Endoluminal ultrasography for rectal tumors: efficacy, sources of error and limitations.

Objectives: endorectal ultrasound (EUS) is currently accepted as the best technique for the preoperative study of patients with rectal tumors, and surgical decisions depend increasingly on EUS staging. The main pitfalls in staging rectal tumors are over- or understaging as well as errors in imaging lymph nodes. Being aware of such errors and their causes may help to improve the overall results. The aim of the present study was to evaluate the accuracy of EUS in staging rectal neoplasms, and to study potential sources of error.
Methods: from May 1996 to December 1998, 120 patients with rectal tumors were studied preoperatively by EUS. The uTNM classification described by Hildebrandt and coworkers was used. The EUS findings were compared prospectively with the results of pathological examination. When there was no correlation, both the specimen and the EUS findings were carefully reviewed to look for potential sources of error.
Results: 41 out of 120 patients were classified as uT1, 10 as uT2, 60 as uT3 and 9 out of 120 as uT4. 31 patients had positive lymph nodes (uN1). On comparing these data with the results of the pathological report, we found 90% accuracy in staging rectal wall penetration, and 70% accuracy in the diagnosis of lymph nodes. Errors were due basically to technical problems, characteristics of the tumor itself, and difficulties in staging lymph nodes.
Conclusions: it is important to identify the potential source of errors as well as the current limitations of EUS to improve the overall results with this technique.