ARTIFICIAL INTELLIGENCE IN DRUG REPURPOSING FOR PARKINSON’S DISEASES: A COMPREHENSIVE REVIEW

Harshadha B.*, Revathi S., Dinesh C., Gokul S., Sinega C.

ABSTRACT

Parkinson’s disease (PD), the second most common neurodegenerative disorder, remains a major challenge in drug discovery due to its complex pathophysiology involving α-synuclein aggregation, oxidative stress, mitochondrial dysfunction, and neuroinflammation. Conventional drug development is costly, time-intensive, and limited by low success rates, highlighting the urgent need for innovative strategies. Drug repurposing, supported by artificial intelligence (AI), offers a transformative approach by leveraging existing safety profiles of approved drugs while reducing both cost and time. This review synthesizes progress between 2015 and 2025 in AI-driven methodologies, including machine learning, deep learning, and network-based frameworks, which enable systematic identification of novel drug-disease associations. Compared with traditional screening (<1% success rate), AI-guided approaches achieve substantially higher hit rates (12.5–21.4%), reduce development costs from $1.3 billion to $250,000, and shorten timelines from 15 years to approximately 4 years. Promising candidates, such as efavirenz, omaveloxolone, and cyproheptadine, demonstrate therapeutic potential by modulating disease mechanisms and offering neuroprotection. Moreover, integration of multi-omics data, knowledge graphs, and precision medicine tools has enabled accurate PD subtyping (95% accuracy), biomarker discovery, and patient stratification. Despite challenges in data integration, regulation, and clinical validation, AI-enabled drug repurposing represents a paradigm shift in PD therapeutics, paving the way for personalized treatments.

Keywords: Artificial intelligence, Parkinson's disease, and Drug repurposing.