Ravi, Gayathri

Background The International Myeloma Working Group (IMWG) defines progression of disease (PD) in multiple myeloma by the increase in serum or urine paraprotein beyond a certain threshold. With more common use of serial bone marrow-based measurable residual disease (MRD) assessment, some patients will evolve with increase in MRD burden without meeting IMWG criteria for PD and no consensus exist on the best approach for this circumstance. Understanding the implication of increasing MRD burden is particularly relevant in patients with NDMM treated with quadruplet therapy (QUAD, consisting of a PI, an IMiD, an anti-CD38 mAb and dexamethasone) and autologous stem cell transplantation (ASCT), since most will reach MRD negativity. Methods We analyzed a large dataset of patients with NDMM patients treated with QUADs + ASCT +/- QUAD consolidation +/- single or dual-agent maintenance and monitored with serial MRD assessment by next-generation sequencing (clonoSEQ®) at end of induction, post ASCT, post consolidation and at least yearly thereafter. The dataset was built from clinical trial participants and from patients receiving standard care under an institutional guideline. MRD progression (MRD-P) was arbitrarily defined as quantitative increase in MRD burden by at least 1 log10 above the nadir. Treatment modification in the event of MRD-P without IMWG-defined PD was at discretion of treating physician. For patients with MRD-P without concomitant PD we describe cumulative incidence of PD after MRD-P. To contextualize the clinical implications of MRD-P, we describe survival free of failure of second line therapy (SFF2T), defined as time from progression event (MRD-P or PD not preceded by MRD-P) until occurrence of IMWG-defined PD or death on second line therapy, regardless of whether second line was initiated due to MRD-P or subsequent IMWG-defined PD. Results Among the 216 patients, 113 (52%) were clinical trial participants, median age was 61 years (IQR 56-68), 21% age 70 or older, 32% were non-White, 57% male, 38% had MM with 1 high risk chromosome abnormality [HRCA, gain/amp 1q, t(4;14), t(14:16), del(17p)], 15% had 2+HRCA. Median duration of follow up is 39.5 months, and 966 MRD datapoints were assessed. Overall, 83% of patients achieved sCR, 78% achieved MRD negativity at 10-5, 68% at the 10-6 threshold, and 56% ceased therapy after confirmed MRD negativity. Forty-nine patients had MRD-P(N=19) or PD not preceded by MRD-P (N=30). Patients with MRD-P or PD not preceded by MRD-P often had 1 or more HRCA (73.7% and 82.8%), International staging system III (36.8% and 30.0%), and high LDH at diagnosis (31.6% and 23.3%). Median time from onset of therapy to MRD-P was 24.4 mo (IQR 12.6-35.8) and to PD not preceded by MRD-P was 23.2 mo (IQR 18.4-29.4). Among the 19 patients with MRD-P, 12 (63%) were on observation at time of MRD-P, 2 were receiving single agent lenalidomide, 5 were receiving a QUAD, 17 (89%) had previously obtained MRD negativity at <10-5 including 10 (53%) who had also achieved MRD negativity at 10-6. Upon detection of MRD-P, 12 (63%) pts changed therapy, all to a monoclonal antibody-containing triplet or quadruplet regimen. Yet, median time from MRD-P to PD was 10.1 months. Patients with MRD-P had 1-year SFF2T of 55.6%, in contrast to 35.0% for patients with PD not preceded by MRD-P. OS 2-year from progression event was 78.0% for MRD-P and 53.9% for PD not preceded by MRD-P. Conclusions In the context of NDMM treated with QUAD + ASCT, MRD-P indicates imminent IMWG-defined PD, and is driven by a thriving plasma cell population unlikely to be controlled by reinstitution of anti-CD38 mAb-based combinations. Outcomes of patients with MRD-P are similar to those of patients with IMWG-defined PD. These findings challenge the paradigm of IMWG-defined PD and paraprotein measurability as minimal parameters for change in therapy and introduction of agents with new mechanisms of action.

Introduction QUADs with an anti-CD38 mAb, an IMiD, a proteasome inhibitor and dexamethasone followed by autologous stem cell transplantation (ASCT) in the treatment of newly diagnosed multiple myeloma (NDMM) led to substantial improvements in depth of response, including minimal residual disease (MRD) negativity, and progression-free survival. The therapeutic success of QUAD + ASCT has reinvigorated an interest in fixed-duration therapy, yet optimal short term efficacy endpoint for treatment cessation is unknown. We used a large dataset of patients (pts) treated with QUAD + ASCT, monitored with serial MRD assessment and managed with MRD-adapted treatment duration to determine the best short term efficacy endpoint to identify pts at low risk of treatment failure. Methods We analyzed 221 pts with NDMM treated with QUAD + ASCT and monitored with serial marrow-based MRD by next-generation sequencing. Based on trial design (N=116) and institutional practice (N=105) most pts received post ASCT QUAD consolidation followed by treatment cessation if MRD<10-5 in two consecutive datapoints. Pts were monitored with MRD post induction, ASCT and consolidation, and at least yearly thereafter. We utilized Cox regression to build parallel models of predictors of disease progression or MRD resurgence utilizing the same patient and disease features and one distinct short term efficacy endpoint per model, inputted as time-varying covariate. Efficacy endpoints tested were IMWG-defined stringent complete response (sCR), MRD<10-5 (single datapoint), MRD<10-6, sustained MRD (S-MRD, two consecutive assessments at least 1 year apart) <10-5and S-MRD<10-6. Best fitting models were determined using Akaike's information criterion (AIC, lower value indicates better fitting model), an information theory approach, and Heagerty & Zheng C index. Results Median age of the 221 pts. was 61 years (IQR 55-68), 21% age 70 or older, 37% had MM with 1 high risk chromosome abnormality [HRCA, gain/amp 1q, t(4;14), t(14:16), del(17p)], 16% had 2+HRCA, 30% were non-White and 58% male. We followed pts for a median of 40.6 mo. and performed 979 MRD assessments. One hundred and eighty-three pts (83%) achieved sCR, 170(77%) MRD<10-5, 146 (66%) MRD<10-6 , 110 (50%) S-MRD<10-5 and 90 (41%) S-MRD<10-6. Average time to achieve the endpoint was 7.8 mo for sCR, 7.1 mo. for MRD <10-5, 9.4 mo. for MRD<10-6, 19.3 mo. for S-MRD<10-5 and 20.5 mo. for S-MRD<10-6. We first tested models to predict PFS among all 221 pts. Predictive models included same covariates (ISS, high LDH and number of HRCA) other than the efficacy endpoint. The best fitting model (AIC 417.2, C-stat 0.757) was based on S-MRD<10-5 (HR=0.23, 95% C.I. 0.11-0.47) and included presence of 1 HRCA (HR=2.93, 95% C.I. 1.31-6.52), 2+ HRCA (HR=7.60, 95% C.I. 3.31-17.43), high LDH (HR=1.64, 95% C.I. 0.81-3.33), ISS2 (HR 1.31, 95% C.I. 0.62-2.77) and ISS3 (HR 1.96, 95% C.I. 0.90-4.28). The model with S-MRD<10-6 had similar, but not better performance (AIC 419.4, C-stats 0.752). We subsequently analyzed the 121 (55%) pts who reached MRD-guided treatment cessation for the best fitting model to predict progression or MRD resurgence in absence of continuous therapy. Among these pts, 117(97%) achieved sCR, all (by definition) MRD<10-5, 116 (96%) MRD<10-6 , 95 (79%) S-MRD<10-5 and 81 (67%) S-MRD<10-6. The best fitting model (AIC 197.2, C-stat 0.766) was based on S-MRD<10-5 (HR 0.01, 95% C.I. 0.002-0.05) and included presence of 1 HRCA (HR=1.41, 95% C.I. 0.49-4.07), 2+ HRCA (HR=5.43, 95% C.I. 1.75-16.84), high LDH (HR=3.19, 95% C.I. 0.98-10.40), ISS2 (HR 1.25, 95% C.I. 0.53-2.96) and ISS3 (HR 0.73, 95% C.I. 0.23-2.37). The model with S-MRD<10-6 did not perform better (AIC 232.4, C-stat 0.727). Among the 94 pts who achieved S-MRD<10-5 and were monitored off therapy, the risk of MRD resurgence or progression 2.5 years after cessation of therapy was 6.1%, 15.0% and 22.2% and risk of progression was 3.0 %, 12.0% and 11.1% for patients with 0, 1 and 2+ HRCA, respectively Conclusion In the setting of QUAD therapy + ASCT, no short term conventional or MRD-based efficacy parameter negates the cumulative impact of HRCA on prognosis. S-MRD<10-5 is the best predictor of PFS and yields the best predictive models for risk of MRD resurgence or progression in the setting of fixed duration therapy. This analysis validates S-MRD<10-5 as efficacy parameter for treatment cessation in trials exploring fixed duration therapy in NDMM.