Successful selection of patients with rectal cancer for local excision requires accurate preoperative lymph node staging. Although endorectal ultrasound is capable of detecting locally advanced disease, its ability to correctly identify nodal metastases in early rectal lesions is less well described. This study examines the accuracy of endorectal ultrasound in determining nodal stage based on depth of penetration of the primary lesion (T stage). Between 1998 and 2003, endorectal ultrasound was performed on 938 consecutive patients; 134 had biopsy-proven rectal cancers and were treated with radical resection, without neoadjuvant therapy. Lymph node metastases were measured pathologically and correlated with endorectal ultrasound and clinicopathologic features. Accuracy and specificity of endorectal ultrasound nodal staging was determined. The overall accuracy of endorectal ultrasound nodal staging for the study cohort was 70 percent, with a 16 percent false-positive rate and 14 percent false-negative rate. Endorectal ultrasound was more likely to overlook small metastatic lymph node deposits. The size of lymph node metastasis and accuracy of endorectal ultrasound nodal staging was related to T stage. The specificity of endorectal ultrasound nodal staging, or the ability to identify patients who were node-negative, was dependent on T stage. Early rectal lesions are more likely to have lymph node micrometastases not detected by endorectal ultrasound. The ability of endorectal ultrasound to correctly identify patients without lymph node metastasis is dependent on the T stage of the primary lesion. The limitations of endorectal ultrasound in accurately staging nodal disease in early rectal lesions may, in part, explain the relatively high recurrence rates seen after local excision.
(C) The ASCRS 2007