Positron emission tomography does not add to computed tomography for the diagnosis and staging of pancreatic cancer

BACKGROUND: Positron emission tomography (PET) has been proposed for pancreatic cancer diagnosis and staging. METHODS: 112 patients with suspected pancreatic cancer underwent 18F-fluoro-2-deoxy-D-glucose gamma camera PET and computed tomography (CT), of whom 62 also had laparoscopic ultrasonography and 70 underwent abdominal exploration for potential resection. The final diagnosis was malignancy in 78 and benign disease in 34 patients (25 with chronic pancreatitis). RESULTS: The diagnostic sensitivity and specificity for PET were 73 and 60% compared to 89 and 65% for CT respectively (Cohen's kappa = 0.59). In 30 patients CT was equivocal with cancer in 14 and benign disease in 16. PET correctly diagnosed 13 of these patients (cancer in 6 and benign disease in 7), interpreted 4 as equivocal (cancer in 3 and benign disease in 1) but was incorrect in the remaining 13 patients (cancer in 5 and benign disease in 8). The sensitivity and specificity for detecting small volume metastatic disease were 20 and 94% for CT and 22 and 91% for PET, respectively. CONCLUSION: PET had a similar accuracy to that of CT for imaging pancreatic cancer but it did not provide any additional information in patients with equivocal CT findings and currently would seem of little benefit for the staging of pancreatic cancer.