Prediction of Malignant Thyroid Nodules Using 18F-FDG PET/CT–Based Radiomics Features in Thyroid Incidentalomas

Abstract

Objective The purpose of the current study was to evaluate the diagnostic performances of 18F-FDG PET/CT–based radiomics features for prediction of malignant thyroid nodules (TNs) in thyroid incidentaloma (TI). Methods PubMed, Cochrane database, and EMBASE database, from the earliest available date of indexing through December 31, 2022, were searched for studies evaluating diagnostic performance of 18F-FDG PET/CT–based radiomics features for prediction of malignant TNs in TI. We determined the sensitivities and specificities across studies, calculated positive and negative likelihood ratios (LRs; positive and negative LRs), and estimated pooled area under the curve. Results Across 5 studies (518 patients), the pooled sensitivity of 18F-FDG PET/CT was 0.77 (95% confidence interval [CI], 0.67–0.84), and a pooled specificity was 0.67. Likelihood ratio syntheses gave an overall positive LR of 2.3 (95% CI, 1.5–3.6) and negative LR of 0.35 (95% CI, 0.26–0.47). The pooled diagnostic odds ratio was 7 (95% CI, 4–12). The pooled area under the curve of fixed effects was 0.763 (95% CI, 0.736–0.791), and that of random effects was 0.763 (95% CI, 0.721–0.805). Conclusion 18F-FDG PET/CT–based radiomics features showed a good diagnostic performance for prediction of malignant TNs in TI.