《柳叶刀-数字醫療》(The Lancet Digital Health)近日颁發一項基于人工智能多使命练習模子展望胃癌患者術後腹膜复發和無病保存期的最新钻研。钻研立异地開辟了一種多使命深度進修模子,可經由過程術前CT图象正确地同時展望腹膜复發和無病保存期,并在多中間表里部数据集中获得超卓的驗證。在人工智能(AI)模子的辅助下,临床大夫展望腹膜复發的正确度获得了显著的提高。别的,此AI模子還可以或许從II、III期胃癌患者中區别出最有可能從辅助化療中获益的人群。该模子的開辟有望于引导胃癌的個别化醫治,并選擇最有可能重新療法(如腹腔热灌注化療)中受益的患者。本钻研由南邊醫科大學南邊病院平凡外科、廣東省胃肠肿瘤精准微创诊療重點實行室、美國斯坦福大學醫學院放射肿瘤科牵頭,小荷醫療康健、南邊醫科大學南邊病院醫學影象中間、中山大學肿瘤防治中間、西北工業大學计较機科學與工程學院、加州大學戴维斯分校、美國斯坦福大學醫學院外科等配合完成。江玉明博士、张志诚博士、袁清玉醫師及王玮博士為配合第一作者,李瑞江传授、李國新传授及周志伟传授為配合通信作者。柳叶刀出格约请作者團队,為读者带来文章解读。辨認二维码或點击文末“浏览原文”,领会原文更多内容。
Figure 1. Study design for the development and validation of a deep learning model to predict peritoneal recurrence and disease-free survival
钻研成果
(1)多使命深度進修模子的開辟與驗證
本钻研操纵胃癌患者術前CT图象练習了一個深度卷积神經收集展望腹膜复發和無病保存期。經由過程迭代方法将监视比拟進修计谋與動态神經收集相連系,多標准交融CT图象摸索了潜伏的特性,從而開辟了展望腹膜复發和無病保存期的人工智能(AI)模子,并在驗證行列步队中获得了超卓的驗證。在练習行列步队中,展望腹膜复發方面的受試者操作特性曲線下面积(AUC)為0.857(95% CI 0.826-0.889),而在内部和外部驗證行列步队中,AUC也到達了0.856(95% CI 0.829-0.882)和0.843(95% CI 0.819-0.866)。同時,操纵该模子,钻研發明在练習行列步队中,腹膜复發評分高的患者的5年积累腹膜复發率显著高于低評分患者,别離為49.60%和0.75%(p<0.0001),而在内部和外部驗證行列步队中也發明了雷同的成果(图2)。
Figure 3: Model accuracy for predicting peritoneal recurrence in the training, internal validation, and external validation cohorts
(2)AI模子可以或许正确地域分保存获益的胃癌患者
本钻研評估了AI模子在展望胃癌患者無病保存期方面的機能,發明模子在练習行列步队、内部驗證行列步队和外部驗證行列步队中展望無病保存期的C指数别離為0.654(95% CI 0.616-0.691)、0.668(95%CI 0.643-0.693)和0.610(95% CI 0.583-0.636)。進一步的阐發成果显示,在练習行列步队中,AI模子高評分患者的無病保存期和总保存期均显著擅长低評分患者,而且在内部和外部驗證行列步队中获得了有用的驗證(p<誘蟑捕捉神器,0.001)(图3)。钻研者同時發明,經由過程整合深度進修模子和临床病理學危害身分构建的列線图,可以或许為當前临床實践中利用的TNM分期體系在個别化的保存展望方面供给显著的附加利用價值。
Figure灰指甲外用藥, 4: Kaplan-Meier analyses of disease-free survival and overall survival
(3)AI模子可以或许在II和III期胃癌患者中區分解療获益人群
本钻研除評估模子在展望腹膜复發和保存期方面的機能外,還進一步評估了模子與II期和III期胃癌患者術後化療療效的相干性。經由過程整合展望腹膜复發和保存的評分将患者從新分為4類後,發明對付第一類患者(腹膜复發率高且保存率低),辅助化療可以或许显著提高II期(HR 0.543(95% CI 0.362-0.815),p=0.003)和III期患者(HR 0.531(95% CI 0.432-0.652),p<0.0001)的無病保存率。相反,對付第四類患者(腹膜复發率低且保存率高),辅助化療其實不影响II期(HR 1.559((95% CI 0.915-2.655),p=0.098)或III期患者(HR 1.001(95% CI 0.697-1.439),p=0.990)的無病保存期(图4)。
Figure 4: Relationship between the deep learning model and benefit from adjuvant chemotherapy in matched patients with stage養生茶推薦, II and III gastric cancer
1.Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: A Cancer Journal for Clinicians 68, 394-424 (2018).
2.Noh SH, et al. Adjuvant capecitabine plus oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): 5-year follow-up of an open-label, randomised phase 3 trial. The Lancet Oncology 15, 1389-1396 (2014).
3.Jiang Y, et al. Association of Adjuvant Chemotherapy With Survival in Patients With Stage II or III Gastric Cancer. JAMA Surgery 152, e171087-e171087 (2017).
4.Ikoma N, et al. Patterns of Initial Recurrence in Gastric Adenocarcinoma in the Era of Preoperative Therapy. Annals of surgical oncology 24, 2679-2687 (2017).
5.Thomassen I, et al. Chemotherapy as palliative treatment for peritoneal carcinomatosis of gastric origin. Acta Oncol 53, 429-432 (2014).
6.Mi D-H, et al. Surgery combined wi治療甲溝炎,th intraoperative hyperthermic intraperitoneal chemotherapy (IHIC) for gastric cancer: A systematic review and meta-analysis of randomised controlled trials. International Journal of Hyper祛斑筆,thermia 29, 156-167 (2013).
7.Yang X-J, et al. Cytoreductive Surgery and Hyperthermic Intraperitoneal Chemotherapy Improves Survival of Patients with Peritoneal Carcinomatosis from Gastric Cancer: Final Results of a Phase III Randomized Clinical Trial. Annals of surgical oncology 18, 1575-1581 (2011).
8.Mukherjee P, et al. A shallow convolutional neural network predicts prognosis of lung cancer patients in multi-institutional computed tomography image datasets. Nature Machine Intelligence 2, 274-282 (2020).
9.Lou B, et al. An image-based deep learning framework for individualising radiotherapy dose: a retrospective analysis of outcome prediction. The Lancet Digital Health 1, e136-e147 (2019).
10.Zhao W, et al. HOTAIR is a predictive and prognostic biomarker for patients with advanced gastric adenocarcinoma receiving fluorouracil and platinum combination chemotherapy. Am J Transl Res 7, 1295-1302 (2015).