MDCT imaging findings for characterization pancreatic cystic lesion: differentiation between benign and malignant pattern.

OBJECTIVE: The purpose of the present study was to determine the characteristic imaging findings for diagnosis of the benign or malignant nature of pancreatic cystic lesions by 64 slice multidetector computed tomography (MDCT) for further investigation or treatment. MATERIAL AND METHOD: A retrospective study was performed in 33 patients with pancreatic cystic lesion who underwent 64 slice MDCT to characterize and establish the diagnosis. The patients were enrolled between January 2004 and March 2009. The MDCT scan of abdomen from patients with pancreatic cystic lesion was retrospectively evaluated by two gastrointestinal radiologists without knowledge of the patient's history, clinical data, and final diagnosis. Sensitivity, specificity, accuracy, PPV, and NPV of CT scan for discriminate benign and malignant pancreatic cystic lesion were calculated. Weighted-kappa statistics were used to estimate agreement between readers. RESULTS: Thirty-three patients with pancreatic cystic lesion were included in the present study. Benign pancreatic cystic lesion are pancreatic pseudocyst (n = 16), serous cystadenoma (n = 4) and benign intraductal papillary mucinous neoplasms IPMN (n = 2). Premalignant and malignant pancreatic cystic tumor are mucinous cystic tumor (n = 5) and include solid pseudopapillary epithelial neoplasm of pancreas (n = 3) and malignant intraductal papillary mucinous neoplasms IPMN (n = 3). The final diagnosis was established either by pathological diagnosis (20 patients), EUS with FNA (11 patients) or F/U clinical and imaging findings (2 patients). Pseudopod sign is a helpful finding for diagnosis pancreatic pseudocyst (3/16) and not observed in other types pancreatic cystic lesions. Type of cyst (unilocular, multilocular microcystic, multilocular macrocystic and solid component) is the only finding that has statistical significance for differentiating between the benign and malignant groups (p < 0.005). The overall sensitivity, specificity and accuracy of 64-slice MDCT to discriminate benign and malignant pancreatic cystic lesion were 36.3%, 100% and 78.8%, respectively. In addition, the PPV and NPV were 100% and 75.9%, respectively. CONCLUSION: The 64 sliced MDCT is a noninvasive method that can be used for discriminating between benign and malignant pancreatic cystic lesions, which is an important finding for further investigation and proper treatment.