Prognostic Role of Mucin Antigen MUC4 for Cholangiocarcinoma: A Meta-Analysis

Abstract

Background and objective: Surgery carries the best hope for cure in the treatment of cholangiocarcinoma (CC), whereas surgical outcome is not fully satisfactory. Bio-molecular markers have been used to improve tumor staging and prognosis prediction. Mucin antigen MUC4 (MUC4) has been implicated as a marker for poor survival in various tumors. However, prognostic significance of MUC4 for patients with CC remains undefined. The aim of the present meta-analysis was to investigate the association between MUC4 expression and overall survival (OS) of patients with resected CC.

Methods: The meta-analysis was conducted in adherence to the MOOSE guidelines. PubMed, Embase databases, Cochrane Library and the Chinese SinoMed were systematically searched to identify eligible studies from the initiation of the databases to April, 2016. OSs were pooled by using hazard ratio (HR) with corresponding 95% confidence interval (CI). Random effect models were utilized because of the between-study heterogeneities.

Results: Five studies reporting on 249 patients were analyzed: 94 (37.75%) were in positive or high expression group and 155 (62.25%) in negative or low expression group. The pooled HR for positive or high expression group was found to be 3.04 (95% CI 2.25-4.12) when compared with negative or low expression group with slight between-study heterogeneities (I2 3.10%, P = 0.39). The result indicated that a positive or high expression level of MUC4 was significantly related to poor survival in patients with resected CC. A commensurate result was identified by sensitivity analysis. The main limitations of the present meta-analysis were the rather small size of the studies included and relatively narrow geographical distribution of population.

Conclusion: The result of this meta-analysis indicated that a positive or high expression level of MUC4 was significantly related to poor survival in patients with resected CC.

Fig 1. Flowchart of the study selection.

Systematic search and selection of relevant articles.

Fig 2. Results of the meta-analysis on pooled HR values.

Each square denotes the HR for that trial comparison with the horizontal lines showing the 95% CIs. The size of the square is directly proportional to the amount of information contributed by the trial. The hollow blue diamond gives the pooled HR from the random effect model; the centre of this diamond denotes the HR and the extremities the 95% CI.

Fig 3. Result of sensitivity analysis.

The middle vertical line indicates the combined HR, and the two vertical lines represent the corresponding 95% CI values. The middle small circle and two ends of the dotted lines indicate the pooled HR and 95% CI values, respectively, when the study on the left was omitted after each round of analysis.

Fig 4. Begg's funnel plot to evaluate OS.

Funnel plot showing symmetry indicative of no evidence of publication bias for OS.