Abstract
BackgroundThere is a growing need for a reproducible and effective imaging method for the quantitative differentiation of benign from malignant thyroid nodules. This study aimed to investigate the performances of intravoxel incoherent motion (IVIM) parameters and the apparent diffusion coefficient (ADC) in differentiating malignant from benign thyroid nodules derived from the most repeatable region of interest (ROI) delineation.MethodsForty-three patients with 46 pathologically confirmed thyroid nodules underwent diffusion-weighted imaging (DWI) with 8 b values. Two observers measured the intravoxel incoherent motion (IVIM) parameters (D, f and D*) and the apparent diffusion coefficient (ADC), ADC600 and ADC990 values using whole-lesion (W-L) ROI and IVIM parameters using single-section (S-S) ROI delineation. The intraclass correlation coefficients (ICCs) and Bland-Altman plots were used to evaluate the intra- and interobserver variability. The diagnostic performance of these parameters was evaluated by generating receiver operating characteristic (ROC) curves.ResultsThe ICC values of all IVIM with W-L ROI delineation were higher than those with S-S ROI delineation, and excellent intra- and interobserver reproducibility was obtained. According to the Bland-Altman plots, the 95% limits of agreement of the IVIM parameters determined by the W-L ROIs revealed smaller absolute intra- and interobserver variability than those determined by S-S ROIs. The D and ADC600 values obtained from the W-L ROIs were the most powerful parameters in differentiating benign from the malignant nodules [area under the ROC curve = 0.962 and 0.970, P = 0.771].ConclusionsThe W-L ROI of the thyroid was considered an effective method for obtaining IVIM measurements with excellent reproducibility for differentiating benign from malignant nodules.
Highlights
There is a growing need for a reproducible and effective imaging method for the quantitative differentiation of benign from malignant thyroid nodules
Inclusion criteria were as follows: (1) thyroid nodules or masses detected by physical examination or incidentally; (2) thyroid nodules > 1 cm detected by ultrasonography; (3) clinicians and/or patients believed that the lesion could be removed surgically; and (4) written informed consent for magnetic resonance imaging (MRI) examination was provided
The intraclass correlation coefficient (ICC) values of all intravoxel incoherent motion (IVIM) parameters (D, f and D*) using W-L region of interest (ROI) delineation were higher than those using S-S ROI delineation, and excellent intraand interobserver reproducibility was obtained
Summary
There is a growing need for a reproducible and effective imaging method for the quantitative differentiation of benign from malignant thyroid nodules. This study aimed to investigate the performances of intravoxel incoherent motion (IVIM) parameters and the apparent diffusion coefficient (ADC) in differentiating malignant from benign thyroid nodules derived from the most repeatable region of interest (ROI) delineation. On the basis of a recent metaanalysis of studies differentiating benign from malignant thyroid nodules, the sensitivity of quantitative DWI and ADC was 0.90~0.91, and the specificity was 0.93~0.95 [6, 7]. Considering the limitations of applying lower or higher b values in a linear form, investigations into head and neck cancers to differentiate benign from malignant neoplasms and predict treatment responses have been performed using intravoxel incoherent motion (IVIM), which could better reflect true diffusion in the tissue [8, 9]. Based on large-sample and multi-centre research, artificial intelligence and machine learning could gradually be applied to magnetic resonance imaging of thyroid nodules
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
