Detection of pericolonic lymph nodes in patients with colorectal cancer: an in vitro and in vivo study of the efficacy of endosonography.

OBJECTIVE

Rectal endosonography is useful for the preoperative detection of perirectal lymph nodes in patients with rectal carcinoma, but some lymph nodes may be overlooked. The aim of this study was to compare the number of lymph nodes detected at sonography with the actual number and the pathology of the nodes found on examination of surgical specimens.

MATERIALS AND

METHODS

Colorectal specimens from 20 patients who had undergone resection of the colon (six patients) or rectum (14 patients) for carcinoma were selected for study. Sonography with 7-MHz endoprobes was performed with the bowel opened and submerged in water. The location and size of all lymph nodes that could be visualized in the pericolonic fat were marked on a drawing of the specimen. Distinction between malignant and benign nodes was not attempted. After fixation, pericolonic lymph nodes were identified by careful palpation, and their location and size were noted on a similar line drawing of the specimen. This was performed "blindly" without knowledge of the sonographic findings. By comparing the sonographic localization and size of the nodes with the histologic findings, it was possible to determine which nodes had been identified and which had been overlooked on sonograms. Preoperative endosonography had also been performed in 10 patients with rectal carcinoma, and the number of lymph nodes detected with this examination was compared with the results of the in vitro sonographic and histopathologic examinations.

RESULTS

Histologic examination detected 205 lymph nodes in the 20 specimens. Sixty-four (31%) of them were seen on in vitro sonograms (25% of the benign nodes and 56% of the malignant nodes). The overall detection rates were 16% for nodes less than 5 mm in diameter, 56% for nodes between 6 and 10 mm, and 82% for lymph nodes larger than 10 mm. Preoperative rectal endosonography generally showed fewer lymph nodes (average, 1.2 nodes) than what was subsequently found with in vitro sonography (average, 2.7 nodes).

CONCLUSION

Our study suggests that only approximately half of the malignant nodes in patients with colorectal carcinoma are visualized with endosonography of the surgical specimen. Even fewer nodes are visualized on preoperative in vivo examinations. The inability to detect more than 56% of metastatic lymph nodes makes preoperative endosonographic assessment less valuable than hitherto believed. This should be considered if sonographically guided biopsy of lymph nodes is considered as part of preoperative staging.