Objective: The aim of this study was to evaluate the ability of endoscopic ultrasound (EUS) alone to predict and differentiate malignant from benign cystic lesions of the pancreas.
Methods: From January, 1995, to August, 1999, 98 cases of pancreatic cystic lesions were evaluated by EUS; all of these were originally imaged by cross-sectional modalities that were not diagnostic. Among these, surgical/pathological correlation was available in 48 patients. The original endosonographic images were reviewed by two endosonographers who were blinded to each other's interpretation and to the surgical and pathological interpretation. The EUS images were assessed for the presence or absence of the following characteristics: 1) wall, 2) solid component, 3) septae, 4) lymphadenopathy, and 5) number of cysts. These characteristics were then correlated with the surgical and pathological findings and were assessed to determine if any were predictors of the lesion being benign or malignant.
Results: For reviewer A, the presence of a solid component by EUS was the only statistically significant predictor of malignancy (odds ratio = 4.73, 95% CI = 1.13-19.68, p = 0.03). However, 61% of patients with benign lesions were also interpreted by EUS to have a solid component. For reviewer B, none of the features were found to be significant predictors of a malignant lesion. When the results of both reviewers were combined, the presence of a solid component was not found to be a statistically significant predictor of malignancy (odds ratio = 1.046, 95% CI = 0.99-1.09, p = 0.07).
Conclusion: Endosonographic features cannot reliably differentiate between benign and malignant cystic lesions of the pancreas after a nondiagnostic cross-sectional modality.