Does Rate of Growth Differentiate Between Mucinous and Non-Mucinous Pancreatic Cysts?: 172

Purpose: The natural history of pancreatic cysts is not clearly defined. It has been suggested in the surgical and radiologic literature that surveillance with cross-sectional imaging for cysts less than 3 cm in size is adequate management. The goal would be to identify cysts that are increasing in size, warranting early intervention. We sought to identify clinical or radiologic parameters that might predict the presence of mucinous cysts (MC) and, more importantly, if the rate of growth can reliably differentiate a MC from a non-mucinous cyst. Methods: The charts of patients that underwent EUS for a pancreatic cyst from 1999–2006 were reviewed. Each subjects' EUS report was reviewed and the location, size, morphology and results of cyst fluid analysis were obtained. Each cyst was characterized as mucinous, non-mucinous or malignant based on cyst fluid analysis, fluid cytopathology and surgical pathology (when available). Patients with multiple cross-sectional imaging studies prior to EUS were identified and the rate of annual growth of the cystic lesion was calculated. Receiver operating characteristic (ROC) analysis was performed to evaluate the ability of rate of growth to differentiate between mucinous and non-mucinous cysts and between malignant and non-malignant lesions. Results: We identified 194 consecutive patients that underwent EUS for the evaluation of pancreatic cysts at our center. Mean age of subjects was 66 years (±13.7) and 75% were female. The size of the cyst on EUS correlated with the presence of a mucinous cyst (P= 0.02). Age, gender, the presence of symptoms, and location of the cyst in the pancreas did not predict the presence of mucinous lesions. Forty-seven patients (24%) had multiple imaging studies prior to EUS. ROC analysis demonstrated that the rate of growth did not help to distinguish MC from non-mucinous cysts (AUC = 0.61). However, a rate of growth greater than 0.5 cm per year predicted a malignant cyst (ROC AUC = 0.89, sensitivity 83%, specificity 78%). Conclusion: A rate of growth greater than 0.5 cm per year suggests the presence of malignancy in a pancreatic cystic lesion. However, rate of growth by cross-sectional imaging cannot reliably distinguish between mucinous and non-mucinous cysts. Stability in size on cross-sectional imaging does not rule out the possibility of a pre-malignant mucinous cyst. Therefore further evaluation by EUS-FNA of cysts that are amenable to aspiration is warranted.