Objective: This study was performed to assess the accuracy of colonoscopic endoanal ultrasound scanning (EUS) in the selection of patients with rectal neoplasia suitable for local excision by transanal endoscopic microsurgery (TEM). Our policy is to offer TEM to patients with premalignant (T0) lesions or with T1 tumours that have early disease.
Patients and methods: Data were collected prospectively on all patients undergoing EUS for the assessment of rectal neoplasia at our institution over a six-year period. A colonoscopic EUS probe was used to determine whether the tumour breached the muscularis propria (the interface between T1 and T2 disease). Subsequently patients underwent surgical resection, including TEM for those with T0/1 disease. The preoperative stage predicted by EUS (uT stage) was compared to the postoperative histopathological stage of the resected specimens (pT stage).
Results: One hundred and fifty-six EUS examinations were evaluated. Sixty-two patients went on to have TEM whilst the remaining 94 had another form of surgery. Of the 62 patients undergoing TEM, 3 were overstaged on EUS. No patients were understaged, giving an accuracy of 95%. Of the 94 patients undergoing an alternative procedure, 5 were overstaged on EUS as having T2 tumours when in fact their histology was T1. Accuracy of EUS at predicting more advanced disease fell to 89%, giving an overall accuracy of 92%.
Conclusions: EUS is accurate at predicting T0/1 vs T2 disease in our institution, and we believe that it is a useful modality in assessing patient suitability for local excision.