首页|人工智能技术在妇科超声教学中的应用

人工智能技术在妇科超声教学中的应用

扫码查看
目的 探讨人工智能技术应用于附件肿块超声诊断教学工作中的前景,提高教学质量及年轻医师诊断水平.方法 收集2020年1月1日-2022年3月1日解放军某医学中心接受经阴道或经直肠妇科超声检查的附件肿块102例患者超声图像,以病理结果为金标准,分别比较人工智能技术与初级超声医师、高级超声医师评估附件肿块的诊断效能.结果 AI技术准确诊断附件肿块88例,高级医师准确诊断83例,初级医师准确诊断53例.AI技术附件肿块良恶性的准确率0.863、敏感度0.824、特异度0.902均高于初级超声医师(0.578、0.451、0.706),差异具有统计学意义,P<0.05.结论 AI技术应用于附件肿块超声诊断教学工作,有利于提高年轻医师诊断水平,对于建立全国规范化教学体系亦有重要意义.
Application of Artificial Intelligence Technology in Gynecological Ultrasound Teaching
Objectives To explore the prospects of artificial intelligence(AI)technology applied in the teaching of ultrasound diagnosis of adnexal masses,thus to improve the teaching quality and the diagnostic ability of younger doctors.Methods Ultrasound images of 102 patients with adnexal masses underwent transvaginal ultrasound examination at the General Hospital of the People's Liberation Army from Juanuary 1st,2020 to March 1st,2022 were collected.The pathological results were used as the gold standard,and the diagnostic efficacy of artificial intelligence technology was compared with that of primary ultrasound physicians and senior ultrasound physicians in evaluating adnexal masses.Results AI technology accurately diagnosed 88 cases of adnexal masses,83 cases were accurately diagnosed by senior physicians,and 53 cases were accurately diagnosed by junior physicians.AI The accuracy,sensitivity,and specificity of the AI techonologyfor benign and malignant tumors were higher than those of junior ultrasound physicians,and there was no statistically significant difference compared to senior ultrasound physicians.Conclusions The application of AI technology in the teaching of ultrasound diagnosis of accessory masses was beneficial for improving the diagnostic level of younger physicians,and also had great significance for establishing a national standardized teaching system.

Artificial intelligent techonologyUltrasoundTeaching

栗嘉楠、汪龙霞、李秋洋

展开 >

解放军总医院第一医学中心超声诊断科,北京市,100853

人工智能技术 超声 教学

军队后勤科研项目计划生育专项课题

23JSZ10

2024

中国病案
中国医院协会

中国病案

CSTPCD
影响因子:1.197
ISSN:1672-2566
年,卷(期):2024.25(6)