RESUMEN
To compare the visual and semiquantitative analysis of carbon-11-methionine (C-MET) PET/computed tomography (CT) images in patients with primary brain tumors and suspected recurrence, persistence, or necrotic post-therapeutic changes. A total of 41 consecutive C-MET-PET/CT scans on 35 (21 men, mean age 44.1±16.6 years) patients were requested for MRI suspicion of recurrent or persistent primary tumor after therapy. The C-MET PET/CT were obtained 20 min after an intravenous injection of 555-740 MBq (15-20 mCi) of C-MET. Both visual and semiquantitative evaluations were performed comparing C-MET uptake between suspicious areas and different lesion/normal-to-background ratios. The final diagnosis was established by histological examination in 12 cases and clinical and MRI follow-up in 29 cases. Visual analyses were positive in 27 (63.4%) and negative in 14 (36.6%) of the C-MET PET/CT. The sensitivity was 83.9%, specificity was 90.0%, positive predictive value was 96.3%, negative predictive value was 64.3% and accuracy was 71.4%. For the semiquantitative analysis, all the lesion/normal-to-background ratios could differentiate between tumor and nontumor (P<0.001), the lesion/contralateral parenchyma (L/CP) maximum standardized uptake value (SUVmax) being the index with the highest area under de curve (0.938). Applying an L/CP SUVmax index of 1.21, the sensitivity was 89.3%, specificity was 90.0%, positive predictive value was 96.1%, negative predictive value was 75%, and accuracy was 82.9%. C-MET-PET/CT was a useful technique to differentiate post-therapeutic changes from tumor presence in treated patients with brain neoplasm in whom cerebral MRI is nonconclusive, showing a high diagnostic performance. Our results showed only slight differences between visual analysis methods and the L/CP SUVmax ratio, the best of the semiquantitative methods.