超声AI影像诊断系统辅助诊断甲状腺结节性质的价值  

Diagnostic value of an ultrasound AI imaging diagnostic system in aiding resident physicians with the malignant risk stratification of thyroid nodules

作　　者：蔡丽萍 张姣 CAI Liping;ZHANG Jiao(Department of Ultrasound,General Hospital of Eastern Theater Command,Nanjing,Jiangsu 210002,China)

机构地区：[1]中国人民解放军东部战区总医院超声诊断科,江苏南京210002

出　　处：《影像研究与医学应用》2025年第5期35-38,42,共5页Journal of Imaging Research and Medical Applications

摘　　要：目的:探讨超声人工智能(AI)影像诊断系统辅助不同年资医师进行甲状腺恶性结节超声特征的识别及分类诊断效果。方法:选取2023年1月—12月中国人民解放军东部战区总医院收治的329例甲状腺结节患者(共381个甲状腺结节)为研究对象,均进行超声检查,所有超声图像均由不同年资医师[高年资医师、住院医师规范化培训(简称规培)医师]、超声AI影像诊断系统辅助诊断方法单独或联合对甲状腺结节超声图像进行分析诊断。以病理结果为金标准,分别计算不同中国超声甲状腺影像报告和数据系统(C-TIRADS)评分截断值下,不同年资医师单独或联合超声AI影像诊断系统对甲状腺结节恶性风险分级的诊断效能。结果:以C-TIRADS评分4A、4B为截断值,AI辅助高年资医师组的诊断效能均优于高年资医师组,差异有统计学意义(P<0.05);以C-TIRADS评分4A为截断值,AI辅助规培医师组的诊断灵敏度、特异度、准确率、阴性预测值高于规培医师组,差异有统计学意义(P<0.05);以C-TIRADS评分4B为截断值,AI辅助规培医师组的诊断效能优于与规培医师组,差异有统计学意义(P<0.05)。结论:在甲状腺结节诊断中,超声AI影像诊断系统辅助不同年资医师诊断的效能均优于传统诊断。Objective To explore the effectiveness of ultrasound artificial intelligence(AI)imaging diagnosis system in assisting physicians of different seniority in identifying and classifying ultrasound features of thyroid malignant nodules.Methods A total of 329 patients with thyroid nodules(381 thyroid nodules in total)admitted to the General Hospital of Eastern Theater Command from January to December 2023 were selected as the study subjects for ultrasonography.All ultrasound images of thyroid nodules were analyzed and diagnosed by sonographers with different experience(experienced sonographers,junior sonographers)and ultrasound AI imaging diagnosis system assisted diagnostic methods alone or jointly.Used pathological results as the gold standard,the diagnostic performance of different Chinese thyroid imaging reporting and data system(C-TIRADS)score cut-off values was calculated for the diagnosis of malignant risk classification of thyroid nodules by sonographers alone or in combination with the ultrasound AI imaging diagnostic system.Results At C-TIRADS score cut-off values of 4A and 4B,the diagnostic performance of the AI-assisted experienced sonographers group was superior to that of the experienced sonographers group alone,with statistically significant differences(P<0.05).At a C-TIRADS score cut-off value of 4A,the sensitivity,specificity,accuracy,and negative predictive value of the AI-assisted junior sonographers group were higher than those of the junior sonographers group alone,with statistically significant differences(P<0.05).At a C-TIRADS score cut-off value of 4B,the diagnostic performance of the AI-assisted junior sonographers group was superior to that of the junior sonographers group alone,with statistically significant differences(P<0.05).Conclusion In the diagnosis of thyroid nodules,the efficacy of ultrasound AI imaging diagnosis system assisted by different seniority doctors is better than that of traditional diagnosis.

关 键 词：人工智能 甲状腺恶性结节 超声诊断 住院医师规范化培训 高年资医师 中国超声甲状腺影像报告和数据系统 

分 类 号：R445.1[医药卫生—影像医学与核医学]