Recurrent nerve nodal involvement is associated with cervical nodal metastasis in thoracic esophageal carcinoma

Background: Because three-field dissection for esophageal carcinoma is one of the most invasive operations, this procedure should be selected only when strictly indicated; but there are no useful criteria for it. The goal of this study was to identify the useful clinicopathologic factors for indicating three-field dissection. Study Design: In this study, we reviewed the survival of patients after three-field dissection and identified factors associated with metastases in cervical nodes (CN), especially internal jugular nodes and supraclavicular nodes. Eighty-six patients who underwent curative esophagectomy with three-field dissection for squamous cell carcinoma of the thoracic esophagus were enrolled in this study. Survival rates were compared with respect to the presence of nodal metastasis. The relationship between recurrent nerve nodal (RNN) involvement and CN metastasis (bilateral internal jugular nodes, supraclavicular nodes, or both) was examined. Clinicopathologic factors possibly influencing CN metastasis were studied by multivariate logistic regression analysis. Results: The overall 5-year survival rate was 45.1%. The 5-year survival rate for patients without metastatic nodes was 67.5%, for patients with one to four metastatic nodes it was 53.1%, and for patients with five or more it was 9.1%. The prognosis of those with five or more metastatic nodes was significantly poorer than those of the other two groups. In the positive-node group, the 5-year survival rate for patients with RNN metastasis was 21.7%, and for patients with negative RNN it was 47.0% (p = 0.2). In the positive-node group, the 5-year survival rate for patients with positive CN was 13.7% and for patients with negative CN it was 45.8% (p = 0.01). Fifty-six (88.9%) of 63 patients without RNN metastasis had no CN metastasis in contrast to 13 of 23 patients (56.5%) with RNN metastasis who had no CN metastasis (p = 0.001). The positive predictive value, negative predictive value, sensitivity, and specificity were 43.5%, 88.8%, 58.8%, and 81.2%, respectively. The number of metastatic nodes (5 or more versus 0–4) (odds ratio: 2.9, 95% Confidence Interval (CI) = 1.6–5.5, p = 0.0008) and RNN involvement (odds ratio: 4.5, 95% CI = 1.3–15.9, p = 0.02) were the significant factors associated with CN metastasis in the multivariate analysis. Conclusions: RNN involvement is associated with CN metastasis as is the number of metastatic nodes and may be an indicator for the selection of three-fiel