Genetic variability in DNA repair and clinical outcome in stage III non-small cell lung cancer (NSCLC) treated with concurrent platinum-based chemoradiation

Abstract

e14621 Background: Concurrent platinum-based chemoradiation (CRT) is the standard of care for stage III NSCLC. CRT is associated with multiple toxicities and substantial interpatient variability in clinical outcomes, with a 20% cure rate. We evaluated the relationship of genetic polymorphisms in platinum metabolic and DNA repair pathways with clinical outcomes in patients treated for stage III NSCLC. Methods: This pilot study was IRB approved and all patients provided written informed consent. Patients with stage III NSCLC who completed CRT were eligible for enrollment. Genomic DNA was extracted from peripheral blood. Genotyping using Pyrosequencing™ was performed for 5 single nucleotide polymorphisms (SNPs): the detoxification enzyme GSTP1 I105V, and the DNA repair genes ERCC1 N118N, ERCC2 K751Q, ERCC2 R156R and XRCC1 R399Q. Pre-treatment variables, progression-free survival (PFS) and toxicity data were obtained from retrospective review of medical records. Results: 23 pts have been enrolled to date. Median age was 58 yrs (49–81). 5 pts received cisplatin and 18 carboplatin. Median radiation dose was 6480 cGy (5000–7020) and median V20 was 27%. 10 pts had grade 3/4 acute CRT toxicities and 7 pts required treatment breaks. Median duration of follow-up was 21 months (4–100). 14 pts recurred and median PFS was 12 months (4–100). 15 pts remain alive. Controlling for age, V20 and radiation dose and using multivariate cox proportional hazards regression, the ERCC2 K751Q and R156R genotypes were independent predictors of PFS (HRs= 0.404 & 2.184, p-values= 0.0172 & 0.0152, respectively, subjects with at least 1 variant allele being reference group). Using multiple logistic regression, the ERCC2 R156R variant showed a trend towards improved response (OR=2.98, p-value= 0.09). Overall survival data is limited by the majority of pts still being alive. Conclusions: Clinical and treatment variables were typical of a stage III NSCLC population. This pilot pharmacogenomic study indicates that outcomes were independently affected by genetic polymorphisms. Our findings, if validated prospectively, could have significant impact on optimizing treatment selection for stage III NSCLC pts, without need for additional tumor tissue. No significant financial relationships to disclose.