Biomarker-Driven Phase Ib Clinical Trial of OPB-111077 in Acute Myeloid Leukemia Increases Overall Response Rates

Background: Phase I clinical trial for oncology drugs have the highest clinical failure rate. Drug candidates may be discarded if they don´t show activity. Yet these drugs may have valuable activity in patient subgroups. Biomarkers attempt to identify these sensitive subpopulations, but are normally evaluated and validated in Phase II when good drugs may have been discarded.

OPB-111077 is a novel, oral, low-molecular-weight compound that inhibits STAT3 and mitochondrial electron transport. A prior Phase I trial on 145 patients from all tumor types gave only 1 partial response (PR). A strong anti-proliferative activity was identified in Acute Myeloid Leukemia (AML) samples ex vivo. This new biomarker was used for patient selection in this trial.

We present a novel design of a Phase Ib trial directly selecting patients with a potential biomarker, that can benefit the patient without increasing toxicity risks, in Relapsed/Refractory (RR) AML patients.

Method: This is an open-label, phase Ib dose-escalation clinical trial to evaluate the safety profile, maximum tolerated dose (MTD) and preliminary efficacy of oral OPB-111077 in AML relapsed or refractory to chemotherapy patients. Patients > 18 years old with high risk AML, ECOG ≤2 and adequate organ functions were selected for biomarker screening. Biomarker consisted on a dose response curve for antiproliferative activity of agent in a bone marrow sample of patients incubated 72-96-120h (Figure), quantified as the Area Under the Curve (AUC). Patient samples whose ex vivo biomarker showed low sensitivity (70% worst) we excluded, since they were potentially predicted as resistant to the agent. The remaining 30% with higher sensitivity were included. OPB-111077 was administered orally on a once daily dose schedule in 28-day cycles until intolerable toxicity or disease progression. Two dose schemas were evaluated: 200 (DL1) and 250 mg/day (DL2). Dose limiting toxicities were any Grade ≥ 3 non-hematologic toxicity and any unexpected non-tolerable grade II that requires delay beyond 1 week until recovery and evaluated during the first 28 days of treatment. Adverse events were graded according to NCI-CTCAE vs 4.03 and response criteria was based on those given by Cheson et al. IC50 and area under the curve (AUC) determined by the Vivia ex vivo biomarker were correlated with clinical response.

Results: We screened 47 RR AML patients, twelve patients were included and selected by personalized medicine test (5 in DL1 and 7 in DL2), median age 76 years, 92% men, 42% ECOG 0 were included in the study. AML patients were refractory in 41.7% and patients were in median of relapses 2 (range 1-6). No DLT was reported. Adverse events related to study treatment were reported in three patients, all G 1-2: in DL1 one patient had vomiting; in DL2 one patient had extrasystoles and other had anorexia, diarrhoea, epigastric discomfort, nausea and vomiting. Two patients died during study treatment due to disease progression and respiratory infection. Half of the patients discontinued study treatment due to disease progression and 25% due to adverse events (respiratory failure G5, respiratory infection G5 and extrasystoles G2). Over the 6 patients evaluable for efficacy, 3 of them achieved PR and 3 treatment failure as best response. Progression free survival and overall survival were 57 days (95% CI: 37 - 77) and 95 days (95% CI: 27 - 163) respectively. The biomarker AUC and IC50 values stratified patients consistent with their clinical response, although without enough sampling potency for statistical significance. Figure shows the relative position of patients included in the study in the context of the overall population of samples considered to build up the mixed effect population pharmacodynamic model used to analyze samples results and define inclusion criteria. In green those that correspond to patients who showed a positive response; in orange those who failed.

Conclusions: OPB-111077 is generally well tolerated and has a manageable toxicity profile. The MTD was not reached. The 50% PR rate achieved by OPB-111077, albeit in a few patients, is substantially higher that the 0.7% (1/145) PR rate achieved in a prior phase I trial on all tumor types¹. This innovative Phase Ib design selecting patients using a biomarker may enable to rescue drug failures in the future.

1. Tolcher A, Flaherty K, Shapiro GI et al. Oncologist. 2018, 23(6):658-e72