| 牛蕾,何晓琴,梁博诚,洪礼伟,练敏,洪思婕,廖慧芳,周萍,李雯婷,李胜利,钟晓红.对比人工智能辅助培训(AIST)与传统人工培训(TMT)用于产前超声标准切面规范化培训[J].中国医学影像技术,2025,41(12):1950~1953 |
| 对比人工智能辅助培训(AIST)与传统人工培训(TMT)用于产前超声标准切面规范化培训 |
| Comparison on artificial intelligence supplementary training (AIST) and traditional manual training (TMT) for standardized training of prenatal ultrasound standard sections |
| 投稿时间：2025-08-24 修订日期：2025-11-02 |
| DOI：10.13929/j.issn.1003-3289.2025.12.003 |
| 中文关键词: 超声检查,产前 标准切面 人工智能 培训 |
| 英文关键词:ultrasonography, prenatal standard sections artificial intelligence training |
| 基金项目:国家重点研发计划(2022YFF0606301)。 |
| 作者 | 单位 | E-mail | | 牛蕾 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)妇产科学系超声医学科, 福建 厦门 361000 | | | 何晓琴 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)妇产科学系超声医学科, 福建 厦门 361000 | | | 梁博诚 | 南方医科大学妇女儿童医学中心深圳市妇幼保健院超声科, 广东 深圳 518028 | | | 洪礼伟 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)科教部, 福建 厦门 361000 | | | 练敏 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)妇产科学系超声医学科, 福建 厦门 361000 | | | 洪思婕 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)妇产科学系超声医学科, 福建 厦门 361000 | | | 廖慧芳 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)妇产科学系超声医学科, 福建 厦门 361000 | | | 周萍 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)妇产科学系超声医学科, 福建 厦门 361000 | | | 李雯婷 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)妇产科学系超声医学科, 福建 厦门 361000 | | | 李胜利 | 南方医科大学妇女儿童医学中心深圳市妇幼保健院超声科, 广东 深圳 518028 | | | 钟晓红 | 厦门大学附属妇女儿童医院(厦门市妇幼保健院)妇产科学系超声医学科, 福建 厦门 361000 | 1369946030@qq.com |
|
| 摘要点击次数: 99 |
| 全文下载次数: 14 |
| 中文摘要: |
| 目的 对比人工智能辅助培训(AIST)与传统人工培训(TMT)用于产前超声标准切面规范化培训的效果。方法 对6名无产前超声经验医师(学员)进行产前超声理论、仪器及操作技术教学,待其通过基线水平评估后随机划分AIST组 与TMT组(由工作10年以上导师进行一对一指导),完成目标为在15 min内采集22种共28个标准切面的15次培训后,脱离AI或导师验证培训效果,计算标准率、完整率及重复率等并进行组间比较。结果 培训前6名学员基线水平差异均无统计学意义(P均＞0.05)。培训期间及培训后AIST组标准率(67.30% vs. 45.95%)和完整率(74.60% vs. 55.08%)均显著高于TMT组(P均＜0.01)。 AIST组在培训期间的重复率为0,且在前4次培训中进步更快。完成培训后2组标准率和完整率均显著高于培训前(P均＜0.05),且AIST组提高绝对值均高于TMT组。结论 AIST用于产前超声标准切面规范化培训的效果优于TMT。 |
| 英文摘要: |
| Objective To compare the effectiveness of artificial intelligence supplementary training (AIST) and traditional manual training (TMT) for standardized training of prenatal ultrasound standard sections. Methods Six physicians without prenatal ultrasound experience (trainees) were selected and randomly divided into AIST group (using artificial intelligence ultrasound workstation) and TMT group (guided by a mentor with over 10 years of experience in a one-on-one manner) after undergoing prenatal ultrasound theory, instrument and operational technique teaching and passing baseline level assessment. Each trainee completed 15 training sessions, with the goal of acquiring 28 standard sections across 22 types within 15 min per session. Then the trainees were evaluated independently of AI or mentor to observe the training effect, and indicators such as standardization rate, completeness rate and repetition rate were calculated and compared between groups. Results Before training, no significant difference of baseline levels was found among trainees (all P＞0.05). During and after training, standardization rate (67.30% vs. 45.95%) and completeness rate (74.60% vs. 55.08%) in AIST group were both significantly higher than those in TMT group (both P＜0.01). During training, repetition rate in AIST group was 0, and trainees in AIST group progressed faster in the first 4 training sessions. After training, standardization and completeness rates in both groups were significantly higher than those before training (both P＜0.05), and the absolute improvement value in AIST group was higher than that in TMT group. Conclusion AIST was more effective than TMT for standardized training of prenatal ultrasound standard sections. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
