Are endoscopic procedures able to predict the benignity of ampullary tumors?

Background Malignant ampullary tumors (AT) require pancreatico-duodenectomy (PD) for curative treatment whereas benign AT can be appropriately treated by local resection. This study evaluated the accuracy of endoscopic exploration combining side-viewing duodenoscopy (SVD) with forceps biopsies, endoscopic sphincterotomy (ES), and endoscopic ultrasonography (EUS) to distinguish benign AT from malignant one. Patients and Methods Twenty-six patients with AT had preoperative SVD with forceps biopsies, including 9 with ES, and EUS. Nodal status was evaluated by EUS in all patients, but could not evaluate parietal spread in 6 in whom ES was previously done. Results of endoscopic examination were compared with pathologic examination after resection (2 local excisions for 2 benign AT, and 24 PD for 20 malignant and 4 benign AT). Results: Accuracy of histologic examination of the 26 biopsies of the papilla was 69%. After ES, accuracy of intra-ampullary biopsies was 77%. The EUS had a 75% accuracy for the parietal spread (tumor limited to ampulla or not) and a 69% accuracy for the nodal status. In 11 patients, all explorations were consistent with a benign lesion, but 6 of these patients had a histologically proven malignancy, including one with nodal metastases and two with duodenal involvement. Conclusions: Side-viewing duodenoscopy with biopsies, even after ES, combined with EUS is not accurate enough to preoperatively ensure that an AT is benign. Indication for a local resection based on these explorations alone is not safe.