Erschienen in:
01.10.2012 | Pancreatic Tumors
Searching for Indicators of Malignancy in Pancreatic Intraductal Papillary Mucinous Neoplasms: The Value of 18FDG–PET Confirmed
verfasst von:
Gian Luca Baiocchi, Francesco Bertagna, Federico Gheza, Luigi Grazioli, David Calanducci, Raffaele Giubbini, Nazario Portolani, Stefano M. Giulini
Erschienen in:
Annals of Surgical Oncology
|
Ausgabe 11/2012
Einloggen, um Zugang zu erhalten
Abstract
Background
Malignancy in intraductal papillary mucinous neoplasms (IPMN) of the pancreas may be predicted on the basis of a number of clinical and radiologic features, which have raised sensitivity but result in a specificity as low as 20–50%. We sought to confirm the additional value of 18F-18-fluorodeoxyglucose–positron emission tomography (18FDG–PET) in diagnostic accuracy of imaging-based IPMN malignancy assessment.
Methods
This prospective uncontrolled case series contained 44 patients with IPMN undergoing comprehensive diagnostic evaluation, including magnetic resonance cholangiopancreatography and 18FDG–PET. Average follow-up time was 39.3 months (range 3–97 months). Diagnostic performance regarding the diagnosis of malignancy was evaluated for the classic preoperative assessment, including clinical signs, CA 19-9, imaging (computed tomography and magnetic resonance cholangiopancreatography), and International Consensus Guidelines criteria, as well as 18FDG–PET scan.
Results
Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 100, 22, 32, 100, and 43%, and 83, 100, 100, 94, and 96%, respectively, for comprehensive assessment without and with 18FDG–PET [maximum standardized uptake value (SUVmax) cutoff of 2.5 MBq]. Elevated CA 19-9 values and positive PET scan were the only independent prognostic factors for malignancy (odds ratio 2.11, 95% confidence interval 1.15–2.74 and 5.49, 95% confidence interval 3.98–21.44, respectively).
Conclusions
18FDG–PET is useful for detection of malignancy in IPMN, improving the differential diagnosis with benign cases by functional data. The choice of SUVmax cutoff should maximize specificity.