(PA-186) Deciphering Genomic Correlates of Differential Treatment Response Kinetics in Multiple Myeloma

Introduction: Multiple myeloma (MM) is an incurable blood cancer often marked by treatment resistance. Emerging evidence suggests that differential treatment response kinetics (DTRK) – the rate of patients' response – can predict outcomes, with slower responders faring better. This study investigates the prognostic value of DTRK and its underlying genetic drivers.

Methods: Patients were classified as fast (≤4 months) or slow ( >4 months) responders based on time to achieve a partial response or better during first-line therapy, consistent with clinical investigations into DTRK. Clinical and genomic data were used from CoMMpass, and an independent clinical dataset was used to validate the clinical outcomes. Kaplan-Meier survival analysis and multivariate Cox regression, adjusting for age, sex, ISS stage, cytogenetics, drug classes (i.e., IMID, PI, PI+IMID), an upfront autologous transplant, and response depth, were used to compare progression-free survival (PFS) and overall survival (OS) between response groups. To investigate the molecular correlates of DTRK, a robust bioinformatics pipeline was used to analyze mutation, structural variant, copy number alteration, and gene expression data. Machine learning based analysis of single-cell RNA sequencing data was performed to characterize the tumor microenvironment.

Results: Patients were classified into fast (n=284) and slow (n=425) responders and their clinical outcomes were compared. Analysis shows that slow responders have significantly longer PFS and OS compared to fast responders (HR=0.68, 95% CI 0.52-0.89, p=0.005; and HR=0.66, 95% CI 0.50-0.89, p=0.006, respectively). These results were independently validated in a separate real-world clinical cohort. Genomic analysis identified alterations in myeloma driver genes; such as fast responders exhibiting a higher frequency of TP53 aberrations (Chi-squared test, p value < 0.05), which may contribute to worse outcomes. Pathway analysis revealed upregulation of stemness-associated and lineage differentiation pathways in fast responders, suggesting a more aggressive tumor phenotype. Machine learning based exploration of immune microenvironment identified that that slow responders were enriched for natural killer cells and CD4⁺ and CD8⁺ T cells, indicating distinct immune landscapes.

Conclusions: Our findings demonstrate that slower treatment response is associated with significantly improved prognosis in multiple myeloma, independent of covariates. Multi-omics analyses identify DTRK to be associated with differences in tumor stemness, differentiation status, and immune microenvironment composition. These findings underscore the clinical relevance of response kinetics as a biomarker and provide insights into the underlying mechanisms, which may inform more personalized therapeutic strategies in MM.