人工智能术中实时识别可解剖层在腹腔镜腹股沟疝修补术中的应用
该研究对比了54例使用实时AI导航(RAIN)的TAPP手术与454例常规TAPP手术,经倾向性评分匹配后每组53对,发现RAIN组手术时间显著缩短(49.2 vs. 38.5 min),术后并发症无显著差异,提示AI实时识别可解剖层安全且有助于缩短手术时间。
要点
- AI实时导航辅助TAPP可显著缩短手术时间。
- RAIN组术后并发症与常规组无显著差异。
- 该技术临床应用尚处早期,需进一步研究验证。
核心结论
- TAPP术中实时AI手术支持与缩短手术时间相关,且不增加术后并发症。
主要发现
- RAIN组手术时间显著短于常规组(48.6 vs. 38.6 min)。
- 术后并发症两组间无显著差异。
- 倾向性评分匹配后每组53对,RAIN组平均手术时间仍显著更短(49.2 vs. 38.5 min)。
临床怎么用
- 在TAPP术中可考虑使用AI实时识别可解剖层以辅助解剖,可能缩短手术时间。
局限性
- EUREKAα的临床使用仍处于早期阶段,其完整价值需未来研究进一步明确。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Accurate identification of anatomical landmarks such as loose connective tissue, nerves, the vas deferens, and microvessels is essential for transabdominal preperitoneal inguinal hernia repair (TAPP). However, this skill largely depends on the surgeon's experience. EUREKAα (Anaut Inc., Tokyo, Japan), an artificial intelligence (AI)-based surgical support system, was approved in Japan in April 2024 to provide real-time intraoperative recognition and visualization of loose connective tissue. METHODS: Of the 508 TAPP procedures performed between February 2020 and December 2025, 54 were carried out using real-time AI navigation (RAIN), whereas 454 were performed using conventional techniques. Patient characteristics, surgical outcomes, and postoperative complications were statistically analyzed to evaluate the effectiveness and safety of RAIN. Propensity score matching (PSM) was conducted. RESULTS: The RAIN group showed significantly shorter operative times (48.6 vs. 38.6 min) compared with the conventional group, with no significant differences in postoperative complications. After PSM, 53 matched pairs of patients for each group. Mean operative time (49.2 vs. 38.5 min) was significantly shorter in the RAIN group compared to the conventional group. CONCLUSIONS: Real-time intraoperative AI surgical support in TAPP was associated with shorter operative time without increasing postoperative complications. As the clinical use of EUREKAα is still in its early stages, its full value will be more clearly defined through future studies.
引用
- Mita K, Mineta S, Takahashi K, et al. Intraoperative real-time recognition of dissectible layers by artificial intelligence is useful in laparoscopic inguinal hernia repair. Surg Endosc. 2026. PMID: 42622652. doi:10.1007/s00464-026-13241-2
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。