox Foundation study, Crary and other researchers found success in identifying the distinctive Parkinson’s protein using biopsies. But they had to do so manually, sorting through the biopsied slides under a microscope. The method was “extremely laborious,” according to Crary, and not realistic for large-scale adoption. The latest study proved AI could assist in that process.

The hope, Crary said, is for the AI to one day be able to identify and diagnose Parkinson's in living patients without biopsies altogether.