The role of primary tumour sidedness, EGFR gene copy number and EGFR promoter methylation in RAS/BRAF wild-type colorectal cancer patients receiving irinotecan/cetuximab

Background: The data from randomised trials suggested that primary tumour sidedness could represent a prognostic and predictive factor in colorectal cancer (CRC) patients, particularly during treatment with anti-epidermal growth factor receptor (EGFR) therapy. However, an in-deep molecular selection might overcome the predictive role of primary tumour location in this setting. Methods: We conducted a retrospective analysis in which tumour samples from RAS/ BRAF wild-type (WT) metastatic CRC patients treated with second-third-line irinotecan/ cetuximab were analysed for EGFR gene copy number (GCN) and promoter methylation. Study objective was to evaluate the correlation of tumour sidedness, EGFR promoter methylation and EGFR GCN with clinical outcome. Median follow-up duration was 14.3 months. Results: Eighty-eight patients were included in the study, 27.3% had right-sided CRC, 72.7% had left-sided CRC; 36.4% had EGFR GCN<2.12 tumour, 63.6% had EGFR GCN >= 2.12 tumour; 50% had EGFR promoter-methylated tumour. Right-sided colorectal cancer (RSCRC) were associated with reduced overall response rate (ORR) (4.2% for RSCRC vs 35.9% for left sided colorectal cancer (LSCRC), P = 0.0030), shorter progression-free survival (PFS) (3.0 vs 6.75 months, P<0.0001) and shorter overall survival (OS) (8 vs 13.6 months, P<0.0001). EGFR GCN<2.12 tumours were associated with reduced ORR (6.2% for EGFR GCN<2.12 vs 39.3% for EGFR GCN >= 2.12 tumours, P = 0.0009), shorter PFS (3.5 vs 6.5 months, P = 0.0006) and shorter OS (8.5 vs 14.0 months, P<0.0001). Epidermal growth factor receptor-methylated tumours were associated with reduced ORR (9.1% for methylated vs 45.5% for unmethylated, P = 0.0001), shorter PFS (3 vs 7.67 months, P<0.0001) and shorter OS (8 vs 17 months, P<0.0001). At multivariate analysis, EGFR GCN and EGFR promoter methylation maintained their independent role for ORR (respectively P = 0.0082 and 0.0025), PFS (respectively P = 0.0048 and<0.0001) and OS (respectively P = 0.0001 and<0.0001). Conclusions: In our study, an accurate molecular selection based on an all RAS and BRAF analysis along with EGFR GCN and EGFR promoter methylation status seems to be more relevant than primary tumour sidedness in the prediction of clinical outcome during cetuximab/irinotecan therapy. However, these data need to be validated with future prospective and translational studies.