Targeted first-line therapies for advanced colorectal cancer: a Bayesian meta-analysis

Abstract

Background: Colorectal cancer is common and deadly. First-line treatments for patients with metastatic disease include FOLFIRI and FOLFOX, which have been combined with anti-EGFR or anti-VEGF antibodies to achieve benefit in selected populations. However, optimal therapy remains unclear.

Results: Fifteen publications on 10 trials were identified. There was a lack of decisive evidence that FOLFIRI or FOLFOX impact efficacy of either anti-EGFR or anti-VEGF, across mutational status groups. On the other hand, evidence suggests both anti-EGFR and anti-VEGF may be more effective for KRAS WT than MT patients. KRAS WT results provided evidence that anti-EGFR treatments may be more effective than anti-VEGF treatments when combined with FOLFIRI or FOLFOX. Further, evidence suggests that both anti-EGFR and anti-VEGF therapies, when combined with FOLFIRI or FOLFOX, may be harmful as compared to chemotherapy for KRAS MT patients.

Materials and methods: Literature was searched for randomized trials comparing anti-EGFR or anti-VEGF antibodies, paired with FOLFIRI or FOLFOX, as first-line therapy for advanced colorectal cancer. Meta-estimates were generated via Bayesian hierarchical log-linear model. The primary endpoint was overall survival.

Conclusions: Further studies examining impact of all-RAS mutation status, left or right side location of primary tumor, and combination anti-VEGF with modern bolus fluoropyrimidine are needed.

Keywords: Bayesian; colorectal cancer; decision analysis; meta-analysis; targeted therapy.

Figure 1. Search diagram for randomized clinical trials comparing FOLFOX or FOLFIRI chemotherapy, potentially paired with anti-EGFR (cetuximab, panitumumab) or anti-VEGF (bevacizumab) treatments, as first-line therapy for patients with inoperable advanced or metastatic colorectal cancer

Figure 2. OS comparisons of anti-EGFR and anti-VEGF therapies in addition to chemotherapy regimens FOLFIRI or FOLFOX as compared to chemotherapy alone as first-line treatment for advanced or metastatic colorectal cancer

(A) and (C) Chemotherapy regimen comparison for all-RAS and KRAS WT. (B) and (D) Mutation status comparison for all-RAS and KRAS. Treatments are compared to chemotherapy alone. Bayesian network estimates reported as hazard ratio (95% credible intervals in black and 95% predictive intervals in red). *Probability that the treatment arm is better than chemotherapy alone.