Li Xiangxiang, Han Xu, Liu Nan, Wang Shen, Zheng Hongyuan, Ma Ziyu, Zhang Ruiguo, Jia Qiang, Zheng Wei
Department of Nuclear Medicine, Tianjin Medical University General Hospital, Tianjin, China.
Front Endocrinol (Lausanne). 2025 Jul 25;16:1617229. doi: 10.3389/fendo.2025.1617229. eCollection 2025.
Thyroid mass is crucial for I treatment of Graves' disease (GD). However, the accuracy of ultrasound (US) - based thyroid mass measurement remains controversial.
This retrospective study included patients who underwent thyroid US and CT scans. The differences correlation, and agreement in thyroid mass measurements between the two methods were analyzed. Data from GD patients who received their first I treatment were collected and evaluated at a 6-month follow-up. Regression analyses identified clinical factors for treatment efficacy and developed a predictive model.
A statistically significant difference was observed in thyroid mass measurements exceeding 20 g between US and CT. (Z = -11.493, P<0.001). Despite a strong correlation between the two methods (r = 0.9809, P=0.001), the average relative error remained substantial (0.19 ± 11.65%). Poor agreement was observed between CT and US (mean bias: 16.65g; ICC = 0.179, p = 0.087). Disease duration, FT level, 24 - hour radioactive iodine uptake, I dose and thyroid mass were identified as independent risk factors influencing the efficacy of the initial I treatment (p<0.05). Based on these factors, a predictive model was developed and evaluated using ROC curves, DCA and CAL. The model demonstrated an AUC of 0.663 (95% CI = 0.631-0.695).
US may underestimate the true thyroid mass in large-mass cases; therefore, CT calibration is recommended before initiating I treatment. The proposed predictive model provides valuable guidance for optimizing initial I treatment in patients with GD.
甲状腺质量对格雷夫斯病（GD）的碘治疗至关重要。然而，基于超声（US）的甲状腺质量测量的准确性仍存在争议。
这项回顾性研究纳入了接受甲状腺超声和CT扫描的患者。分析了两种方法在甲状腺质量测量方面的差异、相关性和一致性。收集接受首次碘治疗的GD患者的数据，并在6个月的随访中进行评估。回归分析确定了治疗效果的临床因素，并建立了预测模型。
在美国和CT之间，甲状腺质量测量超过20g时观察到统计学上的显著差异。（Z = -11.493，P<0.001）。尽管两种方法之间有很强的相关性（r = 0.9809，P = 0.001），但平均相对误差仍然很大（0.19 ± 11.65%）。CT和US之间观察到一致性较差（平均偏差：16.65g；ICC = 0.179，p = 0.087）。疾病持续时间、FT水平、24小时放射性碘摄取、碘剂量和甲状腺质量被确定为影响初始碘治疗效果的独立危险因素（p<0.05）。基于这些因素，使用ROC曲线、DCA和CAL建立并评估了一个预测模型。该模型的AUC为0.663（95%CI = 0.631 - 0.695）。
在大质量病例中，超声可能低估甲状腺的真实质量；因此，建议在开始碘治疗前进行CT校准。所提出的预测模型为优化GD患者的初始碘治疗提供了有价值的指导。
