The learning curve for endorectal ultrasonography in rectal cancer staging

Abstract

Endorectal ultrasound (ERUS) is an emerging technique for preoperative rectal cancer staging. It is an operator-dependent examination with accuracy closely related to endosonographer experience. In this study, we prospectively analyzed our results of ERUS staging for rectal cancer, aiming to determine its accuracy and to define the learning curve of the procedure. Between July 2007 and August 2009, consecutive patients with rectal cancer were recruited for preoperative ERUS staging performed by a single colorectal surgeon. We compared results of ERUS tumor (uT) and nodal (uN) staging with pathological staging of surgical specimens in patients who had surgery without neoadjuvant chemoradiation. To evaluate the learning-curve effect on ERUS, patients were divided into two equal halves for analysis (early group and late group). In the 26-month study period, 50 patients (36 males) with median age of 67 years (range 47-89 years) underwent ERUS staging. The overall accuracy rates of uT and uN staging were 86 and 66%. For uT staging, 10% of tumors were overstaged and 4% were understaged. For uN staging, 22% of patients were overstaged and 12% were understaged. With experience accumulation from early group to late group, accuracy improvement was observed in uN staging (52 vs. 80%, P = 0.037), while the accuracy rate remained consistently high in uT staging (84 vs. 88%, P = 1.0). ERUS was accurate in preoperative staging of rectal cancer. It was an easy-to-learn procedure for accurate tumor staging, but considerable experience was required to attain accuracy for nodal staging.