Multi-omics profiling and AI-driven clinically deployable risk models in MGUS and smoldering myeloma

Monoclonal gammopathy of undetermined significance (MGUS), smoldering multiple myeloma (SMM), and multiple myeloma (MM) form a continuum of plasma cell disorders, with progression from MGUS to MM being difficult to predict. Current risk stratification models, largely based on clinical, laboratory, and cytogenetic markers, fail to capture the molecular complexity underlying disease progression, limiting their predictive accuracy. Recent advancements in multi-omics technologies, encompassing genomics, transcriptomics, proteomics, and metabolomics, have provided deeper insights into the molecular drivers of these conditions. The integration of artificial intelligence (AI) and machine learning (ML) further enhances this understanding, offering new avenues for dynamic, personalized risk prediction. AI-based approaches that incorporate multi-omics data have the potential to identify novel biomarkers and predict disease outcomes with greater precision. These advancements could revolutionize risk stratification by providing a more individualized and dynamic framework for patient monitoring and treatment. However, the clinical adoption of AI and multi-omics tools is fraught with challenges, including the integration of complex data types, the need for standardized protocols, and concerns surrounding data privacy and algorithmic bias. Furthermore, evolving regulatory frameworks must accommodate the continuous learning capabilities of AI systems. This article explores the current limitations of risk stratification models in MGUS and SMM and examines the potential of multi-omics and AI to improve predictive accuracy. It also discusses the technical, ethical, and regulatory hurdles that must be overcome to enable the clinical implementation of these technologies, offering a roadmap for their future integration into patient care.