Systematic review and meta-analysis of tumor biomarkers in predicting prognosis in esophageal cancer

Abstract

Background: Esophageal cancer (EC) is a frequently occurring cancer with poor prognosis despite combined therapeutic strategies. Many biomarkers have been proposed as predictors of adverse events. We sought to assess the prognostic value of biomarkers in predicting the overall survival of esophageal cancer and to help guide personalized cancer treatment to give patients the best chance at remission.

Methods: We conducted a systematic review and meta-analysis of the published literature to summarize evidence for the discriminatory ability of prognostic biomarkers for esophageal cancer. Relevant literature was identified using the PubMed database on April 11, 2012, and conformed to the REMARK criteria. The primary endpoint was overall survival and data were synthesized with hazard ratios (HRs).

Results: We included 109 studies, exploring 13 different biomarkers, which were subjected to quantitative meta-analysis. Promising markers that emerged for the prediction of overall survival in esophageal squamous cell cancer included VEGF (18 eligible studies, n=1476, HR=1.85, 95% CI, 1.55-2.21), cyclin D1 (12 eligible studies, n=1476, HR=1.82, 95% CI, 1.50-2.20), Ki-67 (3 eligible studies, n=308, HR=1.11, 95% CI, 0.70-1.78) and squamous cell carcinoma antigen (5 eligible studies, n=700, HR=1.28, 95% CI, 0.97-1.69); prognostic markers for esophageal adenocarcinoma included COX-2 (2 eligible studies, n=235, HR=3.06, 95% CI, 2.01-4.65) and HER-2 (3 eligible studies, n=291, HR=2.15, 95% CI, 1.39-3.33); prognostic markers for uncategorized ECs included p21 (9 eligible studies, n=858, HR=1.27, 95% CI, 0.75-2.16), p53 (31 eligible studies, n=2851, HR=1.34, 95% CI, 1.21-1.48), CRP (8 eligible studies, n=1382, HR=2.65, 95% CI, 1.64-4.27) and hemoglobin (5 eligible studies, n=544, HR=0.91, 95% CI, 0.83-1.00).

Conclusions: Although some modest bias cannot be excluded, this review supports the involvement of biomarkers to be associated with EC overall survival.

Figure 1

Flow diagram of the literature search and selection of eligible studies.

Figure 2

Characteristics of included studies in the systematic review. A. Frequencies with which adjustments were made for various clinicopathologic parameters. B. Distributions of the total number of clinicopathologic covariates that were adjusted for across the 109 eligible studies. NOTE: nr, not report.

Figure 3

Quality of individual study reports (n = 16 items, n = 109 studies), based on the REMARKER guidelines. Definition items of each item are given in Additional file 1: Table S1.

Figure 4

Forest plots of the data for four biomarker-outcome comparisons for which eligible data were presented in ten or more studies. Forest plots of HR for OS of (A) VEGF, (B) cyclin D1, (C) p53 and (D) E-cadherin. For each study, the hazard ratio (HR), 95% confidence interval (CI), and relative weight are show. Combined fixed effect HRs and tests for heterogeneity (I²) were based on the generic inverse variance (I-V) method. Combined random effect HRs were calculated according to the DerSimonian-Laird (D + L) method.