Citations

Abstract Abstract 1881 Background: Drug-disease modeling and simulation has recently been proposed to support drug development decisions in patients with liquid and solid tumors (Claret et al J Clin Oncol, 27:4103-4108, 2010; Wang et al. Clin Pharmacol Ther, 86:167-174, 2009). We hypothesized that similar modeling based on a decrease in M-protein may be applied in multiple myeloma. The objective was to use a modeling and simulation approach to leverage available lenalidomide multiple myeloma data and support pomalidomide development decisions by, 1) developing a drug-independent link between tumor burden reduction (as assessed by change from baseline in serum M-Protein) and survival and PFS and 2) simulating expected survival and PFS based on interim M-protein data of the phase 1/2 trial (CC-4047-MM02, NCT00833833) of pomalidomide (POM) in patients with relapsed and refractory multiple myeloma who have received at least 2 prior therapies including lenalidomide (LEN) and bortezomib. Methods: M-Protein measurements were modeled as a function of time from 704 patients included in two phase 3 clinical studies of LEN plus dexamethasone (DEX) vs. DEX (MM010, Dimopoulos, NEJM 357, 2123–2132, 2007 and MM009, Weber, NEJM 357, 2133–2142, 2007). Models were developed for survival and PFS times as a function of change in M-protein level at end-of-cycle 2 (week 8) from baseline as well as other prognostic factors. Interim M-protein data from the ongoing MM-002 POM study (217 patients) were modeled to simulate clinical endpoints. Results: Based on lenalidomide Phase 3 data, week-8 change in M-Protein (p < 0.00001), ECOG performance status (p < 0.0009), baseline albumin, hemoglobin and creatinine levels (p < 0.01) were significant independent predictors of survival when week-8 change in M-Protein (p < 0.00001) and baseline hemoglobin (p < 0.001) were significant independent predictors of PFS. Observed survival and PFS distributions over 100 weeks in lenalidomide studies and difference between the two treatments (LEN + DEX vs. DEX) were consistent with the 95% confidence intervals (CI) of the models. Model predictions (95% PIs) of median survival and PFS based on week-8 change in M-Protein following treatment with POM and POM + DEX are presented in the table below. The recently disclosed observed medians and 95% CIs for all endpoints and all arms from MM-002 are included in the model. Conclusions: Change in M-protein level as a continuous longitudinal biomarker appears to be a potential indicator to assess drug effect in multiple myeloma studies using modeling and simulation based on early interim data. PFS predicted by the model and the actual PFS are similar and thus may support end-of-Phase 2 decision (Bruno et al Clin Pharmacol Ther 86:136-138, 2009). Current simulations indicate encouraging results for POM in a refractory multiple myeloma patient population who have received prior treatment that includes lenalidomide and bortezomib but these exploratory methods require confirmation as part of ongoing studies. Disclosures: Bruno: Pharsight: Employment; Celgene Corporation: Research Funding. Off Label Use: Pomalidomide is not an FDA licensed product for multiple myeloma treatment. Jonsson:Pharsight: Employment; Celgene Corporation: Research Funding. Zaki:Celgene Corporation: Employment. Jacques:Celgene Corporation: Employment. Swern:Celgene Corporation: Employment. Richardson:Millennium Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees; Celgene Corporation: Membership on an entity's Board of Directors or advisory committees; Novartis Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees; Johnson & Johnson: Membership on an entity's Board of Directors or advisory committees; Bristol Myers Squibb: Membership on an entity's Board of Directors or advisory committees. Claret:Celgene Corporation: Consultancy, Research Funding; Pharsight: Employment.

AbstractA population pharmacokinetic (PK) model and exposure–response (E‐R) analysis was developed using data collected from 5 phase 1b/2 and 2 phase 3 studies in subjects with multiple myeloma. Subjects receiving intravenous infusion on 2 consecutive days each week for 3 weeks (days 1, 2, 8, 9, 15, and 16) in each cycle at doses ranging from 15 to 20/56 mg/m2 (20 mg/m2 in cycle 1 and, if tolerated, escalated to 56 mg/m2 on day 8 of cycle 1). The population PK analysis indicated that among all the covariates tested, the only statistically significant covariate was body surface area on carfilzomib clearance; however, this covariate was unlikely to be clinically significant. Despite inclusion of different populations (relapsed or relapsed/refractory), treatments (carfilzomib monotherapy or combination therapy), infusion lengths (2 to 10 minutes or 30 minutes), and different doses, the E‐R analysis of efficacy showed that after adjusting for baseline characteristics, higher area under the concentration–time curve was associated with improved overall response rate (ORR), from 15 to 20/56 mg/m2. No positive relationships between maximum concentration and ORR were identified, indicating that ORR would not be expected to be impacted by infusion length. For safety end points, no statistically significant relationship between exposure and increasing risk of adverse events was identified. The results of an E‐R analysis provided strong support for a carfilzomib dose at 20/56 mg/m2 as a 30‐minute infusion for monotherapy and combination therapy. This article illustrates an example of application of E‐R analysis to support labeling dose recommendation in the absence of extensive clinical data.