Purpose: Multiple studies examining the relationship between loss of E-cadherin expression, a pivotal event for evolving metastatic behavior among epithelially derived cancers, and 5-year survival in infiltrating ductal breast carcinoma have yielded inconclusive and contradictory results.
Experimental design: We conducted a systematic review of the PubMed database through August 2005 with no language restrictions to identify cohort studies that evaluated E-cadherin immunohistochemical expression as a prognostic marker for ductal breast carcinoma. 5-year all-cause mortality or 5-year breast cancer-specific mortality were the primary study outcomes. Meta-analysis was conducted using the REVMAN software and summary hazard ratios assuming both fixed effect and random effect models were calculated.
Results: Ten retrospective cohort studies were identified. Reduced or absent E-cadherin expression significantly increased the risk of all-cause mortality [combined HR = 1.55; 95% CI = 1.08-2.23] whereas a non-significant association was observed for breast cancer-specific mortality [combined HR = 0.70; 95% CI = 0.39-1.27]. We documented substantial inter-study heterogeneity with respect to all aspects of clinical data collection, immunohistochemical staining and interpretation as well as statistical modeling. These factors could not be formally analyzed but they challenge the robustness of our calculated summary estimates.
Conclusions: Loss of E-cadherin expression may be an independent negative prognostic indicator for infiltrating ductal breast carcinoma and randomized, controlled studies evaluating this finding are justified. We encourage standardization of immunohistochemical techniques, data interpretation algorithms across laboratories and use of all-cause mortality to increase data compatibility and facilitate future efforts summarizing the utility of alternate prognostic markers in cancer.