RESUMEN
PURPOSE: Pemetrexed has shown varied response rates in advanced breast cancer. This randomized, double-blind, phase II study was conducted to assess the efficacy and safety of two doses of pemetrexed in a homogeneous population. A secondary objective was to identify molecular biomarkers correlating with response and toxicity. EXPERIMENTAL DESIGN: Patients with newly diagnosed metastatic breast cancer or locally recurrent breast cancer received 600 mg/m(2) (P600 arm) or 900 mg/m(2) (P900 arm) of pemetrexed on day 1 of a 21-day cycle. All patients received folic acid and vitamin B(12) supplementation. RESULTS: The P600 (47 patients) and P900 (45 patients) arms had response rates of 17.0% (95% confidence interval, 7.7-30.8%) and 15.6% (95% confidence interval, 6.5-29.5%) with approximately 50% stable disease per arm, median progression-free survival of 4.2 and 4.1 months, and median times to tumor progression of 4.2 and 4.6 months, respectively. Both arms exhibited minimal toxicity (grade 3/4 neutropenia <20%, leukopenia <9%, and other toxicities <5%). Tumor samples from 49 patients were assessed for the expression levels of 12 pemetrexed-related genes. Folylpolyglutamate synthetase and thymidine phosphorylase correlated with efficacy. Best response rates and median time to tumor progression for high versus low thymidine phosphorylase expression were 27.6% versus 6.3% (P = 0.023) and 5.4 versus 1.9 months (P = 0.076), and for folylpolyglutamate synthetase were 37.5% versus 10.0% (P = 0.115) and 8.6 versus 3.0 months (P = 0.019), respectively. gamma-Glutamyl hydrolase expression correlated with grade 3/4 toxicities: 78.6% for high versus 27.3% for low gamma-glutamyl hydrolase (P = 0.024). CONCLUSION: The two pemetrexed doses yielded similar efficacy and safety profiles. Exploratory biomarker analysis identified efficacy and toxicity correlations and warrants further evaluation.