Abstract
BACKGROUND: Most patients undergo follow-up after surgery for cancers of the gastro-oesophageal junction, stomach or pancreas, but data to support which modalities to use and the frequency of investigation are limited. METHODS: Patients in the EUFURO study were randomized to either visits to the outpatient clinic at 3, 6, 9, 12, 18, and 24 months after surgery (standard), or to the addition of [18F]fluorodeoxyglucose (FDG) PET-CT and endoscopic ultrasonography (EUS) with guided fine-needle aspiration biopsy to clinical assessments (intervention). Data from the intervention arm were used to analyse the diagnostic performance of endosonography or [18F]FDG PET-CT in detecting recurrences. RESULTS: During the scheduled follow-up, 42 of 89 patients developed recurrence; PET-CT and EUS in combination detected 38 of these recurrences. EUS detected 23 of the 42 patients with recurrent disease during follow-up and correctly diagnosed 17 of 19 locoregional recurrences. EUS was able to detect isolated locoregional recurrence in 11 of 13 patients. In five patients, EUS was false-positive for isolated locoregional recurrence owing to missed distant metastases. PET-CT detected locoregional recurrence in only 12 of 19 patients, and isolated locoregional recurrence in only 7 of 13. False-positive PET-CT results in 23 patients led to a total of 44 futile procedures. CONCLUSION: Accuracy in detecting recurrences by concomitant use of PET-CT and EUS was high (90 per cent). PET-CT had moderate to high sensitivity for overall recurrence detection, but low specificity. EUS was superior to PET-CT in the detection of locoregional and isolated locoregional recurrences.