AB1528 APPLICATION OF BONE MARROW EDEMA SCORING SYSTEM DERIVED ARTIFICIAL INTELLIGENCE IN RAPID RADIOGRAPHIC PROGRESSION RESCUE STUDY

Background We demonstrated new bone marrow edema (BME) scoring system in X-ray (specificity 90%, sensitivity 65-70%) using deep learning with artificial intelligence (AI) reflects MRI BME previous EULAR meeting[1]. Objectives In this study, to ascertain whether score by film might reflect Biologics therapy, we retrospectively applied the rapid. radiographic progression rescue study previously reported[2]. Methods RA patients inadequate response MTX who have extensive single or bilateral wrist joints and DAS28-ESR>3.2 were treated enhanced group, of conventional synthetic (cs) DMARDs (26 patients) biologics (23 patients). film. two groups after 3-6 months treatment scoringsystem (cut off:0.4) compared BME. Results J-PEG file hand Xray uploaded system(Figure 1). are expressed as Positive Negative index [0-1]. number positive rate reduced from 65.3% 15.3% improvement group 78.5% 57.1% unchanged group(Table csDMARDs 75% 25% 59.7% 53.1% group. both groups, false negative (N: Figure 1, white color back) observed. Conclusion AI-derived may help early recognition efficacy during treatment. References [1]Katayama. K, ARD meeting supple AB 1335,2022 [2]Katayama. FRI 0124, 2018 Table 1. RED(P): positive, Black(N):BME negative, Before therapy (0M), (3/6M) Acknowledgements: NIL. Disclosure Interests None Declared.