摘要：2025年3月12日~3月14日，第97届日本胃癌协会（JGCA）年会在日本名古屋隆重召开，来自日本、中国、韩国、新加坡等多个国家的胃癌领域专家、学者汇聚一堂，共同探讨胃癌的最新研究成果、前沿技术和治疗策略。在大会现场，我们有幸邀请到国际知名胃癌外科教授、韩国

2025年3月12日~3月14日，第97届日本胃癌协会（JGCA）年会在日本名古屋隆重召开，来自日本、中国、韩国、新加坡等多个国家的胃癌领域专家、学者汇聚一堂，共同探讨胃癌的最新研究成果、前沿技术和治疗策略。在大会现场，我们有幸邀请到国际知名胃癌外科教授、韩国外科学会秘书长、韩国胃癌协会主席、首尔大学医学院胃癌中心主任Han-Kwang Yang教授解读胃癌手术领域数字化创新的发展状况和前景。

肿瘤瞭望消化时讯：您如何评估目前胃癌手术领域数字化创新的发展状况？哪些数字技术在胃肠癌领域最具潜力？

Han-Kwang Yang教授：随着数字时代的逐渐到来，我们可以利用人工智能来理解并帮助治疗疾病。首先，人工智能可以用于CT扫描和病理学的解读，也可以应用于内窥镜诊断。胃镜对于非常早期的病变非常模糊，但AI可以帮助检测和诊断。在治疗方面，我们可以利用AI进行手术区域的评估。目前腹腔镜手术的视频录制已经很常见，机器人手术也是如此。在这些手术过程中，正如大会演讲中提到的，如果外科医生缺乏经验，可能会损伤血管或器官，人工智能可以通过识别手术中的血管来帮助确定手术平面。此外，在手术完成后，外科医生需要准备手术记录，而AI可以自动完成所有手术描述。我的团队正在研究AI是否可以评估手术（例如，D2淋巴结清扫术）的质量。我们现在也在研究手术质量与并发症结果之间的关系，以期回答“外科医生进行出血量较大的手术是否会导致更高的术后并发症率”这一问题。每次手术大约需要4~5小时，经过上百次手术后用肉眼评估非常困难，客观性也是一个问题。AI可以帮助进行手术评估，这就是我们目前在AI研究领域的探索。

Dr Yang: As we have entered the digital era, which means we can use artificial intelligence for our understanding of disease and to help in treatment. Firstly, we already know that artificial intelligence is used for the interpretation of CT scans and pathology. It is also applied for diagnoses by endoscopy. Gastroscopy is very vague for a very early lesion, but AI can help in detection and diagnosis. For treatment, we are using surgical field assessment. Video- recording of laparoscopic procedures is now very common, and surely robotic, too. During these operations, as we heard today in different presentations, if surgeons don’t have much experience intraoperatively, there can be injury to vessels or organs. Artificial intelligence can help with the surgical plane by identifying vessels during the operation. Also, when an operation is finished, the surgeon needs to prepare an operation record of what has been done, but with AI, all the operation description can be automatically done. What my team is working on is whether the operation quality can be evaluated by AI. If you do a proper D2 dissection, AI can determine if it is a good D2 dissection or not. Another application we are looking at now is comparing operation quality versus outcomes with regard to complications. If a surgeon does very bloody surgery, does this lead to a higher complication rate? We want to see this. Each operation is 4-5 hours, and over a hundred operations, it is very difficult to evaluate with the human eye. Objectivity is also a concern too. AI can help with this kind of evaluation. That is our current activity in the field of AI studies.

肿瘤瞭望消化时讯：您认为胃癌领域数字化创新的主要挑战是什么？

Han-Kwang Yang教授：正如我提到的，胃癌领域数字化确实具有前景，但问题是准确性。如果要投入使用，必须保证准确性。这正是我们所担忧的。为了更有效地进行相关评估，我们需要获取大量的手术视频数据。数据越多，越容易制定结果评估。然而在这一过程中，确实有一些困难需要克服，因为在许多国家，由于隐私保护等原因，将手术视频带出医院非常困难。这无疑是一个巨大的挑战。

Dr Yang: That is a difficult question. As I mentioned, the prospects are there, but the concern is accuracy. If it is made available, this must be truly accurate. This is our concern. We need a lot of data from operation videos. There are some hurdles to meeting these challenges, because bringing an operation video out of a hospital is very difficult in many countries due to protection of privacy. The more data there is, the easier it is to formulate outcome evaluation, but, I feel that is a big challenge at this moment because of this limitation.

肿瘤瞭望消化时讯：您认为各国数字化创新的发展有哪些差异？

Han-Kwang Yang教授：在全球范围内，临床医生和科学家们都有各自的研究进展，在发展方向和所使用的开发工具上自然会存在差异。这也正是我们希望开展合作的原因。例如，我们团队正在与美国的一家AI公司合作进行分析。这种合作加速了基于AI的计算机视觉对手术过程评估的发展。如果我们能找到某种手术行为与并发症之间的相关性，这意味着我们可以减少那些并发症的发生。在日本，他们已经开发了术中辅助系统，利用AI确定合适的手术计划。而在欧洲国家，他们强调解剖结构，如可视化血管结构。此外，术前虚拟规划已经研究了几十年。通过共同努力，我们可以实现更好的应用。由于我们在地理上非常接近（仅几个小时的距离），因此我希望年轻的中国医生和韩国医生能够加强合作，加快AI新平台的开发和利用。

Dr Yang: All over the world, clinicians and scientists have their own developments. It is quite natural to see differences in development direction and development tools. This is why we want to work together. For example, my team is working with an AI company in the Unites States for analysis. This kind of collaboration speeds up development of AI-based computer vision evaluation of surgical procedures. If we can find a correlation between a certain surgical behavior versus complications, that means we can reduce those complications. In Japan already, they have developed intraoperative assistance whereby AI determines a proper surgical plan. In Japan, they have emphasized this kind of approach. In European countries, they have emphasized anatomy, such as visualizing vascular structures. Also, preoperative virtual planning has been worked on for decades. By combining forces, we can make for much better application. That is what I believe. I hope young Chinese doctors and young Korean doctors can cooperate together. We are geographically very close (only hours away), so hopefully we can work together to speed up development of new platforms for using AI.

来源：肿瘤瞭望消化时讯