Ex Vivo, Nodule-Based Shear Wave Elastography as a Possible Alternative to Intraoperative Frozen Section for Predicting Malignancy in Thyroid Nodules

Background and Objectives: Intraoperative frozen section (FS) examination is used to differentiate malignant from benign thyroid nodules but is limited by pathologist dependence, time, and cost. This study evaluated whether ex vivo shear wave elastography (SWE), applied separately to mapped thyroid nodules, could serve as an alternative to FS and investigated the effects of nodule number and tumour size on SWE performance. Materials and Methods: A total of 343 thyroid nodules from 100 thyroidectomy patients were mapped and numbered ex vivo. Five two-dimensional SWE measurements were obtained from each nodule using a Philips EPIQ Elite ultrasound system with ElastQ Imaging. Mean, minimum, and maximum stiffness values (kPa) were recorded, and each nodule underwent individual pathological examination. Diagnostic performance was assessed using ROC analysis and the Youden index. Performance was compared according to nodule number (1, 2–3, ≥4) and tumour size (≤10, 11–20, >20 mm). Generalised Estimating Equations (GEE) logistic regression was used to assess independent effects while accounting for within-patient clustering. Results: Of 343 nodules, 229 (66.76%) were benign and 114 (33.23%) malignant. Mean SWE was significantly higher in malignant than benign nodules (49.87 ± 21.30 vs. 18.59 ± 5.53 kPa; p < 0.0001). ROC analysis yielded an AUC of 0.9396 (95% CI: 0.9110–0.9650). At the 25.40 kPa cutoff, sensitivity, specificity, and accuracy were 84.21%, 90.39%, and 88.33%, respectively. Sensitivity decreased with increasing nodule number (92.30%, 86.11%, 81.53%), whereas tumour size correlated strongly with SWE (ρ = 0.7807; p < 0.0001). Sensitivity increased from 70.17% for malignant nodules ≤ 10 mm to 100% for those >20 mm. In GEE analysis, nodule number was not independently significant after adjustment for tumour size (OR = 0.93; 95% CI: 0.58–1.47; p = 0.75), whereas tumour size remained significant (p < 0.0001). Conclusions: Ex vivo nodule-based SWE accurately predicts thyroid malignancy. The apparent effect of nodule multiplicity on sensitivity is largely explained by tumour size. SWE may therefore represent a useful adjunct or potential alternative to intraoperative FS, particularly when nodule size is considered.