Downstaging after chemoradiotherapy for locally advanced rectal cancer: is there more (tumor) than meets the eye?

PURPOSE

Preoperative chemoradiotherapy can lead to pathologic complete response of rectal cancer. This study was designed to determine the relationship between postchemoradiotherapy pathologic T stage (ypT stage) and nodal metastases and to evaluate whether pathologic complete response of the primary tumor results in sterilization of mesorectal lymph nodes.

METHODS

Clinicopathological data from 1997 to 2007 of a prospectively maintained colorectal cancer database were examined. Inclusion criteria were patients with extraperitoneal rectal cancer who underwent preoperative chemoradiotherapy and subsequent radical resection. Statistical analysis was performed by use of Kruskall-Wallis and Wilcoxon rank-sum tests.

RESULTS

Two hundred forty-two patients were identified (73.1% male, median age, 57 y (range, 36-85 y)). Data regarding preoperative chemoradiotherapy were available for 177 patients (73.1%). The median dose of radiotherapy was 5040 cGy (3060-6100 cGy). The mean preoperative radiotherapy dose and interval between chemoradiotherapy and surgery are similar when stratified by ypT stage (P = .55 and P = .72, respectively). Low anterior resection was performed in 174 patients (71.6%), and the remainder underwent abdominoperineal resection. A mural pathologic complete response was achieved in 62 patients (25.6%). In this pathologic complete-response group, positive lymph nodes were found in 2 patients (3.2%). The rate of metastatic lymph nodes increased as ypT stage increased (ypT1 = 11.1%, ypT2 = 29.2%, ypT3 = 37.3%).

CONCLUSION

Patients with a mural pathologic complete response have a low rate of positive lymph nodes. These findings may have implications for the management strategies of these patients, including the use of local resection or a watch-and-wait policy. When the response to chemoradiotherapy is not complete, radical surgery should remain the treatment based on high rates of lymph node involvement.