Normalization of Serum B-Cell Maturation Antigen Levels from Treatment Predicts Progression Free and Overall Survival in Multiple Myeloma Patients

Introduction: B-cell maturation antigen (BCMA) is increased in the serum (s) of multiple myeloma (MM) patients (pts). We have previously shown that baseline sBCMA levels predict outcomes for MM pts. The purpose of this study was to determine whether a decrease of sBCMA to normal levels (< 82.69 ng/mL) after initiating treatment predicts both progression free survival (PFS) and overall survival (OS) among MM pts and its relationship to clinical response status in this pt population.

Methods: sBCMA levels were determined using an ELISA (R&D Systems; Minneapolis, MN) in 147 consecutive MM pts whose first treatment was in the frontline (n=87 [59%]) or salvage (n=60 [41%]) settings in a single clinic specializing in MM from February 2009 to April 2019 (observation cut-off: May 31, 2019). We determined sBCMA weekly during the first cycle and then monthly thereafter (median follow-up= 28 mo). An age-matched analysis of 206 healthy donors was used to determine a median normal sBCMA level of 37.6 ng/mL (SD 22.5 ng/mL). A reference threshold value of 82.7 ng/mL was determined using 2 SDs above the median. We defined normalization of sBCMA as a decrease in levels to < 82.7 ng/mL on two consecutive measurements. Kaplan-Meier analysis was used to compare PFS and OS among these pts and compared with responses as defined using the IMWG criteria. All pt samples were obtained following informed consent in accordance with the Declaration of Helsinki.

Results: One hundred thirteen pts (77%) had a baseline sBCMA ≥ upper threshold of normal (82.7 ng/mL) with a median of 435.3 ng/mL (range, 84.5-9,153.8 ng/mL) whereas the remaining 34 pts (23%) had baseline levels below 82.7 ng/mL and were excluded from analyses in relationship to normalizing sBCMA during their treatment. sBCMA levels normalized (< 82.7 ng/mL) in over half (55%) of evaluable pts (n=62), and these pts showed a markedly longer PFS (median 33.2 mo) than pts that did not (n=51; median 4.0 mo; p<0.0001). Furthermore, those normalizing their sBCMA levels showed improved OS (p<0.0001). Among those evaluated during their frontline treatment (n=75), those who normalized (n=50) also showed a longer PFS (median 33.9 mo) than those who did not (n=25; median 12.8 mo; p<0.0001). Only 12 of 38 pts undergoing salvage treatment normalized their sBCMA, but these pts experienced both improved PFS (median 8.3 mo vs 3.1 mo; p=0.0115) and OS (p=0.0345).

We also compared the PFS and OS of pts in CR to normalization of sBCMA. Pts whose sBCMA normalized (n=62) had both similar PFS (p=0.6257) and OS (p=0.7346) to those who achieved CR (n=26). Notably, every pt who achieved CR had normalized sBCMA and all pts who failed to normalize did not achieve CR (n=51). Pts who failed to normalize had shorter PFS (median 4.0 mo; n=51) than those who did not achieve CR (median 12.7 mo; n=87; p=0.0091). Among the pts who normalized sBCMA, pts who achieved CR (n=26) had similar PFS (p=0.4029) and OS (p=0.6354) to those who did not achieve CR (n=36). Conversely, among the 87 pts who did not achieve CR, pts who normalized their sBCMA (n=36) had markedly longer PFS (median 29.4 mo) than those who did not (n=51; median 4.0 mo; p<0.0001); normalizing pts also showed improved OS (p=0.0072).

We examined pts whose best clinical response was ≥ MR to determine if normalization of sBCMA improved upon determination of their PFS and OS. We found that among pts who achieved PR, those who normalized sBCMA (n=26) experienced improved PFS (median 33.2 mo) than those who did not (n=18; median 12.8 mo; p=0.0173); OS was also improved in those PR pts who normalized their sBCMA (p=0.0350). Similar findings were found when analyzing those achieving MR or PR, with PFS (p=0.0006) and OS (p=0.0326) for pts who normalized (n=30) being longer than for pts who did not normalize sBCMA (n=28). Of those achieving at least an MR, both PFS (median 33.9 vs 12.8 mo; p<0.0001) and OS (p=0.0052) were longer for normalizing pts (n=59) than pts who failed to normalize (n=28).

Conclusion: We have demonstrated that among MM pts starting new treatment with baseline elevated sBCMA levels, normalization of sBCMA levels (< 82.7 ng/mL) predicts markedly longer PFS and OS. Patients who achieve normalization have similar outcomes to those in CR. Additionally, our findings suggest that normalization of sBCMA predicts longer PFS and OS for pts who achieve ≥ MR. Therefore, normalization of sBCMA may be a novel approach to predict clinical outcomes for MM pts starting new treatment.