Personalized Treatment for Parkinson's Disease Using Artificial Intelligence

KAIST Develops Personalized Type Classification Platform

Domestic researchers have developed a technology that enables personalized treatment of chronic degenerative brain diseases such as Parkinson's disease using artificial intelligence (AI).

KAIST announced on the 15th that Professor Min Choi's research team from the Department of Brain and Cognitive Sciences, in collaboration with the Francis Crick Institute in the UK, has developed an AI-based platform that predicts individual disease subtypes in Parkinson's disease patients.

Since direct access to the brain cells of living patients with chronic degenerative brain diseases like Parkinson's is limited, attempts to predict mechanistic subtypes of patient diseases using AI based on cellular data from brain disease patients have not been made before.

The research team developed a platform that accurately predicts pathological subtypes of Parkinson's patients by learning only the nuclear, mitochondrial, and ribosomal image information of neurons differentiated from patient-derived human induced pluripotent stem cells (hiPSC). This allows classification of Parkinson's disease manifestations by biological mechanisms rather than the outwardly visible phenotypes that vary by patient. Through this, it becomes possible to diagnose molecular and cellular subtypes of Parkinson's disease patients with unknown causes, opening the door to personalized treatment. Additionally, since this platform uses a high-speed, large-scale screening system, it can also be utilized as a customized drug development pipeline suitable for pathological subtypes.

Until now, Parkinson's disease treatment has used a 'one-size-fits-all approach' based on probabilities without considering the individual patient's pathological state. This approach made it difficult to improve treatment efficacy due to mismatches between pathological causes and treatment methods.

The platform developed by the research team can precisely profile molecular and cellular information of individual patients' brain cells. Based on this, it can accurately diagnose patients' disease subtypes, ultimately enabling 'Precise medicine.' This leads to personalized medicine tailored to each individual, which is expected to significantly improve treatment outcomes.

This platform follows the 'disease in a dish' paradigm, using brain cells differentiated from induced pluripotent stem cells (iPSC), a technology awarded the Nobel Prize in Physiology or Medicine in 2012. It is recognized as one of the technologies that can overcome the limitations of animal models, which cannot accurately replicate human brains or obtain lesions directly in degenerative brain diseases. In particular, sequential imaging of target disease cells cultured in a dish allows tracking of a series of pathological events, providing the advantage of predicting drug response outcomes according to disease progression.

Professor Choi said, "This study specifically introduced a method to effectively train AI with biological data obtained in the laboratory to create a highly accurate disease subtype classification model," adding, "It will also be possible to classify subtypes of brain diseases with distinctly different individual patient symptoms, such as autism spectrum disorder, enabling the development of effective treatments."

The research results were published in the August issue of the international journal Nature Machine Intelligence (IF = 25.8). (Paper title: Prediction of mechanistic subtypes of Parkinson’s using patient-derived stem cell model)

This content was produced with the assistance of AI translation services.