Abstract
BACKGROUND: Crizotinib produces high response rates and prolonged PFS in ALK+ NSCLC. Retrospective analyses suggest enhanced sensitivity to pemetrexed in crizotinib naive ALK+ NSCLC. Cross-resistance between crizotinib and pemetrexed has not been previously investigated. PATIENTS AND METHODS: Patients with stage IV ALK+ NSCLC treated with PEM-CRIZ, or CRIZ-PEM were identified. Overall PFS and PFS excluding central nervous system events (eCNS) were compared. RESULTS: Objective response rates in evaluable patients were 66% (PEM-CRIZ) and 75% (CRIZ-PEM) for pemetrexed and 84% (CRIZ-PEM) and 66% (PEM-CRIZ) for crizotinib. For PEM-CRIZ (n = 29), median PFS and eCNS PFS were both 6 months with pemetrexed, and 10 and 14.5 months, respectively, with crizotinib. For CRIZ-PEM (n = 9), median PFS and eCNS PFS were 4.5 and 3 months, respectively, with pemetrexed, and 8.5 and 7.5 months, respectively, with crizotinib. There was a statistically significant increase in the risk of an overall PFS event with pemetrexed when administered after crizotinib (P = .0277; hazard ratio [HR], 2.5898; 95% confidence interval [CI], 1.1100-6.0424), but differences in the risk of an eCNS PFS event were not significant (P = 0.4913; HR, 1.3521; 95% CI, 0.5727-3.1920). Neither overall nor eCNS PFS for patients while taking crizotinib was associated with a sequence effect relative to pemetrexed. CONCLUSION: Crizotinib and pemetrexed are active drugs in ALK+ NSCLC. PFS benefit appeared higher with crizotinib than with pemetrexed. PFS benefit from pemetrexed was less after crizotinib compared with before crizotinib, however, this difference was only statistically significant for overall and not eCNS PFS. Pemetrexed exposure did not seem to affect crizotinib outcomes.