Immune Signatures in Older Patients with Newly Diagnosed Multiple Myeloma are Associated with Survival Outcomes of First-Line Therapy Irrespective of Frailty Levels
The treatment landscape for older patients with multiple myeloma (MM) has rapidly evolved with the introduction of CD38-targeting antibodies. Yet, outcomes remain highly variable and are only partially explained by frailty status. Recent studies suggest that ‘immune fitness’ significantly contributes to the outcomes of immunotherapy in MM patients. Nonetheless, the immune system of older MM patients remains poorly studied. To address this, we investigated the impact of the immune system on survival outcomes of 89 newly-diagnosed MM patients in the HOVON-143 trial.
We conducted comprehensive immune profiling of peripheral blood (PB) and bone marrow (BM) samples from 89 patients enrolled in the HOVON-143 clinical trial. Computational data clustering was performed using the unsupervised clustering algorithm FlowSOM. Immune subset counts were compared between frail and intermediate-fit patients using the Mann-Whitney U test. Unsupervised clustering of Z-score standardized immune data was performed using k-means, with optimal clusters identified using Elbow and Silhouette methods. Associations with survival were assessed using Cox regression analyses. LASSO-regularized Cox models identified the most predictive immune subsets to construct the immune risk scores. Immune risk scores for PFS and OS were developed by multiplying the LASSO-selected predictor coefficients (β) by the z-score standardized counts (χ) of the corresponding immune subsets, according to the formula: immune risk score = Σ(βi*χi). To account for other predictor variables, the associations between immune risk scores and survival outcomes (PFS or OS) were evaluated in two-sided multivariable Cox regression analyses.
Comprehensive immunophenotyping of lymphoid and myeloid subsets as relative or absolute counts in peripheral blood (PB) and bone marrow (BM) revealed comparable immune composition between frail and intermediate-fit patients at diagnosis, except for reduced naive CD4+ and CD8+ T-cells and increased effector memory CD4+ T-cells and CD56bright NK-cells in frail patients.
Unsupervised clustering of T- and NK-cell subsets revealed that older patients with NDMM could be clustered into distinct immunotypes, which were associated with differential survival outcomes. To further investigate which specific immune subsets were linked to superior survival, multivariable analyses revealed that higher absolute counts of naive CD8+ T-cells, CD38+CD4+ T-cells, and CD56dimCD57+ NK-cells and reduced EM CD8+ T-cells, were significantly associated with favorable PFS and OS.
Using the most predictive immune parameters, we subsequently developed two immune risk scores - one for PFS and one for OS - which remained strongly associated with survival after adjusting for frailty status, disease stage and cytogenetic risk.
Our findings underscore the importance of a composite analysis of the immune system and demonstrate the association of baseline immune parameters with survival outcomes of first-line therapy in non-fit MM patients.