Among patients with metastatic melanoma, detectable circulating levels of BRAFV600 mutated circulating tumor DNA and elevated levels of circulating hepatocyte growth factor separately predicted significantly worse survival, regardless of treatment.
After adjusting for lactate dehydrogenase, ECOG (Eastern Cooperative Oncology Group) status, disease stage, and treatment, hazard ratios for overall survival were 1.75 (95% confidence interval, 1.35-2.28) for detectable versus undetectable circulating tumor DNA (ctDNA) and 1.24 (95% CI, 1.00-1.53) for high versus low circulating hepatocyte growth factor (cHGF), reported William Lu, PhD, of Genentech in South San Francisco, with his associates. Their retrospective, exploratory analysis of the phase 3 BRIM-3 trial was published in JCO Precision Oncology.
BRIM-3 was a multicenter, open-label trial of 675 patients with previously untreated unresectable stage IIIC or stage IV melanoma that tested positive for the BRAFV600 mutation. Patients received either vemurafenib or dacarbazine therapy.
Baseline ctDNA, baseline cHGF, and ECOG performance status most strongly predicted overall survival, said the investigators. Patients whose mutant ctDNA fraction exceeded 0.039 had an 11.5-month shorter median overall survival on vemurafenib (P less than .001) and a 13.9-month shorter median overall survival on dacarbazine (P less than .001) than patients whose mutant ctDNA fraction was less than 0.039, they explained. Similarly, median overall survival times were 5.4 months shorter for vemurafenib (P = .002) and 12.3 months shorter for dacarbazine (P less than .001) when patients had high (more than 438 pg/mL) versus low circulating cHGF.
Median overall survival for vemurafenib was shortest (7.3 months) in the subgroup of 64 patients with detectable BRAFV600 ctDNA and elevated cHGF, according to the researchers. In contrast, median overall survival for vemurafenib was longest (22.0 months) when patients had undetectable ctDNA and an ECOG status of 0.
“Although similar studies have been performed, to our knowledge, our study is unique in its size and in its design as a randomized trial,” Dr. Lu and associates concluded. “The biomarkers in this study can be readily acquired and measured, and require only small volumes of plasma.” They suggested validating the findings in a separate dataset.
F. Hoffman-La Roche provided funding. Roche Molecular Systems makes the Cobas 4800 BRAFV600 Mutation Test used in the study. Dr. Lu disclosed employment and research funding from Genentech/Roche.
SOURCE: Lu W et al. JCO Precis Oncol. 2018 Apr 25. doi: 10.1200/PO.17.00168.