P187 Prediction of flares in psoriatic arthritis using a machine learning approach to improve patient care

Le résumé fourni par la source

Abstract Background/Aims Treatment of psoriatic arthritis (PsA) involves the management of flares which can be unpredictable. Patients with flares may have long-term joint damage if it is recurrent or not promptly treated. Artificial intelligence (AI) and natural language processing (NLP) methods provide algorithms which may support early identification of patients with a risk of having a flare. In previous work, we presented an approach using NLP to identify patients who may have rheumatic disease. With the same aim, we extended the panel of predictors by combining clinical and laboratory data to predict flares in PsA. Methods We studied 191 patients with PsA and all patients met CASPAR criteria. There are 94 patients who had flares and 97 patients who did not have flares. We utilized patient blood test results, demographic information, electronic patient-reported outcomes scores, comorbidity, weight and height. We developed a time series dataset and applied Long Gradient Boosting Machine (LGBM) framework and neural networks (NN) to forecast the risk of flare before their future clinics. Results We conducted predictive modelling to forecast flare and non-flare events using data collected prior to patients’ upcoming clinic appointments. The dataset was split into training, validation and testing sets based on observations taken within 3, 6, 9 and 12 months before the clinic visit. Our LGBM model achieved 71.0% accuracy, 84% sensitivity, 57% specificity, 70% precision, 79% Area Under the Receiver Operating Characteristic Curve (AUROC) predicting flare and non-flare events within 12 months before the clinics. The results are summarised in Table 1. Conclusion Machine learning (ML) analysis of the combination of clinical and laboratory data shows promising results for the prediction of flares in PsA and may in the future assist clinicians in seeing the right patient in the right clinic. A study is planned not only to confirm the results but also to refine the ML models developed and specify disease attributes. Flares have significant impact on PsA patients and experiencing it adversely affects quality of life, disability and work productivity. The results of this study aims to enable early intervention to prevent flares of PsA in the future. Disclosure P. Moon: None. W. Li: None. E. Bazuaye: None. A. Chan: Honoraria; Novartis, Amgen, UCB, Lilly. Member of speakers’ bureau; Novartis, Jenssen, UCB, Amgen.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.