0714 Machine learning model for Predicting phenoconversion in patients with Rem behavior disorder using clinical markers

Abstract Introduction Idiopathic Rapid eye movement (REM) sleep behavior disorder is a condition that can be an early sign of alpha-synuclein-mediated neurodegenerative diseases, and the course disease vary greatly from patient to patient. It important identify patients who are at risk developing diseases in future for purpose clinical trials plan their lives accordingly. Previous research has identified various factors phenoconversion RBD patients, but these studies not practical use settings due resource availability or rarity certain features. Additionally, most have been conducted on non-Asian populations, which may different genetic backgrounds than Asian populations. This study aimed develop machine learning model predict survival using features commonly available routine settings. Methods recruited diagnosed with based polysomnography results collected 34 each Missing data were imputed models applied improve performance. The model's predictive performance was evaluated integrated Brier score concordance index. Mean indicators calculated 5-fold cross-validation results. A web application hosting final prediction developed deployed server by physicians patients. Results 173 included study. We used likelihood ratio test calculate p-values all variables selected following 8 less 0.1: UPDRS part III, age, history antidepressant use, alcohol MoCA (Montreal Cognitive Assessment), PSQI-TST (Pittsburgh Sleep Quality Index - total time), AHI-REM (apnea-hypopnea index REM sleep), education level. random forest had best mean IBS 0.07 C-index 0.93 Conclusion showed it possible Support (if any)