AI-Based Prediction of Psoriasis Progression to Psoriatic Arthritis

This study is a technical report that investigates the use of artificial intelligence (AI) to predict the progression of psoriasis (PsO) to psoriatic arthritis (PsA) using a real-world dataset. The report outlines the challenges in identifying which psoriasis patients are likely to develop psoriatic arthritis, emphasizing the importance of early diagnosis and intervention. The study analyzes a subset of a large multi-source dataset, identifying two cohorts: the Positive Cohort, consisting of patients with both PsO and subsequent PsA diagnoses, and the Non-Positive Cohort, comprising patients with PsO but no PsA diagnoses. An AI-based digital phenotyping tool was utilized to isolate a phenotype associated with PsO-to-PsA progression. The findings indicate strong predictive performance of the tool, with a notable area under the receiver-operating characteristics curve (AUC) of 0.78. The report concludes that this approach can significantly enhance the identification of high-risk patients, potentially leading to improved clinical outcomes.