Abstract
PURPOSE The aim of this study is to investigate and compare the performance of different PET radiopharmaceuticals for the detection of recurrent medullary thyroid carcinoma (MTC) by performing a network meta-analysis (NMA) using direct comparison studies with 2 or more PET radiopharmaceuticals. METHODS PubMed and EMBASE were searched for the studies evaluating the performance of PET or PET/CT for the detection of recurrent MTC. The NMA was performed for different PET radiopharmaceuticals in both patient- and lesion-based analyses and with a threshold of serum calcitonin or carcinoembryonic antigen (CEA) levels and calcitonin doubling time. The consistency was evaluated by examining the agreement between direct and indirect treatment effects, and publication bias was assessed by funnel plot asymmetry tests. The surface under the cumulative ranking curve values were obtained to calculate the probability of each PET modality being the most effective diagnostic method. RESULTS A total of 306 patients from 14 direct comparison studies using 5 different PET radiopharmaceuticals (F-FDG, F-DOPA, Ga-somatostatin analogs, 3-O-methyl-6-[F]fluoro-DOPA, and C-methionine) for the detection of recurrent MTC was included. The detection rate of F-DOPA PET was significantly higher than that of FDG PET in both patient- and lesion-based analyses (patient-based analysis: odds ratio, 2.44; 95% confidence interval, 1.4-4.31; lesion-based analysis: odds ratio, 5.74; 95% confidence interval, 1.65-23.4). Among all PET radiopharmaceuticals, F-DOPA showed the highest surface under the cumulative ranking curve value in both patient- and lesion-based analyses regardless of serum calcitonin or CEA levels and calcitonin doubling time. CONCLUSIONS The results from this NMA indicate that F-DOPA PET clearly showed a best performance for the detection of recurrent MTC in both patient- and lesion-based analyses regardless of serum calcitonin or CEA levels and calcitonin doubling time.
