EUS-guided cytologic diagnosis of pancreatic endocrine tumors: Immunocytochemical features and the impact of on-site cytopathological assessment

Introduction/Aims: Pancreatic Endocrine Tumors (PETs) are rare. Cytologic studies are scarce, in literature there are till now about 440 cases (15 studies from 1999 to 2010) diagnosed by means of EUS-FNA. In less than 200 a cytopathologist was present during the procedure for adequacy assessment on the aspirated material. Our aim is to report our single center experience in detecting and diagnosing PETs by means of EUS-FNA and to describe their cytopathologic and immunocytochemical features. Secondary aim is to outline the impact of cytopathological on-site assessment on the EUS-FNA accuracy in preoperative PETs diagnosis.

Methods: In this retrospective cohort study we reviewed clinical data, EUS findings, cytological and immunocytochemical features of all cases of PETs diagnosed by means of EUS-FNA between August 2003 and May 2010. Part of the aspirated material was used for preparing direct smears by the cytopathologist, partly was fixed in 95% alcohol and subsequently stained with haematoxylin-eosin for Rapid On-Site Evaluation (ROSE) and partly sent for cell block preparations.

Results: We report data on 41 pts (21M, 20F) aged between 35 and 77 years. Mean size of the tumors was 2.8cm (range 0.7–10.3cm); 18/41 lesions were localized in the pancreatic head. The adequacy assessment of the aspirated material on direct smear has been obtained in 35/41 cases (85%). Cytological features mostly pathognomonic for PETs diagnosis were the presence of a variously cellular sample, represented by an uniform cell population, small/medium-size scarcely cohesive cells, with abundant cytoplasm and eccentric nuclei with granular chromatin. Immunocytochemical stains were performed on cell block; antibodies against Chromogranin A and Synaptophysin confirmed neuroendocrine differentiation in all cases. Data about proliferative activity of the tumors were obtained by Ki-67 immunostaining in all samples with high cellularity. Surgical resection was performed in 24/41 patients. Histological analysis confirmed the cytological diagnosis in all cases.

Conclusions: Our results confirm that EUS-FNA is a valuable method in the detection and diagnosis of PETs even of small size (biopsy on a 7mm lesion was performed); PETs possess evident cytomorphological features on FNA samples and all adequate cases showed these features on ROSE: in most cases the cytopathologist could reach a correct diagnosis. When needed, it has been possible to study the hormone producing capability of the tumor and its biological behaviour and receptor expression on aspirated cytologic material (all these data have diagnostic but also prognostic significance). Cytologic on-site assessment can increase the performances of EUS-FNA in the diagnosis of PETs and can have a significant impact on diagnostic accuracy and patient management.