分析测量甲状腺结节剪切波弹性成像（SWE）杨氏模量最大值（Emax）的影响因素，并探讨按最显著影响因素分组是否可提高Emax值对结节性质的鉴别诊断价值。

回顾性选取2017年1月至2019年11月在山西医科大学附属白求恩医院行甲状腺结节手术并经术后病理证实的368例患者（共413个结节），术前均行二维超声及SWE检查。常规超声记录结节的大小、数量、纵横比、位置、内部成分、回声、边缘、钙化；SWE记录结节的Emax值及相对应的深度。应用Spearman秩相关筛选甲状腺恶性结节Emax值的相关因素，对相关因素行多元线性回归分析，依据最显著影响因素分组，采用ROC曲线进一步分析Emax值对结节性质的鉴别诊断价值。

413个甲状腺结节中，良性106个，恶性307个。相关性分析结果表明：甲状腺恶性结节大小、边缘、钙化与Emax值呈正相关（r=0.477、0.244、0.256，P均＜0.05），结节深度与Emax值呈负相关（r=-0.132，P＜0.05）；多元线性回归分析结果表明：结节大小对Emax值影响最显著（标准化偏回归系数为0.537）。按结节最大径将甲状腺结节分为A组（最大径≤1 cm）、B组（1 cm＜最大径≤2 cm）、C组（2 cm＜最大径≤3 cm）。未按结节大小分组Emax值诊断甲状腺恶性结节的截断值为36.0 kPa，ROC曲线下面积（AUC）为0.830，诊断敏感度、特异度分别为73.0％、81.1％；依据结节大小分组的A、B、C组Emax值诊断甲状腺恶性结节的截断值分别为33.7 kPa、37.8 kPa、57.1 kPa，AUC分别为0.832、0.889、0.952，A组诊断敏感度、特异度分别为68.0%、87.9%，B组诊断敏感度、特异度分别为89.0%、82.1%，C组诊断敏感度、特异度分别为94.7%、90.9%。与未按结节大小分组比较，A、B、C组的AUC均有不同程度增大，A组Emax值诊断的敏感度降低，特异度增加，B、C组Emax值诊断的敏感度及特异度均增加。

甲状腺结节Emax值受结节大小的影响最为显著，按结节大小分组，可提高Emax值对结节最大径＞1 cm的甲状腺结节的诊断价值，应用Emax值诊断甲状腺结节性质时，不同大小结节建议采用不同的截断值。

To analyze the influencing factors of maximum Young's modulus (Emax) of shear wave elastography (SWE) of thyroid nodules, and to explore whether grouping by the most significant influencing factors can improve the diagnostic value of Emax.

A total of 368 patients (with 413 nodules in total) who underwent surgery for thyroid nodules at Bethune Hospital Affiliated to Shanxi Medical University from January 2017 to November 2019 were retrospectively selected. All patients underwent preoperative two-dimensional ultrasound and SWE examination. The size, number, shape, position, internal composition, echo, margin, and calcification of nodules were recorded by conventional ultrasound. Emax values and the depths of nodules were recorded by SWE. Spearman rank correlation was used to screen the influencing factors of Emax value of thyroid malignant nodules. Multiple linear regression analysis was performed on the related factors, and the diagnostic value of Emax value was further analyzed according to the grouping of the most significant influencing factors.

Among 413 thyroid nodules, 106 were benign and 307 were malignant. The results of correlation analysis showed that the size, margin, and calcification of thyroid malignant nodules were positively correlated with Emax value (r=0.477, 0.244, and 0.256, respectively; P＜0.05 for all), while the depth of thyroid nodules was negatively correlated with Emax value (r=-0.132, P＜0.05). Multiple linear regression analysis showed that nodule size had the most significant effect on Emax value (standardized partial regression coefficient, 0.537). The nodules were divided into group A (maximum diameter ≤1 cm), group B (1 cm < maximum diameter ≤2 cm), and group C (2 cm < maximum diameter ≤3 cm) according to their maximum diameter. The cutoff value was 36.0 kPa, the area under the ROC curve (AUC) was 0.830, and the sensitivity and specificity were 73.0% and 81.1%, respectively, for the diagnosis of thyroid malignant nodules not grouped by Emax value. The cutoff value of Emax for diagnosis of thyroid malignant nodules in groups A, B, and C according to the size of nodules was 33.7 kPa, 37.8 kPa, and 57.1 kPa, and AUC was 0.832, 0.889, and 0.952, respectively. The diagnostic sensitivity and specificity in group A were 68.0% and 87.9% in group A, 89.0% and 82.1% in group B, and 94.7% and 90.9% in group C, respectively. Compared with the nodules not grouped by nodule size, the AUCs of groups A, B, and C increased to varying degrees, the diagnostic sensitivity and specificity of Emax value decreased in group A, and increased in groups B and C.

The Emax value of thyroid nodule is most significantly affected by the size of thyroid nodules. Grouping by the size of thyroid nodule can improve the diagnostic value of Emax value for thyroid nodules with the maximum diameter >1 cm. When using Emax value to diagnose the nature of thyroid nodules, it is recommended to use different cutoff values for different nodule sizes.