{"title":"在机器人辅助胃切除术中通过语义分割精确标出胰腺:利用人工智能为外科医生提供视觉辅助","authors":"Tatsuro Nakamura, Nao Kobayashi, Yuta Kumazu, Kyohei Fukata, Motoki Murakami, Shugo Kohno, Yudai Hojo, Eiichiro Nakao, Yasunori Kurahashi, Yoshinori Ishida, Hisashi Shinohara","doi":"10.1007/s10120-024-01495-5","DOIUrl":null,"url":null,"abstract":"<h3 data-test=\"abstract-sub-heading\">Background</h3><p>A postoperative pancreatic fistula (POPF) is a critical complication of radical gastrectomy for gastric cancer, mainly because surgeons occasionally misrecognize the pancreas and fat during lymphadenectomy. Therefore, this study aimed to develop an artificial intelligence (AI) system capable of identifying and highlighting the pancreas during robot-assisted gastrectomy.</p><h3 data-test=\"abstract-sub-heading\">Methods</h3><p>A pancreas recognition algorithm was developed using HRNet, with 926 training images and 232 validation images extracted from 62 scenes of robot-assisted gastrectomy videos. During quantitative evaluation, the precision, recall, intersection over union (IoU), and Dice coefficients were calculated based on the surgeons’ ground truth and the AI-inferred image from 80 test images. During the qualitative evaluation, 10 surgeons answered two questions related to sensitivity and similarity for assessing clinical usefulness.</p><h3 data-test=\"abstract-sub-heading\">Results</h3><p>The precision, recall, IoU, and Dice coefficients were 0.70, 0.59, 0.46, and 0.61, respectively. Regarding sensitivity, the average score for pancreas recognition by AI was 4.18 out of 5 points (1 = lowest recognition [less than 50%]; 5 = highest recognition [more than 90%]). Regarding similarity, only 54% of the AI-inferred images were correctly differentiated from the ground truth.</p><h3 data-test=\"abstract-sub-heading\">Conclusions</h3><p>Our surgical AI system precisely highlighted the pancreas during robot-assisted gastrectomy at a level that was convincing to surgeons. This technology may prevent misrecognition of the pancreas by surgeons, thus leading to fewer POPFs.</p>","PeriodicalId":12684,"journal":{"name":"Gastric Cancer","volume":null,"pages":null},"PeriodicalIF":6.0000,"publicationDate":"2024-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Precise highlighting of the pancreas by semantic segmentation during robot-assisted gastrectomy: visual assistance with artificial intelligence for surgeons\",\"authors\":\"Tatsuro Nakamura, Nao Kobayashi, Yuta Kumazu, Kyohei Fukata, Motoki Murakami, Shugo Kohno, Yudai Hojo, Eiichiro Nakao, Yasunori Kurahashi, Yoshinori Ishida, Hisashi Shinohara\",\"doi\":\"10.1007/s10120-024-01495-5\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<h3 data-test=\\\"abstract-sub-heading\\\">Background</h3><p>A postoperative pancreatic fistula (POPF) is a critical complication of radical gastrectomy for gastric cancer, mainly because surgeons occasionally misrecognize the pancreas and fat during lymphadenectomy. Therefore, this study aimed to develop an artificial intelligence (AI) system capable of identifying and highlighting the pancreas during robot-assisted gastrectomy.</p><h3 data-test=\\\"abstract-sub-heading\\\">Methods</h3><p>A pancreas recognition algorithm was developed using HRNet, with 926 training images and 232 validation images extracted from 62 scenes of robot-assisted gastrectomy videos. During quantitative evaluation, the precision, recall, intersection over union (IoU), and Dice coefficients were calculated based on the surgeons’ ground truth and the AI-inferred image from 80 test images. During the qualitative evaluation, 10 surgeons answered two questions related to sensitivity and similarity for assessing clinical usefulness.</p><h3 data-test=\\\"abstract-sub-heading\\\">Results</h3><p>The precision, recall, IoU, and Dice coefficients were 0.70, 0.59, 0.46, and 0.61, respectively. Regarding sensitivity, the average score for pancreas recognition by AI was 4.18 out of 5 points (1 = lowest recognition [less than 50%]; 5 = highest recognition [more than 90%]). Regarding similarity, only 54% of the AI-inferred images were correctly differentiated from the ground truth.</p><h3 data-test=\\\"abstract-sub-heading\\\">Conclusions</h3><p>Our surgical AI system precisely highlighted the pancreas during robot-assisted gastrectomy at a level that was convincing to surgeons. This technology may prevent misrecognition of the pancreas by surgeons, thus leading to fewer POPFs.</p>\",\"PeriodicalId\":12684,\"journal\":{\"name\":\"Gastric Cancer\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":6.0000,\"publicationDate\":\"2024-04-04\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Gastric Cancer\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1007/s10120-024-01495-5\",\"RegionNum\":1,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"GASTROENTEROLOGY & HEPATOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Gastric Cancer","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1007/s10120-024-01495-5","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"GASTROENTEROLOGY & HEPATOLOGY","Score":null,"Total":0}
Precise highlighting of the pancreas by semantic segmentation during robot-assisted gastrectomy: visual assistance with artificial intelligence for surgeons
Background
A postoperative pancreatic fistula (POPF) is a critical complication of radical gastrectomy for gastric cancer, mainly because surgeons occasionally misrecognize the pancreas and fat during lymphadenectomy. Therefore, this study aimed to develop an artificial intelligence (AI) system capable of identifying and highlighting the pancreas during robot-assisted gastrectomy.
Methods
A pancreas recognition algorithm was developed using HRNet, with 926 training images and 232 validation images extracted from 62 scenes of robot-assisted gastrectomy videos. During quantitative evaluation, the precision, recall, intersection over union (IoU), and Dice coefficients were calculated based on the surgeons’ ground truth and the AI-inferred image from 80 test images. During the qualitative evaluation, 10 surgeons answered two questions related to sensitivity and similarity for assessing clinical usefulness.
Results
The precision, recall, IoU, and Dice coefficients were 0.70, 0.59, 0.46, and 0.61, respectively. Regarding sensitivity, the average score for pancreas recognition by AI was 4.18 out of 5 points (1 = lowest recognition [less than 50%]; 5 = highest recognition [more than 90%]). Regarding similarity, only 54% of the AI-inferred images were correctly differentiated from the ground truth.
Conclusions
Our surgical AI system precisely highlighted the pancreas during robot-assisted gastrectomy at a level that was convincing to surgeons. This technology may prevent misrecognition of the pancreas by surgeons, thus leading to fewer POPFs.
期刊介绍:
Gastric Cancer is an esteemed global forum that focuses on various aspects of gastric cancer research, treatment, and biology worldwide.
The journal promotes a diverse range of content, including original articles, case reports, short communications, and technical notes. It also welcomes Letters to the Editor discussing published articles or sharing viewpoints on gastric cancer topics.
Review articles are predominantly sought after by the Editor, ensuring comprehensive coverage of the field.
With a dedicated and knowledgeable editorial team, the journal is committed to providing exceptional support and ensuring high levels of author satisfaction. In fact, over 90% of published authors have expressed their intent to publish again in our esteemed journal.