PII-059 - MODELING M-PROTEIN DYNAMICS IN SUBJECTS WITH RELAPSED/REFRACTORY MULTIPLE MYELOMA RECEIVING VENETOCLAX: UTILITY OF EARLY M-PROTEIN MONITORING TO PREDICT PROGRESSION FREE SURVIVAL

S. Liu¹, A. Salem², J. Ross¹, O. Bueno³, M. Badawi⁴; ¹AbbVie, ²AbbVie Inc., ³AbbVie Inc.,, North Chicago, IL, USA, ⁴AbbVie Inc.

Background: Multiple myeloma (MM) is a plasma cell malignancy in which monoclonal plasma cells proliferate in the bone marrow, resulting in an overabundance of monoclonal paraprotein (M protein). M-protein in the blood/serum is a marker of tumor burden. Venetoclax is a BCL-2 inhibitor with significant efficacy in MM patients, specifically those harboring the t(11;14) chromosome translocation. Establishing and validating early predictors of survival in MM subjects is important in the clinical development of new agents. The objective of this study was to develop a model to describe M-protein dynamics in MM subjects receiving venetoclax and assess the utility of early M-protein measurements in predicting progression-free survival (PFS) and overall survival (OS).

Methods: M-protein data from MM subjects receiving venetoclax in a Phase 1/2 study (NCT01794520) were fit to a tumor growth inhibition (TGI) model. Key demographic and clinical characteristics were tested in covariate analyses. Individual M-protein profiles were simulated using the final TGI model and % M-protein changes at 3, 4, 5, and 6 weeks were predicted. Parametric survival models were then developed to assess the utility of M-protein changes as well as key demographics and clinical characteristics in predicting PFS and OS.

Results: Among the 117 subjects in the study, 63 (53.8%) had data that could be used to develop the M-protein model. A TGI model described by Kgrow, rate of M-protein increase (0.01 week-1); Kkill, rate of drug-induced M-protein decrease (0.173 week-1/mg/L); and λ, rate of disappearance of drug effect (0.691 week-1), adequately described the data. Body weight was identified as a covariate on Kgrow (0.0004) with patients of larger weight having a higher Kgrow. Additionally, t(11;14) status was identified as a covariate on Kkill and λ with t(11;14)-positive patients having a 3-fold higher Kkill and 71% lower λ than non-positive patients. Survival analysis indicated that % M-protein changes from baseline at weeks 3, 4, 5, and 6 were the only predictors of PFS, each decreasing the AIC by > 29 points. No significant covariates were identified for OS.

Conclusion: This model demonstrates that early changes in M-protein concentrations within the first 6 weeks of treatment could be predictive of PFS in MM subjects receiving venetoclax.