Immunohistochemical Analysis Increases the Yield of Nondiagnostic EUS-Guided Pancreatic FNA

Abstract

Background and Aims: Endoscopic ultrasound (EUS)-guided fine needle aspiration (FNA) has emerged as a main diagnostic modality of pancreatic lesions. FNA pathology results are usually reported as benign, malignant or atypical cells. Atypical cells diagnosis leads to repeated FNAs and ultimately to delay in care. The aim of the study was to evaluate the incidence and long-term outcome of atypical cells diagnosis by EUS-FNA of solid pancreatic masses done in two institutional settings (tertiary academic center and community hospital). Secondly, the aim was to determine if the diagnostic yield can be improved by immunohistochemical staining for various tumor markers. Methods: 225 solid pancreatic mass EUS-FNAs from 2001-2008 were retrospectively reviewed; 151 were done at a tertiary academic medical center (group 1) and 74, at a community hospital (group 2). All EUS-FNAs were performed by two endoscopists (VE and DC) with extensive EUS experience. All pathological diagnosis in group 1 was made by two pathologists (RV and SF); in group 2, one cytopathologist reviewed all EUS-FNAs and the atypical specimens were sent out for second opinion. The final diagnosis was based on clinical follow-up and/or pathology specimens obtained endoscopically, surgically or post-mortem. Results: In group 1, cytology was 49 atypical (32%), 63 malignant (42%), 28 benign (19%), and 11 insufficient for diagnosis (7%). Of the atypicals in this group, 33 (67%) were later determined to be malignant. In group 2, 29 (39%) had atypical diagnosis, 25 (34%) were malignant, and 20 (27%) benign. Of the 29 atypicals, 17 (59%) were diagnosed malignant on long-term follow up. Sensitivity for malignancy in group 1 was 62% if the atypicals were considered benign, and 96% if atypicals were considered malignant; specificity was 100% and 77%, respectively. For group 2, sensitivity was 52% if the atypicals were considered benign, and 88% if malignant, with specificity of 100% and 54%, respectively. Immunohistochemical staining for CEA, p53, MIB, CA 19-9, CAM 5.2 was performed on 69 FNAs from group 1. Adding at least one immunohistochemical marker to routine cytology increased the sensitivity from 62% to 85%. In 24 atypical specimens (from the total 49), diagnosis was changed to malignant after the immunohistochemical analysis. Conclusions: Our results suggest that atypical cell diagnosis is fairly common in both academic and community settings. Immunohistochemical staining improves the yield of EUS-FNA of solid pancreatic masses and should be considered for routine use. The majority of atypical FNAs were finally diagnosed malignant, thus necessitating a close follow up and further testing.