Spatial and Deep Learning Approaches to Predict Early Progression in Smoldering Multiple Myeloma

Description

Smoldering multiple myeloma (SMM) represents a common precursor stage to active multiple myeloma (MM), yet accurately predicting early progression in SMM remains a challenge. Traditional risk models, such as the Mayo 20-2-20 system rely on disease burden parameters and have been validated to predict the risk of progression to active myeloma but fail to capture underlying biological heterogeneity. Multiomic analyses have identified distinct clonal evolution patterns and metabolic
reprogramming in SMM, yet the role of spatial immune architecture in the bone marrow microenvironment and its impact on SMM progression risk remains unexplored. This study aims to integrate spatial and deep learning approaches to predict early progression in SMM. The findings from this study have the potential to advance precision medicine approaches in the care of patients with SMM.