Purpose: Dasatinib has limited single-agent activity in unselected patients with metastatic breast cancer. Several gene signatures predictive of dasatinib response in vitro have been reported. The purpose of this three-arm, phase II study was to prospectively assess the utility of three previously published gene signatures to select patients with clinical benefit from dasatinib.
Experimental design: Patients with metastatic breast cancer underwent biopsy for gene expression profiling in an academic CLIA laboratory; those who were positive for any one of three predictive gene signatures (dasatinib sensitivity signature, SRC pathway activity signature, and dasatinib target index) received dasatinib 100 mg orally daily. The three marker-defined cohorts were analyzed separately, using early stopping rules for futility.
Results: Ninety-seven patients were enrolled, 93 underwent biopsy, and 80% of the biopsies were sufficient for molecular testing. Thirty patients were positive for at least one signature and received treatment. Only 1 patient experienced clinical benefit and had stable disease over 300 days. All three arms were closed early for futility. There was one serious biopsy-related adverse event (hematoma and pain following a liver biopsy). There were no unexpected toxicities from dasatinib.
Conclusion: None of the three predictive gene signatures, although supported by preclinical evidence, defined tumors clinically sensitive to dasatinib as a single agent.
©2014 American Association for Cancer Research.