SP1.2.13Prediction of malignant progression of colorectal adenomatous polyps – A complete systematic review and meta-analysis

Abstract

Aims Only 5% of colorectal adenomas (CA) progress to cancer. Identifying these is highly relevant for colorectal cancer screening risk. Currently, no biomarkers exist that predict this malignant progression. We looked at the effectiveness of common genetic aberrations as potential biomarkers through systematic review and meta-analyses. Methods MEDLINE, EMBASE and Cochrane Library were searched to identify all studies that assessed p53, APC, p21, BRAF, MLH1, MSH2, CIMP and Kras mutations as prognostic markers in CA. Main outcome measure was the development of colorectal cancer. Results 109 clinical studies were included. P53 mutation [DOR 8.37 (5.21-13.42), sensitivity 55% (52-58%), specificity 85% (83-87%), PLR 4.23 (2.60-6.88), NLR 0.57 (0.57-0.64), AUC 0.7532] was found to be superior to all other mutations. Mutations of APC [DOR 0.57 (0.16-2.01), sensitivity 19% (4-25%), specificity 71% (66-76%), PLR 0.71 (0.33-1.54), NLR 1.19 (0.84-1.69)], BRAF [DOR 0.38 (0.22-0.67), sensitivity 12% (11-14%), specificity 60% (58-62%), PLR 0.57 (0.38-0.84), NLR 1.27 (1.12-1.44)], Kras [DOR 1.22 (0.72-2.0), sensitivity 33% (30-37%), specificity 74% (71-76%), PLR 1.17 (0.81-1.68), NLR 0.94 (0.83-1.07)], MLH1 [DOR 2.48 (1.05-5.84), sensitivity 26% (23-30%), specificity 82% (79-84%), PLR 1.74 (0.95-3.19)], MSH2 [DOR 1.06 (0.53-2.13), sensitivity 22% (17-27%), specificity 88% (84-90%), PLR 1.03 (0.62-1.70), NLR 0.99 (0.91-1.07)] and CIMP [DOR 1.88 (0.25-13.88), sensitivity 30% (23-36%), specificity 76% (72-80%), PLR 1.31 (0.29-5.99), NLR 0.76 (0.40-1.45)] failed to demonstrate any advantage over p53. Conclusions p53 mutations effectively predict malignant progression in CA. Panel of biomarkers would be more suited for surveillance programme. This needs confirmation in prospective clinical trials with cost-efficiency analyses.