Meta-analysis of biomarkers predicting risk of malignant progression in Barrett's oesophagus.

BACKGROUND: Barrett's oesophagus is a precursor to the development of oesophageal adenocarcinoma. This study sought to clarify the role of genetic, chromosomal and proliferation biomarkers that have been the subjects of multiple studies through meta-analysis.

METHODS: MEDLINE, Embase, PubMed and the Cochrane Library were searched for clinical studies assessing the value of p53, p16, Ki-67 and DNA content abnormalities in Barrett's oesophagus. The main outcome measure was the risk of development of high-grade dysplasia (HGD) or oesophageal adenocarcinoma.

RESULTS: Some 102 studies, with 12 353 samples, were identified. Mutation (diagnostic odds ratio (DOR) 10·91, sensitivity 47 per cent, specificity 92 per cent, positive likelihood ratio (PLR) 4·71, negative likelihood ratio (NLR) 0·65, area under the curve (AUC) 0·792) and loss (DOR 16·16, sensitivity 31 per cent, specificity 98 per cent, PLR 6·66, NLR 0·41, AUC 0·923) of p53 were found to be superior to the other p53 abnormalities (loss of heterozygosity (LOH) and overexpression). Ki-67 had high sensitivity in identifying high-risk patients (DOR 5·54, sensitivity 82 per cent, specificity 48 per cent, PLR 1·59, NLR 0·42, AUC 0·761). Aneuploidy (DOR 12·08, sensitivity 53 per cent, specificity 87 per cent, PLR 4·26, NLR 0·42, AUC 0·846), tetraploidy (DOR 5·87, sensitivity 46 per cent, specificity 85 per cent, PLR 3·47, NLR 0·65, AUC 0·793) and loss of Y chromosome (DOR 9·23, sensitivity 68 per cent, specificity 80 per cent, PLR 2·67, NLR 0·49, AUC 0·807) also predicted malignant development, but p16 aberrations (hypermethylation, LOH, mutation and loss) failed to demonstrate any advantage over the other biomarkers studied.

CONCLUSION: Loss and mutation of p53, and raised level of Ki-67 predicted malignant progression in Barrett's oesophagus.