1. ¹Department of Pathology Roswell Park Cancer Institute, State University of New York, Buffalo, NY 14263, USA
2. ²Department of Laboratory Medicine Roswell Park Cancer Institute, State University of New York, Buffalo, NY 14263, USA
3. ³Department of Medicine Roswell Park Cancer Institute, State University of New York, Buffalo, NY 14263, USA
4. ⁴Department of Surgery Roswell Park Cancer Institute, State University of New York, Buffalo, NY 14263, USA

Correspondence: D Tan, Departments of Pathology and Cancer Genetics, Roswell Park Cancer Institute, State University of New York, Buffalo, Elm and Carlton Streets, Buffalo, NY 14263, USA. E-mail: dongfeng.tan@roswellpark.org

Received 3 July 2003; Revised 9 September 2003; Accepted 17 October 2003; Published online 23 January 2004.

Abstract

E-cadherin (E-cad) and epidermal growth factor receptor (EGFR) are important cell adhesion and signaling pathway mediators. This study aimed to assess their expression in lung adenocarcinoma (AdC) and squamous cell carcinoma (SCC) and their association with clinicopathologic variables. In all, 130 resectable lung cancers (stages I–IIIA) were studied using a high-density tissue microarray. Two to three cores from each case were arrayed into three blocks using a Beecher system. Immunohistochemistry was performed using an avidin–biotin complex method and monoclonal antibodies against E-cad and EGFR. Unequivocal membrane staining in >10% of tumor cells was considered as a positive expression of E-cad and EGFR. Markers expression and coexpression were analyzed against clinicopathologic variables (age, gender, smoking status, performance status, weight loss, histology, grade, stage, and lymph node involvement) and patient survival. There were 118, 126, and 115 cases that were fully assessable for E-cad, EGFR, and both markers, respectively. For E-cad, 65 cases (55%) were positive (+), 53 (45%) were negative (-); 23 cases of the negative group had only cytoplasmic staining. For EGRF, 43 cases (34%) were (+), and 83 (66%) were (-). There was no significant association between E-cad or EGFR, and any of the clinicopathologic variables except for an association between EGFR(+) and SCC histologic type. Both negative and cytoplasmic staining of E-cad correlated with shorter patient survival with P=0.008 and 0.002, respectively. EGFR expression did not correlate with patient survival; however, patients with E-cad(-)/EGFR(+) phenotype had poorer survival than those with E-cad(+)/EGFR(-) (P=0.026). Our study suggests that lung AdC and SCC may be stratified based on expression of E-cad and EGFR with the E-cad(-)/EGFR(+) expression having a worse disease outcome. Moreover, the cytoplasmic expression of E-cad may represent an altered localization of this protein in association with tumorigenicity.