Mo1403 Preoperative Endoscopic Tissue Sampling of Intraductal Papillary Mucinous Neoplasms (IPMNs) With Mural Nodules (MNs)

Abstract

Accurate identification of malignant IPMNs on preoperative imaging is difficult. MNs in IPMNs are associated with dysplasia and therefore have been as an indication for surgical resection. However, the utility of endoscopic sampling of MNs is unknown. We sought to assess the ability of endoscopic tissue sampling (endoscopic ultrasound fine needle aspiration (EUS-FNA) or endoscopic retrograde cholangiopancreatography with cytology brushing (ERCP-Br) to preoperatively diagnose high grade dysplasia (HGD) or invasive cancer in IPMNs. Prospective IPMN surgical database was queried for patients with preoperative MN on imaging between 1999 and 2009 at our tertiary referral center. Additional patients were identified from an EUS database referred to our tertiary referral center. Patients were included only if at least one cross-sectional imaging (CT or MRI) demonstrated MNs in addition to EUS or ERCP. IPMN lesions associated with solid mass were excluded. Clinical, endoscopic, radiographic and pathological data were reviewed. Endoscopy reports were reviewed to ensure sampling was done from the lesion with MN. Between 1999 and 2010, 41 patients (18 females, median age 68, range 52-86) underwent EUS-FNA sampling of lesions with MNs. Thirty-four (83%) had symptoms related to IPMN for a median of 6 months (range 1-84). 38/41 patients also underwent a CT scan, and 34/41 underwent MRI/MRCP. All MNs were described as adherent hypoechoic solid lesions on EUS or as discrete adherent filling defect(s) within the main duct or side branch on EUS. All 41 patients underwent EUS-FNA of the MN within a side branch lesion with MN (n=25, lesion median size 2.7 cm, range 1-5.9 cm) or the main duct (n=16, median duct size 0.9, range 0.5-2.1 cm). 27 underwent ERCP-Br of the main duct segment where MNs were noted. Fifteen patients underwent both EUS-FNA and ERCP-Br. Cytopathology was insufficient or non-diagnostic in 14 (34%), showed non-dysplastic papillary mucinous epithelium in 3 (7%), atypical or dysplastic epithelium in 19 (46%) and malignancy in 5 (12%). Twenty four patients underwent surgical resection. Final surgical pathology confirmed IPMN in all patients, with 12 (50%) adenoma/low grade dysplasia, 4 (17%) high grade dysplasia/carcinoma in situ, and 8 (33%) with invasive carcinoma. There was a significant association between presence of MNs and malignancy on surgical pathology (p=0.03). In the group of patients with HGD or invasive malignancy on surgical pathology (n=12), preoperative endoscopic tissue sampling was 57% sensitive, 80% specific with 80% positive predictive value and 57% negative predictive value in diagnosing malignancy. MNs on EUS or ERCP is associated with malignancy in IPMN. Endoscopic tissue sampling of lesions with MNs appears to help predict malignancy and should be attempted whenever feasible.