Can EUS alone differentiate between malignant and benign cystic lesions of the pancreas?

Abstract

Between January 1995 and August 1999, the authors evaluated 98 patients with pancreatic cystic lesions, in whom conventional cross-sectional imaging modalities were nondiagnostic ( i.e., the interpreting radiologist recommended endoscopic ultrasound [EUS] for additional evaluation), and in whom EUS was performed. Surgical and/or pathological correlation was documented in 48 patients. The resultant EUS images were evaluated by two experienced endosonographers (reviewers A and B) who were blinded to surgical and pathological reports as well as the other endosonographer’s interpretation. Five characteristics of the EUS images were recorded for each individual: 1) presence or absence of a wall; 2) solid component; 3) septae; 4) lymphadenopathy; and 5) number of cysts. These features were then correlated with surgical and pathological attributes to determine their utility in distinguishing between benign and malignant lesions. Reviewer A noted that the EUS presence of a solid component was the only statistically significant harbinger of malignancy. However, 61% of benign lesions also harbored a solid component on EUS. Reviewer B was unable to identify any significant EUS predictors of malignancy. Overall, when both reviewers’ results were combined, a solid component was not determined to be a statistically significant predictor of malignancy. The authors concluded that the EUS features that were evaluated in this study could not accurately distinguish between benign and malignant pancreatic cystic lesions in patients who have undergone a standard cross-sectional imaging modality that is nondiagnostic.