Role of [18F] FDG PET/CT for distinguishing benign and malignant pancreatic lesions

Objectives: To assess the ability of [18F]FDG PET/CT for distinguishing malignant from benign lesions of the pancreas and to evaluate the prognostic value of SUVmax in solid and cystic lesions. Methods: We reviewed 94 [18F]FDG PET/CT examinations performed in three different Institutions in patients with radiological suspicion of malignancy. We studied a total of 96 histologically confirmed lesions (39 malignant, 57 benign); 2 patients had 2 different pancreatic lesions. Follow-up was available for 85/94 patients. Results: The mean SUVmax of malignant lesions (5.6±3, range 1.3-17) was significantly higher than that of benign lesions (1.9±0.9 range 0.5-5). With a cut-off SUVmax <=2.4 for benign lesions, overall sensitivity and specificity of PET/CT for distinguishing malignant from benign pancreatic lesions were 93% and 87%, respectively. PPV was 91.5% and NPV was 89.5%. For cystic lesions (n=49), SUVmax <=2.1 as cut-off resulted in 80% sensitivity and 79.5% specificity, with 50% PPV and 94% NPV. There was a significant difference in SUVmax between malignant lesions (4.2±2.7, range 1.3-9) and benign lesions (1.9±1, range 0.5-5). 45/85 patients with long-term follow-up had pancreatic cancer. Analyzing SUVmax as prognostic factor for survival, Kaplan-Meier curves discriminated 3 subgroups with significantly different survivals: 1) 38 patients with SUVmax <=2.4 and longer survival (no deaths, median follow-up 42.5 months); 2) 22 patients with SUVmax 2.5-4.5 (4 deaths, median follow-up 21 months); and 3) 27 patients with SUVmax >4.6 (17 deaths, median survival 25 months, median follow-up 19 months). Conclusions: [18F]FDG PET/CT is useful to characterize suspicious pancreatic masses and allows differentiation of benign from malignant lesions. Moreover, SUVmax of pancreatic tumors correlates with the overall survival of patients.