Limited value of positron emission tomography in treatment of pancreatic cancer: surgeon's view.

We aimed to assess, from a surgical point of view, the value of positron emission tomography (PET) in the routine preoperative diagnostic evaluation of a pancreatic mass. In particular, we were interested in PET's ability to distinguish a malignancy from a nonmalignant process and its impact on surgical decision making. We documented prospectively the results of preoperative ultrasonography, computed tomography (CT), endoscopic retrograde cholangiopancreatography (ERCP), and PET in terms of a correct diagnosis of malignancy in 103 patients. All patients underwent surgery at our department and had a complete histologic workup. We then related the preoperative findings to the intraoperative and histologic findings and analyzed the advantages and limitations of PET in comparison with conventional diagnostic techniques. Both CT and PET showed high sensitivities (82% and 84%, respectively) but rather low specificities (61%). ERCP showed good specificity at 83% but weak sensitivity at 46%. The positive predictive value was higher than 80% for all methods, whereas the negative predictive value was around 60% at best. Both the CT and the PET diagnoses were wrong in 15% of the cases. In all cases with a false CT diagnosis, the results of the other conventional examinations provided sufficient evidence to indicate the need for surgical intervention. In all cases where PET gave misleading results, CT findings indicated the need for surgery. PET overlooked pT1 cancers in three of the patients. PET does not reliably prove or exclude malignancy in situations where conventional diagnostic procedures leave doubt as to the nature of a pancreatic mass. PET does introduce a new procedure-related spectrum of misinterpretations to the diagnostic process.