This study was performed to verify reports of the decreased accuracy of endorectal ultrasonography (EUS) in preoperative staging of rectal cancer, and to compare the efficacy of 3-dimensional (3D) EUS with that of 2-dimensional (2D) EUS and computed tomography (CT).

Eighty-six consecutive rectal cancer patients undergoing curative surgery were evaluated by 2D EUS, 3D EUS, and CT scan.

The accuracy in T-staging was 78% for 3D EUS, 69% for 2D EUS, and 57% for CT (P < .001–.002), whereas the accuracy in evaluating lymph node metastases was 65%, 56%, and 53%, respectively (P < .001–.006). Examiner errors were the most frequent cause of misinterpretation, occurring in 47% of 2D EUS examinations and in 65% of 3D EUS examinations. By eliminating examiner errors, the accuracy rates in T-staging and lymph node evaluation could be improved to 88% and 76%, respectively, for 2D EUS, and to 91% and 90%, respectively, for 3D EUS. Conical protrusions along the deep tumor border on 3D images were correlated closely with infiltration grade, advanced T-stage, and lymph node metastasis.

We found that 3D EUS showed greater accuracy than 2D EUS or CT in rectal cancer staging and lymph node metastases. Concrete 3D images based on tumor biology appear to provide more accurate information on tumor progression.