Comparative analysis of next generation flow (NGF) and next generation sequencing based clonoseq for MRD detection in multiple myeloma using paired bone marrow samples

Abstract Background: Minimal residual disease (MRD) is a critical prognostic and therapeutic response marker in multiple myeloma (MM). Two main technologies are used for MRD assessment in bone marrow: next-generation flow cytometry (NGF) and next-generation sequencing (NGS), from clonoSEQ. While both methods are FDA-cleared and clinically relevant, they differ in sensitivity, sample handling requirements, and technical differences. Comparative performance in real-world matched samples is understudied. Here we compared the detection rates and concordance between MRD by NGF (flow cytometry) and NGS-based clonoSEQ, using paired bone marrow aspirates collected at the same clinical timepoint. Methods We identified 246 paired bone marrow samples from MM patients, which were evaluated by both flow cytometry (Euroflow, typically sensitivity of 10-5) and clonoSEQ (sensitivity of up to 10-6) on the same collection date.MRD-negativity for clonoSEQ was defined as no detectable clones and below the assay's level of detection (LoD). Continuous variables (clonoSEQ and flow CPS) were analyzed using the Mann-Whitney U test. Categorical variables were assessed using chi-square tests with Yates continuity correction for 2x2. A cox proportional hazards model was used to evaluate the association between overall survival and cytogenetic abnormalities, MRD status, and other clinical covariates including age, sex, treatment type, and disease setting (newly diagnosed, relapsed/refractory). A p-value < 0.05 was considered significant. Results Out of 246 paired bone marrow samples, we identified 173 concordant samples, out of which 26 were concordant positive (clonoSEQ+/NGF+), 147 were concordant negative (clonoSEQ-/NGF-), and 73 were discordant (clonoSEQ+/NGF-). clonoSEQ detected MRD in all cases missed by NGF, including some with relatively high tumor burden, reflecting its higher sensitivity. Among discordant samples, 45 (61.6%) were following treatment of newly diagnosed MM (NDMM) and 28 (38.4%) treatment of relapsed/refractory (RRMM). Among the MRD-positive samples, the concordant group (n=26) showed a mean clonoSEQ count of 11923.5 cells per million (CPM) and a mean NGF CPM of 305.5 (p<0.01). In contrast, the 73 discordant samples had a lower mean clonoSEQ CPM of 156.1. For the 147 MRD-negative samples with concordant results, no residual disease was detected by either assay (CPM = 0 for both). When evaluating sex, race, ethnicity, chromosomal abnormalities, prior treatments to sample collection (immunomodulatory drugs, proteasome inhibitors, CD38 antibodies, CAR Ts/bispecific antibodies), and disease stage (NDMM, RRMM, untreated), patients with t(11;14) were significantly higher in the discordant group (17.9% concordant vs. 32.9% discordant; p = 0.016), while CAR T/bispecific antibody-treated patients were more likely concordant for the tests (27.2% concordant vs. 4.1% discordant; p<0.01). High-risk cytogenetics (t(4;14), t(14;16), gain 1q) were associated with worse survival (HR 6.22, 95% CI 2.02 – 19.2; HR 16.2, 95% CI 3.38-77.1; HR 4.0, 95% CI 1.28 – 12.5, respectively). Discordant MRD status did not impact survival, though only 13/246 patients were deceased (median follow-up approximately 50.6 months).Conclusion: ClonoSEQ demonstrates higher sensitivity compared to NGF for MRD detection in MM, identifying residual disease missed by NGF particularly in t(11;14) cases. Discordance did not affect survival, however, only 13 patients were deceased at the time of last follow up, suggesting longer follow up is needed.