Parkinson’s disease (PD) is a neurological disorder that is increasing at a faster rate than any other neurological disease. The diagnosis of PD is primarily based on clinical criteria, which include motor symptoms such as slowness of movement, resting tremors, and rigidity.
However, nonmotor symptoms like sleep disorder, constipation, apathy, and loss of smell can appear years or even decades before the clinical symptoms manifest. Moreover, the current process for identifying PD in patients with Parkinson-like symptoms can often be inconclusive, with an overall diagnostic accuracy of 80%. Â
As per CS Central Science and reported by Neuroscience News, a new AI tool called CRANK-MS shows promise in diagnosing Parkinson’s disease up to 15 years before the onset of symptoms, according to a study conducted by researchers. The study involved 78 individuals from Spain who provided blood samples between 1993 and 1996 and were followed for 15 years. Out of the participants, 39 eventually received a Parkinson’s diagnosis, while the remaining 39 did not.Â
Researchers used the artificial intelligence program CRANK-MS to evaluate the data and distinguish between those who acquired Parkinson’s disease and those who did not. CRANK-MS, unlike more traditional techniques, did not simplify the investigation of the chemical characteristics of interest.
Instead, it was comprehensive, predicting metabolic pathways based on a model and identifying the underlying molecules at the same time. Within 15 years, scientists were able to identify cause chemical combinations in 96% of new cases of Parkinson’s disease using this approach.Â
Traditional clinical testing by movement disorder experts has an 80% accuracy rate, however CRANK-MS has a 90% accuracy rate in identifying Parkinson’s disease. Using a second cohort of 274 people, the equipment was demonstrated to have an accuracy of 84.3% for diagnosing Parkinson’s disease from skin sebum samples.Â
Dr. Daniel Truong, neurologist and medical director of MemorialCare Orange Coast Medical Center’s Parkinson’s and Movement Disorder Institute, has commented on the potential benefits of the AI tool’s ability to identify Parkinson’s disease years before symptoms appear. Dr. Truong emphasized the need of dependable diagnostic techniques for early detection of risk factors, as well as the possible applicability of similar technologies to the diagnosis of other disorders.Â
Dr. Truong and Dr. Julie Pilitsis, a board-certified neurosurgeon, both emphasized the need of reproducing these findings in bigger cohorts and performing more study into metabolic pathways and their connections to Parkinson’s disease. When considering the use of this AI technology in clinical settings, the issues of sample collecting and data interpretation were also raised.
