Immune-MRD Status Informs Tumor-MRD Outcome Prognostication in Multiple Myeloma Patients on Lenalidomide Maintenance

Introduction: Measurable residual disease negativity (MRD-) is predictive in newly diagnosed (ND) multiple myeloma (MM). However, the clinical significance of MRD resurgence (MRD^res), where MRD- patients show re-emergence of MRD positivity (MRD+), is incompletely understood. Additionally, the significance of patient-specific immune profiling has not been thoroughly explored in MM patients with no evidence of measurable disease.

Methods: We prospectively followed 108 MM patients in a now completed phase 2, single arm study of 5 years of continuous lenalidomide (Len) maintenance following unrestricted frontline therapy. Dynamic MRD assessments were made via annual bone marrow (BM) biopsies (MRD- defined as 10^-5 by multicolor flow). Exploratory peripheral blood (PB) T cell profiling was performed via high-dimensional spectral cytometry using two ~40-color T cell focused panels inclusive of lineage, exhaustion, activation markers and intracellular transcription factors and granzymes. T cell profiling results were analyzed with dimensionality reduction analysis with UMAP and algorithm assisted cell clustering with Phenograph and were correlated to dynamic MRD status and clinical outcomes. Statistical analysis for immune profiling results was performed using unpaired t-tests with multiple comparisons correction via the Boneferroni method.

Results: In line with prior datasets, patients MRD- at the start of Len maintenance (N = 46) had superior progression free survival (PFS) compared to MRD+ patients (N = 62, P = 0.0021). MRD^res was observed in 11 patients; however, only 4/11 had subsequent progression of disease (POD), with a median post-MRD^res follow-up time of ~2 years. Patients with MRD^res had poorer outcomes than those with sustained MRD- status at both the 1-year (P = 0.036) and 2-year landmarks (P = 0.0014); however, there were no significant differences in PFS between MRD^res patients and MRD+ patients at either the 1-year (P = 0.68) or 2-year (P = 0.27) landmarks. PB T cell profiling experiments performed at the time of MRD^res (matched control timepoints in sustained MRD- patients) showed that patients with MRD^res but without subsequent POD had nearly identical T cell characteristics to those with sustained MRD- status. Patients with MRD^res and with subsequent POD within 18 months of MRD^res showed highly distinct T cell profiles enriched with T cell clusters corresponding to CD8⁺CD45RO⁺CCR7^- effector memory (T_EM, 2.5 fold increase, P = 0.003), and CD4⁺CD45RO⁺CCR7^- T_EM cells (1.6 fold increase, P = 0.004). This included enrichment for a subpopulation of CD8 T_EMs with high expression of Ki-67, CD38, granzyme K, and granzyme B (5.2 fold increase, P = 1.1 x10^-5). Patients with sustained remission had enrichment of clusters corresponding to CD4⁺CD45RA⁺CCR7⁺ naïve T cells (3 fold increase, P = 0.037), CD8⁺CD45RA⁺CCR7⁺ naïve T cells (5.4 fold increase, P = 0.020), and CD8⁺CD45RA⁺CCR7⁺CD31⁺CD103⁺ recent thymic emigrants (6.9 fold increase, P = 0.028). Early progressors were also noted to have global enrichment in expression of IRF4 (P = 7 x 10^-5) and PLZF (P = 0.0038) on CD4⁺ T cells as well as high Granzyme B (P = 0.0016) and T-bet (P = 7.4 x 10^-5) expression on CD8⁺ T cells. T cell profiles were independent of known prognostic factors in MM including ISS, cytogenetic risk, and treatment history. Longitudinal T cell profiling in these same patients showed that patients with sustained remission (regardless of dynamic MRD status) had unchanged T cell profiles over ~1 year. However, patients with disease progression showed progressive T cell exhaustion as patients approached the time of POD, characterized by decreased CD38 expression (P = 0.028), and a decreased number of CD8 T_EM cells with high expression of Ki-67, CD38, granzyme K, and granzyme B (53% decrease, P = 0.0029). These results were validated in a separate cohort of MRD+ patients, with patients with sustained MRD+ nonprogression having similar T cell profiles to patients with MRD^res but without subsequent POD.

Conclusion: MRD^res is a negative predictive sign in MM patients on Len. PB T cell profiling results show that an activated PB T cell profile, enriched with activated effectors, associated with early relapse despite low/undetectable levels of disease. This “Immune MRD” status was independent of known risk factors in MM and dynamic MRD status, indicating that it may represent a novel predictive biomarker in MRD- MM patients on Len maintenance.