Size-stratified optimal perinodular stiffness by shear-wave elastography enhances malignancy discrimination in...

To identify the optimal perinodular stiffness metric across thyroid nodule size strata and derive size-specific diagnostic thresholds using a decision-tree model.

In this retrospective multicenter study, 718 nodules were stratified by maximum diameter (MD): MD ≤ 10 mm, 10 < MD ≤ 15 mm, 15 < MD ≤ 20 mm. Shear-wave elastography (SWE) measured intranodular maximum stiffness (E_max) and maximum stiffness within the 0.5 mm/1.0 mm/1.5 mm perinodular shell (E_{shell-0.5mm/1.0mm/1.5mm}). Within each stratum, the parameter with the largest receiver operating characteristic (ROC) area under the curve (AUC) was selected, and a decision tree was used to derive diagnostic thresholds. Biomechanical and histologic relevance was evaluated using atomic force microscopy (AFM)-measured stromal modulus and Masson's trichrome collagen fraction.

Malignant nodules showed higher E_max and E_shell (all P < 0.01). E_shell-1.0mm achieved the highest AUC for MD ≤ 10 mm (0.90) and 10 < MD ≤ 15 mm (0.92); for 15 < MD ≤ 20 mm, E_max performed best (0.86). Thresholds were: MD ≤ 10 mm-E_max > 75.71 kPa and E_shell-1.0mm/E_max > 1.23 (terminal-node malignancy rate 94.8%); 10 < MD ≤ 15 mm-E_max > 88.96 kPa and E_shell-1.0mm > 96.96 kPa (91.8%); 15 < MD ≤ 20 mm-E_max > 66.45 kPa (79.8%). Perinodular metrics correlated with AFM modulus and collagen; E_shell-1.0mm was strongest (r = 0.77-0.82; 0.75-0.77).

Size-stratified perinodular stiffness improves benign-malignant discrimination. E_shell-1.0mm is the preferred diagnostic region, particularly for MD ≤ 15 mm, supported by biomechanical and histologic validation.