Development of a Phenotype-based Predictive Analytic for Acute Myeloid Leukemia

Project Summary The 5-year overall survival from acute myeloid leukemia (AML) is less than 30%. While some patients are cured with initial induction therapy, most patients relapse, and the expected outcomes in patients with relapsed and refractory (R/R) AML are dismal. For this reason, developing methods to identify therapies likely to benefit R/R AML patients is a top priority. This proposal aims to address critical unmet needs with a unique Academic- Industry Partnership (AIP) between investigators at Vanderbilt University Medical Center, Karyopharm Therapeutics, and Notable Labs. The BCL2 inhibitor, venetoclax (VEN), is transforming clinical practice for AML, but activity in R/R AML is less pronounced, and resistance occurs in most patients. AIP investigators currenty lead a multi-site investigator-sponsored study, testing VEN in combination with the selective inhibitor of nuclear export (SINE) selinexor (SEL) in a Phase I trial for R/R AML (NCT03955783). This SEL/VEN trial grew from the discovery that SEL synergizes with VEN and overcomes resistance mechanisms in some VEN-insensitive patient samples. Given this, our AIP team has worked together to develop a precision medicine functional platform for this novel combination in R/R AML. Notable Labs utilizes an automated high-throughput, immunophenotype-based flow cytometry method to provide real time drug sensitivity data on multiple, specific cell populations simultaneously within a given patient sample. Building from the only annotated cohort of patient samples treated with SEL/VEN in the world, we propose to develop a precision medicine functional assay to identify R/R AML patients most likely to benefit from SINE/VEN combination therapy. We will build, refine and optimize the functional platform for SEL/VEN with samples from our current Phase I study and train the platform on samples from the Phase 2 SEL/VEN clinical trial proposed by the AIP. Aim 1 will focus on determining assay parameters specifically for SEL/VEN in R/R AML. Aim 2 will train the model with the phase II clinical trial of SEL/VEN in R/R AML, and Aim 3 will contextualize the predictive analytic model on heterogenous genotypes found in R/R AML. At the conclusion of this study, the functional medicine platform will be a companion diagnostic ready for external validation which we will lead in a phase III efficacy trial of SEL/VEN in R/ R AML, and serve as proof-of-principle for development of similar therapy-specifc precision medicine tools.