RESUMO
PURPOSE: To assess diagnostic accuracy of (18)F-FDG PET/CT at 3 months for the detection of local recurrence after radiofrequency ablation (RFA) of lung metastases. METHODS: The PET/CT scan at 3 months was compared with a baseline PET/CT scan from a maximum of 2 months before RFA, with the reference standard as recurrence diagnosed by CT during a 12-month follow-up. Local recurrence was diagnosed on the PET/CT scan if lesional uptake was greater than the mediastinal background. Maximum standardized uptake values (SUVmax) were recorded. ROC curve analysis for SUVmax was performed. Overall survival (OS) and time to local relapse were computed from the date of RFA using the Kaplan-Meier method (www.clinicaltrials.gov: NCT 00382252). RESULTS: Between 2005 and 2009, 89 patients (mean age 65 years) underwent RFA for 115 lung metastases (mean size 16.2 ± 6.9 mm). The median SUVmax before RFA was 5.8 ± 4. PET/CT at 3 months and the reference standard were available in 77 patients and 100 lesions. Accuracy was 66.00% (95% CI 55.85-75.18%), sensitivity 90.91% (95 % CI 58.72-99.77 %), specificity 62.92% (95% CI 52.03-72.93%), PPV 23.26% (95% CI 11.76-38.63%), and NPV 98.25% (95% CI 90.61-99.96%). One-year OS was 94.2% (95% CI 86.6-97.5%) and the probability of being free of local recurrence 1 year after RFA was 84.6% (95% CI 75.0-90.8%). CONCLUSION: The specificity of PET/CT at 3 months is low because of persistent inflammation, especially when the lesion is close to the pleura. This technique is useful for its negative predictive value, but positive findings need to be confirmed by histology before new treatment is planned.