Abstract INTRODUCTION: Volasertib, a polo-like kinase 1 (PLK1) inhibitor, has a roughly 30% CR/CRi response rate in de novo acute myeloid leukemia (AML) patients when combined with low-dose cytarabine. However, its development was halted due in part to toxicity in a Phase 3 study. We have built a flow cytometry-based, high-throughput predictive precision medicine platform to forecast clinical responses to various treatments. Here, we present our design plan for a Phase 2a clinical trial with volasertib in relapsed/refractory (R/R) AML using a modified protocol to reduce toxicity and the development of a companion diagnostic to predict responders in a follow-up Phase 2b study. METHODS: We will initiate an open-label clinical trial in AML with volasertib in venetoclax-hypomethylating agent (HMA) R/R patients while co-developing a companion diagnostic to selectively enroll predicted responders in a follow-up, Phase 2b, study. Primary and secondary objectives will focus on efficacy/safety of intravenous volasertib and ex vivo responses, respectively. To reduce the toxicity seen in the Phase 3 study, specific strategies like tailored dosing will be implemented. To develop the clinical trial assay, peripheral blood or bone marrow samples were collected from AML patients and treated ex vivo with volasertib for 3 days. Surviving blast cells were enumerated by flow cytometry. Using a dynamic separation model restricted to the 35% most sensitive samples, we estimated optimal blast cutoffs for volasertib sensitivity. RESULTS: In single-agent experiments with 41 primary AML samples (31 de novo, 9 R/R, 1 with disease status unclear), we observed dose-response profiles consistent with volasertib’s role as a cell cycle inhibitor. Irrespective of the disease status, median and mean EC50 values were about 10 nM (1st-3rd quartile from 7-12 nM). To identify an optimal concentration of volasertib for patient stratification, we compared the ex vivo resistant blast fractions at each concentration with the areas under the dose-response curves (AUCs), assuming that in our case, AUCs most accurately reflect clinical responses. At 31.6 and 100 nM, the Pearson correlation coefficients were highest (r > 0.8), suggesting that these are optimal stratifying doses. With these concentrations and corresponding blast cutoff values, 32-33% of the de novo and 25% of the R/R patients would have been predicted to be responders were they given volasertib, assuming a positive predictive value of 100%. CONCLUSION: Volasertib’s EC50 values clustered within a narrow range, suggesting that EC50 may not be a suitable metric to stratify patients into responders and nonresponders, in contrast to AUCs and residual blast fractions at 31.6 and 100 nM of volasertib, which displayed more heterogeneous distributions. Limitations include small sample sizes, especially for the R/R group. Citation Format: Markus Daniel Lacher, Glenn Michelson, Allen Ling, Wade Anderson, Jeaneille Vinas, Christine J. Gu, Chris Leonardi, Joseph Wagner, Gabriel N. Mannis. Guided by a predictive ex vivo test: Bringing the PLK1 inhibitor volasertib back into the clinic for venetoclax-HMA relapsed/refractory acute myeloid leukemia patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5178.
Read full abstract