Disease burden by NGS in bone marrow and peripheral blood samples at baseline and after frontline therapy: Subgroup analysis of the advance randomized multi-center study of carfilzomib, lenalidomide and dexamethasone (KRd) with or without daratumumab

Abstract Introduction: ADVANCE (NCT04268498) is a fully enrolled, ongoing trial for patients with newly diagnosed multiple myeloma (NDMM). Patients were randomly assigned to receive 8 cycles of carfilzomib-lenalidomide-dexamethasone (KRd) with or without daratumumab (D). Patients who were minimal residual disease (MRD) positive at 10-5 after 8 cycles had the option to proceed to transplant. All patients transitioned to lenalidomide maintenance. The primary endpoint was MRD negativity (10-5) by next-generation sequencing (NGS) after up to 8 cycles of combination therapy. Biologically, MM is a disease of the bone marrow (BM), but interest in the potential utility of peripheral blood (PB) for monitoring disease burden is growing, and the prognostic value of circulating disease in blood has been established. We were motivated to compare the disease burden by NGS at 10-6 in bone marrow and peripheral blood samples at baseline and after frontline therapy. Methods: BM and PB samples were analyzed by NGS at 10-6 sensitivity (clonoSEQ, Adaptive Biotechnologies, Seattle) at baseline and at complete response (CR) or after cycle 8 (C8). For these analyses, CR and C8 were considered equivalent timepoints. Numbers of clonotypic sequences per million total cells in PB samples were compared with those from respective BM samples to assess differences in disease burden between the compartments. Results: Of the 306 patients enrolled, 71 patients had paired BM and PB samples sent for NGS at baseline and CR/C8. At baseline, all BM and PB samples were positive by clonoSEQ. In the BM, baseline median disease burden median was 41,400 and average was 116,200 sequences per million cells (range, 2,014 to 618,000). The median log difference between BM and PB disease burden was 2.54 logs (range, -0.32 to 4.9), with BM demonstrating greater values. In the primary analysis at 10-5 sensitivity, CR/C8 73% were MRD negative. In the current analysis at 10-6 sensitivity, CR/C8, 22/71 (31%) BM samples and 58/71 (82%) PB samples were MRD neg. Twenty-one of the 71 (30%) sample pairs were concordantly negative. Of the samples that were MRD pos, the median disease burden in the BM was 7.9 and average was 876 (n=49; range, 0.15 to 22,400) sequences per million cells while the median disease burden in the PB was 2.9 and average was 14.5 (n=13; range, 0.21 to 83) sequences per million cells. Twelve of the 71 (17%) BM/PB sample pairs were positive in both compartments. The median log difference in disease burden between BM and PB in these samples was 1.9 logs (range, 0.18 to 3.4). Ongoing analysis comparing levels of PB MRD at either timepoint to other clinical measures such as disease characteristics, serum biomarkers and clinical response, and will be presented at the meeting. Conclusions: Assessment of disease burden in PB may provide valuable information relative to other standard of care blood tests (such as SPEP, IFE and sFLC assays) and in lieu of frequent BM monitoring in MM. However, the disparate assessment of disease burden between PB and BM NGS testing indicate that, at least in the context of NDMM, PB negativity should not be used as a surrogate for establishing MRD status given the lower disease burden in the blood. Longitudinal samples with clinical follow up data are needed to better define the value of PB-based NGS testing.