Abstract 1076: Plasma microRNAs associated with overall survival in patients with hepatocellular carcinoma treated with galunisertib (LY2157299 monohydrate), an inhibitor of transforming growth factor-β receptor1

Background: Galunisertib, a selective transforming growth factor-β receptor1 inhibitor, is being investigated in clinical trials for hepatocellular carcinoma (HCC). MicroRNAs (miRs) are small (∼22 nucleotide) non-coding RNAs that regulate expression of targeted genes, and are secreted by cells into blood. miRs are differentially expressed during HCC progression, differentiation, and response to therapy. We hypothesize that circulating miRs may be useful to identify patients who benefit from galunisertib treatment. Patients and Methods: Plasma samples from HCC patients (n = 105) treated with galunisertib were analyzed for miR expression. Patients were enrolled in part A, of a multi-part single-arm study in 2nd-line HCC (phase II trial NCT01246986/JBAK). All patients had elevated alpha-fetoprotein (AFP) levels at baseline (AFP ≥1.5 x ULN). The median OS of this cohort of patients was 7.2 mo. Eighty percent of patients received prior sorafenib treatment. Plasma samples were collected during patient screening, cycle 1 day 1 (pretreatment), and cycle 2 day 14. The Exiqon RT microRNA PCR Human panel I+II was used to measure 752 miRs. Expression levels of detectable miRs and their association with overall survival (OS) were investigated. Results: Low plasma levels of miR-665 (HR = 0.50, p = 0.001, q = 0.13), miR-320d (HR = 0.44, p = 0.001, q = 0.13), miR-320a (HR = 0.47, p = 0.001, q = 0.13), and miR-130b-3p (HR = 0.44, p = 0.002, q = 0.13) are associated with better survival. Whereas, low plasma levels of miR-451a (HR = 2.0, p = 0.002, q = 0.13), miR-let7g-5p (HR = 2.3, p = 0.002, q = 0.13), and miR-18a-3p (HR = 2.0, p = 0.002, q = 0.13) are associated with poor survival. To assess within patient baseline biological variation of miR expression, a comparison of the expression of 369 miRs in patients (n = 42) with 2 pre-treatment samples was performed. The proportion of miRs attaining statistical significance was smaller than what we would expect by chance. Conclusions: Circulating miRs may serve as easily accessible markers to identify HCC patients who may benefit from galunisertib treatment, which requires confirmation in randomized controlled study. Given the low intra-patient variability measured at baseline for most of the miRs, circulating miRs may represent reliable molecular markers with prognostic and/or predictive utility. Citation Format: Shawn T. Estrem, Michael Man, Xuekui Zhang, Duytrac Nguyen, Danni Yu, Michael M. Lahn, Ann Cleverly, Durisala Desaia