Pretest prediction and diagnosis of metastatic lesions to the pancreas by endoscopic ultrasound‐guided fine needle aspiration

Background and Aim Early diagnosis of solid pancreatic lesions ( SPLs ) enables prompt treatment. The study aims to identify factors differentiating metastatic lesion to the pancreas ( PMET ) from pancreatic ductal adenocarcinoma ( PDAC ) and pancreatic neuroendocrine tumors ( PNETs ). Methods This is a retrospective study at a tertiary cancer center. Consecutive patients referred for endoscopic ultrasound ( EUS ) of SPLs from 2004 to 2011 were reviewed. The main outcomes were pre‐ EUS‐FNA (endoscopic ultrasound‐guided fine needle aspiration) predictors and diagnostic accuracy of EUS‐FNA for PMETs . Results Among a total of 1108 EUS ‐ FNAs for pancreatic lesions, 672 patients had neoplastic SPLs ( PMETs  = 53; PDACs  = 528, PNETs  = 91). The sensitivity, specificity, positive predictive value, and accuracy of EUS‐FNA for diagnosis of PMETs were 84.9%, 100%, 100%, and 98.8%, respectively. The mean number of EUS‐FNA passes for diagnosis of PMET was 3.1 per patient. For each endosonographer, preceding 3‐year EUS volume (mean/year) significantly correlated with fewer needle passes (r s [−0.30], P  = 0.03). The most common PMET was renal cell carcinoma. Stratified multivariate analyses were performed. Compared with patients with PDACs , PMETs were more common in men (odds ratio [ OR ] = 2.0; 95% confidence interval [ CI ] = 1.0–4.0); located in the pancreatic tail ( OR  = 2.4; 95% CI  = 1.1–5.2); and were less likely with increasing age ( OR  = 0.95; 95% CI  = 0.92–0.99), presence of major symptoms (abdomen pain/diarrhea/weight loss; OR  = 0.2; 95% CI  = 0.1–0.4), elevated bilirubin ( OR  = 0.3; 95% CI  = 0.13–0.69), and imaging evidence of arterial invasion ( OR  = 0.15; 95% CI  = 0.03–0.67). Compared with PNETs , PMETs were more common with increase age ( OR  = 1.05; 95% CI  = 1.02–1.08) and increasing lesion size ( OR  = 1.03; 95% CI  = 1.0–1.1), and were less likely in patients with diabetes ( OR  = 0.34; 95% CI  = 0.11–0.99). Conclusion Among the largest numbers of neoplastic SPLs evaluated at a single center, pre‐test features reliably characterize, and EUS‐FNA provides a highly specific diagnosis of PMETs .