Comparative Analysis of sSLAMF7, sBCMA, and M-Protein As Prognostic, Predictive, and Pharmacodynamic Biomarkers in Relapsed/Refractory Multiple Myeloma Treated with Pomalidomide and Dexamethasone +/- Elotuzumab

Despite advances in therapy, multiple myeloma (MM) remains largely incurable, and therefore identification of predictive or prognostic biomarkers is important for optimal treatment (tx) selection and detection of early progression. Elotuzumab (Elo) is an IgG1 monoclonal antibody that targets SLAMF7 on MM cells and is approved for relapsed and/or refractory (RR) MM in combination with lenalidomide or pomalidomide and dexamethasone (Pd). BCMA, a cell surface protein with expression fairly restricted to MM cells, is being investigated in clinical studies for multiple targeting strategies.

Soluble forms of SLAMF7 (sSLAMF7) and BCMA (sBCMA) are elevated in serum from patients (pts) with MM and are proposed surrogate markers of tumor burden. Serum monoclonal protein (M-prt) is the standard marker to assess tumor burden and monitor response and progression in pts with MM. However, the variable and slow rate of M-prt turnover may complicate early detection of relapse or response to therapy. Previously, we demonstrated that Elo-containing regimens show a more pronounced decrease in free sSLAMF7 compared with non-Elo regimens (Postelnek J et al. Blood 2015). Changes in sBCMA have not been reported for pts receiving Elo-based regimens in clinical trials, but high levels of sBCMA have been associated with poor outcomes for pts with MM receiving a variety of tx. This is the first comparative analysis of these serum markers in a randomized trial to assess their potential utility for predicting and monitoring response to Elo-containing regimens.

Methods:

Of the 117 pts randomized in ELOQUENT-3 (NCT02654132), serum samples were available for 106 (sSLAMF7) and 51 (sBCMA) pts (further data on sBCMA will be generated). Concentrations of sSLAMF7 and sBCMA were measured by ELISA at Cycle [C] 1, Day [D] 1, 8, 15, and 22; C2D1, and C3D1. Serum M-prt was measured by standardized SPEP at C1D1, C2D1, and C3D1 at a central laboratory and was available for all pts. Associations between baseline (BL) sSLAMF7/sBCMA and prognostic factors at study entry (ISS, number of prior therapies, MM risk category; assessed using Kruskal-Wallis rank sum tests) or efficacy in each study arm (responder vs non; median PFS) were assessed. Hazard ratios (HR) for PFS were calculated using univariate Cox models and p-values were calculated using log-rank tests. Linear mixed effect models were used to explore dynamics of sSLAMF7 and sBCMA relative to serum M-prt over time for responders and non-responders within each tx arm.

Results:

At BL, higher sSLAMF7 and sBCMA levels were associated with high-risk disease; higher BL sBCMA levels were also associated with ISS stage III disease. Among pts treated with Pd who had higher BL sSLAMF7 or sBCMA levels there was a trend towards non-response and poorer PFS. No association between BL sSLAMF7 and EPd tx efficacy was observed; however, EPd treated pts with higher BL sBCMA tended to have poorer PFS. Regardless of BL levels of sSLAMF7, PFS was longer with EPd than Pd; this difference was even more substantive in pts with elevated sSLAMF7 (≥median; HR=0.415, p=0.006). Concentration of M-prt at BL did not appear to be associated with response to either tx, but high levels did show a trend for poorer PFS in the Pd arm with no difference observed in the EPd arm.

Reduction of both free sSLAMF7 and sBCMA from BL was observed as early as C1D8 in both EPd and Pd arms; the reduction in free sSLAMF7 was more pronounced in the EPd arm than the Pd arm. Greater reductions of sSLAMF7 and sBCMA distinguished responders from non-responders no later than the beginning of the second cycle. Decreases in M-prt did not distinguish responders from non-responders in either arm until Cycle 3, and the reduction was smaller than either sSLAMF7 or sBCMA.

Summary:

Data from pts enrolled in ELOQUENT-3 show that high BL levels of sBCMA were associated with a trend to poorer PFS in both the Pd and EPd arm. High BL sSLAMF7 levels were also associated with poorer PFS in Pd arm, but not in the EPd arm. A trend toward lower response rate in pts with high BL levels of sSLAMF7 or sBCMA was identified in the Pd arm, but no association was seen in pts receiving EPd. The dynamics of sBCMA and sSLAMF7 revealed an earlier decrease than for serum M-prt, separating responders from non-responders as soon as C1D22. sSLAMF7 and sBCMA may therefore represent a more sensitive parameter for monitoring response to tx in MM, however additional studies are needed to better understand this application.