BACKGROUND: To assess the role of immunohistochemically detectable nodal microinvolvement of patients with "curatively" resected esophageal carcinoma. METHODS: In 73 patients with resectable esophageal carcinoma [squamous cell carcinoma (SCC), n = 45 (61.6%); adenocarcinoma (AC), n = 28 (38.4%)] a total of 2174 lymph nodes (LN) were removed. In each of the 1958 LN classified as negative on conventional histopathology, immunohistochemistry was performed using the anticytokeratin antibody AE1/AE3. To determine the role of the amount of residual tumor load, the patients were grouped according to the percentage of LN affected with micrometastasis (0%, or =11%). RESULTS: Tumor cells were immunohistochemically detected in 47 LN (2.4%) from 25 (34.2%) patients. Five-year overall survival probability (5-YSP) of 30% in pN(0 )patients with detected occult tumor cells in LN was significantly worse than that in those without nodal microinvolvement (76%, P = 0.021), hereby resembling that of pN1-patients (24%, P = 0.84). Median overall survival in patients with no (0%), low (11%) micrometastatic tumor load was 43, 27, and 11 months, respectively. Substratification according to histological type showed that, in patients with AC, the presence of nodal microinvolvement had a significant impact on 5-YSP (0% versus 65%; P = 0.03), whereas in patients with SCC, differences of 5-YSP were only of borderline significance (24% versus 53%; P = 0.081). CONCLUSION: Minimal tumor cell load as assessed by the ratio of micrometastatically affected LN is a complementary tool for better risk stratification of patients with esophageal carcinoma.
