Incidence and prognostic influence of lymph node micrometastases in rectal cancer

The aim of this study was to determine the rate of lymph node micrometastases and evaluate their prognostic significance in rectal cancer.

Patients with either Dukes A or B rectal carcinoma who had undergone curative resection by either low anterior resection or abdominal perineal resection between 1991 and 2000 were selected from a prospectively collated database. None of the patients had metastasis at the time of surgery and none received adjuvant or neoadjuvant therapy. A single section from each lymph node was stained with haematoxylin and eosin (H+E) and with CAM 5.2 by immunohistochemistry. Statistical analyses were performed with Chi-square test.

A total of 774 lymph nodes with a median of 14 lymph nodes per patient were examined, from a cohort of 56 patients with a median age of 66 years. In the 56 patients in whom lymph node metastases were not detected by haematoxylin-eosin staining, cytokeratin staining was positive in 15 lymph nodes from 10 patients. Nine patients had disease recurrence at a median follow-up of 98 months. The presence of lymph node micrometastases by immunohistochemistry did not predict either disease-free (p=0.44) or overall survival (p=0.63).

Immunohistochemical staining detects micrometastases in rectal cancer which are not observed with H+E staining. However, no significant relationship was observed between disease relapse and rectal micrometastases detected by immunohistochemistry.