Accuracy of endoscopic ultrasound to assess tumor response after neoadjuvant treatment in rectal cancer: can we trust the findings?

BACKGROUND

The finding that some rectal cancers respond to neoadjuvant chemoradiation is broadening new surgical options for the treatment of some of these tumors that, until now, required a total mesorectal excision. Nevertheless, a fine match between clinical and pathological response is required when planning conservative surgical approaches.

OBJECTIVE

This study aims to prospectively validate the use of endoscopic ultrasound as a predictor of clinical and pathological tumor response in patients with locally advanced rectal cancer.

DESIGN

: This is an observational study of a cohort of patients undergoing chemoradiation followed by surgery.

SETTINGS

This study was conducted at a tertiary medical center.

PATIENTS

A total of 235 consecutive patients who underwent chemoradiation followed by surgery at a single institution during a 7-year period were included.

MAIN OUTCOME MEASURES

All tumors were staged and restaged at 4 to 6 weeks after neoadjuvant treatment. Downsizing and downstaging were calculated between the initial and posttreatment measures and correlated to the pathological stage. The accuracy of endoscopic ultrasound to predict response was determined.

RESULTS

Findings after chemoradiation showed T-downstaging in 54 patients (23%) and N-downstaging in 110 (47%). Overstaging occurred in 88 (37%) patients and was more commonly observed than understaging (21 patients; 9%). Related to the pathological report, endoscopic ultrasound correctly matched the T stage in 54% and the N stage in 75% of tumors. Sensitivity, specificity, and positive and negative predictive values to predict nodal involvement were 39%, 91%, 67%, and 76%. Accuracy was not influenced by such factors as age, distance of the tumor from the anal verge, or time to surgery.

LIMITATIONS

This study was limited by the lack of comparison with other imaging methods.

CONCLUSIONS

Endoscopic ultrasound allows prediction of involved lymph nodes in 75% of the cases; however, 1 in 5 patients are missclassified as uN0 after neoadjuvant treatment. In our point of view, this percentage is too high to rely only on this diagnostic modality to support a "wait and see" approach.