关键词: 多模态超声, 甲状腺精细解剖, 桥本甲状腺炎, 甲状腺微小乳头状癌, 诊断效能

Abstract:

Objective This study investigates the clinical utility of detailed anatomical segmentation of the thyroid combined with multimodal ultrasound to provide imaging evidence for early clinical diagnosis and risk stratification. Methods A retrospective analysis was conducted on the clinical data of 220 patients with Hashimoto′s thyroiditis （HT） and thyroid micronodules （maximum diameter ≤ 1.0 cm） admitted to the Affiliated Hospital of Xuzhou Medical University from January 2024 to December 2025. All patients met the indications for surgery following clinical risk assessment， with postoperative pathological results serving as the gold standard； The 9-segment detailed anatomical localization method was used to determine nodule locations. Multimodal ultrasound features of benign and malignant nodules were compared， and the distribution patterns of malignant nodules across different anatomical subsections were analyzed. Univariate and multivariate logistic regression were performed to identify independent risk factors for papillary thyroid microcarcinoma （PTMC）， and a combined diagnostic model was constructed to evaluate its diagnostic performance. Results The proportion of malignant nodules in the subcapsular regions of the thyroid is higher than in the central zone of the gland； in the upper anteromedial subcapsular subregion， the proportion of malignant nodules reaches 72.73 percent， whereas in the central zone it is only 26.19 percent； Statistically significant differences were observed between benign and malignant nodules in terms of marginal morphology， aspect ratio， internal echogenicity， calcification characteristics， capsular continuity， blood flow grading， and elasticity score （P < 0.001）； Multivariate logistic regression analysis revealed that irregular margins， an aspect ratio ≥ 1， disruption of capsular continuity， an elasticity score ≥ 4， and nodule location in the subcapsular region were all independent risk factors for HT combined with PTMC （P < 0.05）， with disruption of capsular continuity showing the strongest association （OR= 9.156）； The area under the curve （AUC） for for the diagnosis using multimodal ultrasound combined with anatomical localization was 0.913， with a Youden′s index of 0.827， a sensitivity of 92.86%， and a specificity of 89.81%； Its diagnostic performance was significantly superior to that of any single ultrasound modality （P < 0.001）. Conclusions Combining detailed anatomical thyroid segmentation with multimodal ultrasound can effectively identify the anatomical distribution and imaging characteristics of PTMC in the context of HT. This approach offers superior diagnostic performance compared to single ultrasound modalities and provides an important reference for the early differentiation and individualized diagnostic and therapeutic decision-making for such lesions.