AbstractEndoscopic ultrasound‐guided fine‐needle aspiration (EUS‐FNA) is the preferred method for diagnosing pancreatic masses. While the diagnostic success of EUS‐FNA is widely accepted, the actual performance of EUS‐FNA is not known. This study sought to define the EUS‐FNA accuracy compared with the gold standard, surgically resected specimens. The study was a single institution, retrospective, and chart review of patients with surgically resected pancreatic specimens from 2005 to 2015 with a preoperative EUS‐FNA or biliary brushing. Cytological reports were organized from least concerning (i.e., low chance of malignancy) to most concerning (high chance of malignancy) into eight cytologic categories. We identified 741 cytologic cases: 530 EUS‐FNA and 211 endoscopic brushings. For EUS‐FNA samples, 62.5% of “benign” samples proved to be “benign” on surgical pathology. A cytologic diagnosis of “suspicious for malignancy” or “positive for malignancy” were concordant with a cancer diagnosis on surgical pathology 93.3% and 98.0% of cases, respectively. EUS‐FNA proved to be highly reliable at diagnosing malignancy for cytologic samples that were “suspicious” or “positive” for malignancy. Paired with supportive clinical data, these interpretations may be used to justify cancer treatment.