Heuron, a medical AI company, said Thursday that the results of a study using its Parkinson's disease solutions Heuron IPD and Heuron NI were published in the American Journal of Neuroradiology (AJNR), a journal of the American Academy of Neurological Radiology.
The study, co-authored by Professor Seo Bae-seon of the Department of Radiology at Severance Hospital, Dr. Heo Hwan of Heuron's Neurodegenerative Disease Research Team, and Professor Seo Jong-hyeon of the Department of Radiology at Asan Medical Center, used Heuron's solution to detect and quantify the nigrosome-1 region automatically, a key indicator of Parkinson's disease.
(Courtesy of Heuron)
The researchers analyzed MRI data from 237 patients with Parkinson's disease and controls. Heuron's AI solutions, Heuron IPD and Huron NI, were used to analyze the data. Based on MRI, Heuron IPD automatically detects and visualizes nigrosome regions and classifies them as normal or abnormal, while Heuron NI quantitatively measures the volume of the nigrosome-1 region.
Nigrosomes are characterized by decreased volume due to damage to dopamine neurons in Parkinson's disease and are used as a marker for diagnosis and disease progression. In the study, the Huron IPD had a sensitivity of 97 percent, specificity of 95.7 percent, and accuracy of 96.6 percent in identifying patients. Heuron NI had a sensitivity of 96.4 percent and an accuracy of 89.9 percent.
The team also quantitatively identified a reduction in negrosome volume according to the Hoehn and Yahr Scale (H&Y Scale), which measures the severity of Parkinson's disease. This demonstrates that MRI alone can provide an objective diagnosis, unlike existing diagnostic methods that rely on PET-CT scans or visual inspection.
“This study confirms that our solution can help diagnose Parkinson's disease early and identify its progression,” Heuron CEO Shin Dong-hoon said.
The study, “Deep learning-based algorithm for automatic quantification of nigrosome-1 and Parkinsonism classification using susceptibility map-weighted MRI,” was published in a recent issue of the Journal of Neuroscience.