Background: Tumor recurrence is almost the rule in patients resected for pancreatic cancer, and is also very frequent after resection for periampullary cancer. CT-scan is still the most used imaging modality but it's accuracy is low in this setting. 18FDG PET was seldom used in the postoperative evaluation of tumor relapse, although promising results are reported in small series of patients.
Methods: Between 1998 and 2005, 69 patients resected for pancreatic (53) or periampullary (16) cancer were prospectively evaluated with serial CA 19-9 determinations, CT-scan every 3 months for two years, and 18FDG PET within one year from surgery with a dedicated Siemens machine (ECAT EXACT 47, Erlangen, Germany). A Standardized Uptake Value > 2.5 was considered positive. Actuarial survival curves were constructed using the Kaplan-Meier method and comparison was made with Log-rank test.
Results: 40/53 pancreatic, 5/6 biliary, 3/6 ampullary, and 2/4 duodenal cancers relapsed. Sensitivity of 18FDG PET and CT scan in detecting tumor's relapse was 98% (49/50) and 62% (31/50), respectively. Group 1: 9/19 recurrences detected only by 18FDG PET underwent resection of tumor's relapse (6 potentially radical and 3 non-radical), ten chemotherapy and/or radiotherapy. Group 2: In 31 patients both 18FDG PET and CT scan were positive. All patients underwent chemotherapy and/or radiotherapy. The CT scan positive 18FDG PET negative lever metastasis was resected. The patient died 10 mo later.Mean and median overall survival [and residual survival time from the detection of recurrence] were significantly longer (10 and 5 months respectively) in the group 1 than in the Group 2 (p<0.05), while the disease free survival was the same. Nine patients are still alive after tumor relapse: 6 of the Group 1 (18-54 months), and 3 for the Group 2 (10-22 months).Group 3: 19 patients are still disease free: 5 had increased CA 19-9 levels and 4 equivocal CT findings during follow-up. 18FDG PET was true negative, and discovered a second primary tumor that was successfully resected (1 lung and 2 colon cancers). Overall, 18FDG PET modified the treatment in 16/69 patients (23%): 13/19 in Group 1, 0/31 in Group 2, and 3/18 in Group 3, and allowed a wait and see policy, confirmed by further follow-up, in 9 Group 3 patients (13%).
Conclusions: 18FDG PET is very sensitive in detecting recurrent pancreatic and periampullary cancer. The early detection of tumor's relapse, allowed us to perform potentially radical surgery in 6/19 of Group 1 patients, and in 3/18 Group 3 patients with a second primary cancer. A significantly longer survival can be achieved in these patients.