BACKGROUND: The chemokine and bone marrow-homing receptor CXCR4 has been implicated in metastatic dissemination of various cancers. We investigated CXCR4 expression in esophageal cancer specimens and its association with survival, lymph node microinvolvement, and bone marrow micrometastasis. METHODS: We analyzed frozen tumor specimens from 136 patients with completely resected esophageal cancer for CXCR4 expression by immunohistochemistry. Lymph node microinvolvement and bone marrow micrometastasis were assessed by immunohistochemistry with monoclonal antibodies Ber-EP4 (against epithelial cell adhesion molecule) and pancytokeratin A45-B/B3 (against several cytokeratins), respectively. Associations between CXCR4 expression and clinicopathologic features, including tumor stage, histologic grade, lymph node metastasis and microinvolvement, bone marrow micrometastasis, and survival, were investigated with Fisher's test, log-rank test, and Cox multivariable analysis. All statistical tests were two-sided. RESULTS: CXCR4 protein was expressed in 75 (55%) of 136 esophageal tumors examined. CXCR4 expression was statistically significantly associated with reduced median overall and disease-specific survival, compared with CXCR4 nonexpression (P < .001; log-rank test). The median overall survival of patients with CXCR4-positive tumors was 20 months and with CXCR4-negative tumors, 76 months (difference = 56 months, 95% confidence interval [CI] = 4 to 108 months; P < .001). The median disease-specific survival of patients with CXCR4-positive tumors was 25 months and with CXCR4-negative tumors was 97 months (difference = 72 months, 95% CI = 34 to 110 months; P < .001). CXCR4 expression was statistically significantly associated with increased lymph node microinvolvement (P < .001) and with increased bone marrow micrometastasis (P < .001). In multivariable analysis, CXCR4 expression, compared with its nonexpression, was identified as the independent variable that was most strongly associated with reduced disease-specific survival (relative risk [RR] of death = 2.03, 95% CI = 1.20 to 3.41; P = .008) and overall survival (RR of death = 2.18, 95% CI = 1.33 to 3.59; P = .002). CONCLUSION: CXCR4 expression was associated with poor clinical outcome in esophageal cancer patients. CXCR4 may have a role in early metastatic spread because its expression was associated with micrometastases to both the lymph nodes and bone marrow. Thus, CXCR4 should be explored further as a target for adjuvant therapy for micrometastatic disease.
BACKGROUND: The chemokine and bone marrow-homing receptor CXCR4 has been implicated in metastatic dissemination of various cancers. We investigated CXCR4 expression in esophageal cancer specimens and its association with survival, lymph node microinvolvement, and bone marrow micrometastasis. METHODS: We analyzed frozen tumor specimens from 136 patients with completely resected esophageal cancer for CXCR4 expression by immunohistochemistry. Lymph node microinvolvement and bone marrow micrometastasis were assessed by immunohistochemistry with monoclonal antibodies Ber-EP4 (against epithelial cell adhesion molecule) and pancytokeratin A45-B/B3 (against several cytokeratins), respectively. Associations between CXCR4 expression and clinicopathologic features, including tumor stage, histologic grade, lymph node metastasis and microinvolvement, bone marrow micrometastasis, and survival, were investigated with Fisher's test, log-rank test, and Cox multivariable analysis. All statistical tests were two-sided. RESULTS:CXCR4 protein was expressed in 75 (55%) of 136 esophageal tumors examined. CXCR4 expression was statistically significantly associated with reduced median overall and disease-specific survival, compared with CXCR4 nonexpression (P < .001; log-rank test). The median overall survival of patients with CXCR4-positive tumors was 20 months and with CXCR4-negative tumors, 76 months (difference = 56 months, 95% confidence interval [CI] = 4 to 108 months; P < .001). The median disease-specific survival of patients with CXCR4-positive tumors was 25 months and with CXCR4-negative tumors was 97 months (difference = 72 months, 95% CI = 34 to 110 months; P < .001). CXCR4 expression was statistically significantly associated with increased lymph node microinvolvement (P < .001) and with increased bone marrow micrometastasis (P < .001). In multivariable analysis, CXCR4 expression, compared with its nonexpression, was identified as the independent variable that was most strongly associated with reduced disease-specific survival (relative risk [RR] of death = 2.03, 95% CI = 1.20 to 3.41; P = .008) and overall survival (RR of death = 2.18, 95% CI = 1.33 to 3.59; P = .002). CONCLUSION:CXCR4 expression was associated with poor clinical outcome in esophageal cancerpatients. CXCR4 may have a role in early metastatic spread because its expression was associated with micrometastases to both the lymph nodes and bone marrow. Thus, CXCR4 should be explored further as a target for adjuvant therapy for micrometastatic disease.