Development and Validation of a Prediction Model of Outcome after B-Cell Maturation Antigen-Directed Chimeric Antigen Receptor T-Cell Therapy in Relapsed/Refractory Multiple Myeloma

Although chimeric antigen receptor T therapy (CAR-T) cells are an established therapy for relapsed/refractory multiple myeloma (RRMM), there are no established models predicting outcome to identify patients who may benefit the most from CAR-T. PATIENTS AND METHODS This is an international retrospective observational study including patients with RRMM infused with currently available commercial or academically produced anti–B-cell maturation antigen (BCMA) CAR-T. We describe characteristics and outcomes in Europe (n 5 136) and the United States (n 5 133). Independent predictors of relapse/progression built a simple prediction model (Myeloma CAR-T Relapse [MyCARe] model) in the training cohort (Europe), which was externally validated (US cohort) and tested within patient- and treatment-specific subgroups. RESULTS The overall response rate was 87% and comparable between both cohorts, and complete responses were seen in 48% (Europe) and 49% (the United States). The median time to relapse was 5 months, and early relapse <5 months from infusion showed poor survival across cohorts, with the 12-month overall survival of 30% (Europe) and 14% (the United States). The presence of extramedullary disease or plasma cell leukemia, lenalidomide-refractoriness, high-risk cytogenetics, and increased ferritin at the time of lymphodepletion were independent predictors of early relapse or progression. Each factor received one point, forming the three-tiered MyCARe model: scores 0-1 (low risk), scores 2-3 (intermediate risk), and a score of 4 (high risk). The MyCARe model was significantly associated with distinct 5-month incidence of relapse/progression (P < .001): 7% for low-risk, 27% for intermediate-risk, and 53% for high-risk groups. The model was validated in the US cohort and maintained prognostic utility for response, survival, and outcomes across subgroups.