Clinicopathological, prognostic and predictive value of CD166 expression in colorectal cancer: a meta-analysis

Abstract

CD166 has been identified as an important cancer stem cell (CSC) marker in colorectal cancer (CRC). The purpose of our study was to investigate the relationship between CD166 expression and clinical features and to examine the role of CD166 expression on the survival of patients with CRC. A total of 15 studies with 3,332 cases were identified in this meta-analysis. The pooled OR indicated that CD166 expression was significantly higher in CRC than in colonic adenomas or normal colonic mucosa (OR = 3.48, P = 0.002 and OR = 55.13, P = 0.017, respectively). CD166 expression was found to be negatively correlated with vascular invasion (OR = 0.75, P = 0.017), but it was not associated with gender, tumor location, lymph node status, distant metastasis, clinical stage, T classification or tumor differentiation. Meanwhile, CD166 expression was not associated with the prognosis of overall survival (OS) (HR = 1.20, 95% CI = 0.45-3.22, P = 0.72) in multivariate regression analysis. One study reported that CD166 expression may be a predictor of survival in stage II CRC patients using multivariate logistic regression analysis (OS: OR = 9.97, P = 0.035; disease-specific survival: OR = 29.02, P = 0.011). Our findings suggest that CD166 expression may be correlated with CRC carcinogenesis and a decreased risk of vascular invasion, and it may become a predictive biomarker of survival for stage II CRC patients, but additional studies with large sample sizes are essential to validate the prognostic and predictive values of CD166 expression.

Keywords: CD166; CRC; clinical features; expression; survival.

Figure 1. Flow diagram of the selection procedure for this study

Figure 2. Forest plot of the relationship of CD166 expression between CRC and control groups, cancer vs. colonic adenoma

OR = 3.48, 95% CI = 1.55-7.79, P = 0.002; cancer vs. normal colonic mucosa: OR = 55.13, 95% CI = 2.04-1486.86, P = 0.017.

Figure 3. Forest plot of the relationship of CD166 expression with vascular invasion and gender status in colorectal cancer, vascular invasion (yes vs. no)

OR = 0.75, 95% CI = 0.60-0.95, P = 0.017; gender (male vs. female): OR = 0.89, 95% CI = 0.68-1.17, P = 0.414.

Figure 4. Forest plot of the correlation of CD166 expression with lymph node status and distant metastasis in colorectal cancer, distant metastasis (yes vs. no)

OR = 1.60, 95% CI = 0.83-3.10, P = 0.16; lymph node status (yes vs. no): OR = 1.35, 95% CI = 0.87-2.11, P = 0.183.

Figure 5. Forest plot of the association of CD166 expression with tumor location and T classification in colorectal cancer, tumor location (left vs. right)

OR = 0.59, 95% CI = 0.28-1.21, P = 0.15; T classification (pT3-4 vs. pT1-2): OR = 1.55, 95% CI = 0.82-2.94, P = 0.181.

Figure 6. Forest plot of the relationship between CD166 expression and tumor differentiation and clinical stage in colorectal cancer, tumor differentiation (poor vs. moderate/well)

OR = 1.45, 95% CI = 0.80-2.63, P = 0.217; clinical stage (stage 3-4 vs. stage 1-2): OR = 1.73, 95% CI = 0.91-3.27, P = 0.092.

Figure 7. Forest plot showing the prognostic values of CD166 expression in overall survival (HR = 1.20, 95% CI = 0.45-3.22, P = 0.72), disease free survival (HR = 5.61, 95% CI = 1.82-17.36, P = 0.003), and progression free survival (HR = 0.65, 95% CI = 0.32-1.32, P = 0.233) for multivariate analysis

Figure 8. Forest plot of the possible publication bias in relation to gender status (P = 0.997), tumor differentiation (P = 0.819), clinical stage (P = 0.511), T classification (P = 0.009) and lymph node status (P = 0.010) using Egger's test