Meta-analysis of the prognostic value of CpG island methylator phenotype in rectal cancer

PURPOSE: The pathological and prognostic importance of CpG island methylator phenotype (CIMP) in rectal cancer, as a sub-population of colorectal cancer, is unknown. A meta-analysis was preformed to estimate the prognostic significance of CIMP in rectal cancer. METHODS: A systematic search was performed of PubMed, Embase, MEDLINE, PubMed Central, and Cochrane electronic databases for articles pertaining to CIMP and rectal cancer. Articles were analysed and data extracted according to PRISMA standards. RESULTS: Six studies including 1529 patients were included in the analysis. Following dichotomisation, the prevalence of CIMP-positive tumours was 10 to 57%, with a median of 12.5%. Meta-analysis demonstrated the pooled odds ratio for all-cause death for CIMP-positive tumours vs CIMP-negative tumours was 1.24 (95% CI 0.88–1.74). Z test for overall effect was 1.21 (p = 0.23). Heterogeneity between the studies was low (X(2) 5.96, df 5, p = 0.31, I(2) = 16%). A total of 15 different loci were used for assessing CIMP across the studies, with a median of 6.5 loci (range 5–8). CONCLUSIONS: No significant association between CIMP and poor outcomes in rectal cancer was demonstrated. There was a high degree of heterogeneity in CIMP assessment methodologies and in study populations. Rectal cancer datasets were frequently not extractable from larger colorectal cohorts, limiting analysis.