ASH 2021 | Predicting MRD outcomes in myeloma using machine learning

Bruno Paiva, PhD, University of Navarra, Pamplona, Spain, discusses the development of a machine learning model to predict undetectable measurable residual disease (MRD) in patients with newly diagnosed transplant-eligible multiple myeloma. Nearly 300 patients treated with a standard of care therapy were included in the study. MRD status after treatment was determined and associations with 37 clinical and biological parameters were evaluated. Of the 17 variables found to associate with MRD outcomes, the most effective model was created using cytogenetic, tumor burden, and immune related biomarkers. It was demonstrated that patients predicted to achieve undetectable MRD showed longer progression-free and overall survival compared to those who did not. The model has been made available online to enable its use in clinical practice. Using this model to individualize treatment decisions based on the probability of achieving undetectable MRD may represent a new method for personalized care in myeloma. This interview took place at the 63rd ASH Annual Meeting and Exposition congress in Atlanta, GA.