胸“星”团队成果展示|基于影像组学的机器学习可术前预测临床IA期纯实性非小细胞肺癌患者的生存结局

学术   科学   2024-05-31 19:58   四川  


胸小星第二十六期成果

如期见面

让我们走进本期推文

一起来看看吧


文章题目:Preoperatively Predicting Survival Outcome for Clinical Stage IA Pure-Solid Non-Small Cell Lung Cancer by Radiomics-based Machine Learning
发表杂志:The Journal of Thoracic and Cardiovascular Surgery
影响因子:6.0
第一作者:严浩吉
通讯作者:Kenji Suzuki
DOI:10.1016/j.jtcvs.2024.05.010.

Preoperatively Predicting Survival Outcome for Clinical Stage IA Pure-Solid Non-Small Cell Lung Cancer by Radiomics-based Machine Learning

Haoji Yan, MD1, Takahiro Niimi, MD1, Takeshi Matsunaga, MD1, Mariko Fukui, MD1, Aritoshi Hattori, MD1, Kazuya Takamochi, MD1, Kenji Suzuki, MD2

1. Department of General Thoracic Surgery, Juntendo University School of Medicine, 1-3 Hongo 3-chome, Bunkyo-ku, Tokyo, Japan, 113-8431.

2. Department of General Thoracic Surgery, Juntendo University School of Medicine, 1-3 Hongo 3-chome, Bunkyo-ku, Tokyo, Japan, 113-8431.


Objective:

Clinical stage IA non-small cell lung cancer (NSCLC) showing a pure-solid appearance on computed tomography (CT) is associated with a worse prognosis. This study aimed to develop and validate machine learning models using preoperative clinical and radiomic features to predict overall survival (OS) in clinical stage IA pure-solid NSCLC.


Methods:

Patients who underwent lung resection for NSCLC between January 2012 and December 2020 were reviewed. The radiomic features were extracted from the intratumoral and peritumoral regions on CT. The machine learning models were developed using random survival forests (RSF) and XGBoost algorithms, while the Cox regression model was set as a benchmark. Model performance was assessed using the integrated time-dependent area under the curve (iAUC) and validated by 5-fold cross validation.


Results:

In total, 642 clinical stage IA pure-solid NSCLC patients were included. Among 3748 radiomic and 34 preoperative clinical features, 42 features were selected. Both machine learning models outperformed the Cox regression model (iAUC, 0.753 [95% CI: 0.629, 0.829]). The XGBoost model showed a better performance (iAUC, 0.832 [95% CI: 0.779, 0.880]) than the RSF model (iAUC, 0.795 [95% CI: 0.734, 0.856]). The XGBoost model showed an excellent survival stratification performance with a significant overall survival (OS) difference among the low-risk (5-year OS: 100.0%), moderate low-risk (5-year OS: 88.5%), moderate high-risk (5-year OS: 75.6%), and high-risk (5-year OS: 41.7%) groups (P< 0.0001).


Conclusions:

Radiomics-based machine learning model can preoperatively and accurately predict OS and improve survival stratification in clinical stage IA pure-solid NSCLC.

向下滑动查看所有内容

基于影像组学的机器学习可术前预测临床IA期纯实性非小细胞肺癌患者的生存结局



目的

临床ⅠA期非小细胞肺癌(non-small cell lung cancer, NSCLC)在计算机断层扫描(computed tomography, CT)上表现为纯实性时,预后更差。本研究旨在利用术前临床特征和影像学特征开发和验证机器学习模型,以预测临床ⅠA期纯实性NSCLC的总体生存(overall survival, OS)。



方法

本研究回顾性分析了于2012年1月至2020年12月行肺切除术的NSCLC患者。从CT图像中提取肿瘤内和肿瘤周围区域的影像学特征。使用随机生存森林(random survival forests, RSF)和XGBoost算法开发机器学习模型,同时将Cox回归模型设为评估基准。使用整合时间依赖性曲线下面积(integrated time-dependent area under the curve, iAUC)评估模型性能,并采用5折交叉检验进行验证。



结果

共纳入642名临床ⅠA期纯实性NSCLC患者。在3748个影像学特征和34个术前临床特征中筛选出了42个特征用于拟合模型。两种机器学习模型性能均优于Cox回归模型(iAUC, 0.753 ;95% CI, 0.629-0.829)。XGBoost模型(iAUC, 0.832 ;95% CI, 0.779-0.880)性能优于RSF模型(iAUC, 0.795 ;95% CI,0.734-0.856)。XGBoos模型在低危组(5年OS:92.9%),中危组(5年OS:78.9%)和高危组(5年OS:41.7%)的OS存在显著差异(P<0.0001),显示出优异的生存分层预测性能。



结论

基于影像组学的机器学习模型可以在术前准确预测临床IA期纯实性NSCLC的OS并改善生存分层。

Figure 2.Assessment of Discrimination and Calibration for Prognostic Models.

A. Discrimination ability of three models illustrated by time-dependent AUC curve. The XGBoost model had the highest time-dependent AUC from 6 to 60 months among all models. B. Calibration curve of three models illustrated by observed and predicted survival curve. The XGBoost model predicted survival curve is mostly consistent with the observed survival curve among all models.


Figure 6.Examples of the Application of the XGBoost Model.

A. A 64-year-old female with clinical stage IA3 NSCLC shows a high predicted 5-year survival of 97%, matched by actual survival more than 56 months. B. A 72-year-old male with clinical IA3 NSCLC has a lower predicted 5-year survival of 51.4%, with actual survival at 21 months.

主要研究者


严浩吉,日本顺天堂大学呼吸器外科博士生,胸“星”外科学术团队第三届队长,主要研究方向为肺癌临床研究、人工智能与医学结合研究、肺移植基础及临床研究,发表中文论文10篇,其中发表在中文核心杂志6篇,发表英文论文共计23篇 (包括Lancet Gastroenterol Hepatol, JAMA Netw Open, J Heart Lung Transplant, J Thorac Cardiovasc Surg等),以第一作者或共同第一作者发表英文论文10篇,并多次在国内外学术年会做大会发言(包括AATS 104th Annual Meeting),参编(译)“十四五”重点出版工程《临床肺移植》等医学专著3部。

胸“星”外科学术团队简介


      胸“星”外科学术团队创立于2017年,是国内较早组建的医学本科生科研团队。秉承“Never Stop Making Challenges ”的科研精神,历经7年的建设和磨砺,团队现已成长为一个锐意进取、精诚团结的医学本科生早期科研训练团队。团队以“三全育人-科研育人”综合改革理论为指导,以培养科研兴趣为价值导向,以提升科研水平为行动目标,于国内首次提出医学本科生“科研前置”培养模式。团队积极开展医学本科生早期高水平科研训练,为建设成全国领先的医学本科生科研团队不懈努力

      团队一直致力于胸外科相关领域的基础与临床研究(肺移植、食管癌、肺癌、胸腺瘤等),现有成员共计38人。在团队指导老师田东教授的带领下,团队深入践行科研育人理论,积极投身科研平台建设,持续培养科技创新意识,不断强化医学科研能力(包含但不限于文献检索、科研选题、课题设计、论文撰写和外语应用等),最终形成了系统性的医学本科生科研训练模式,进而促进了“科研训练-成果产出”之间的良性循环。

      目前团队学生以第一作者或共同第一作者发表中英文论文共计55篇,其中,SCI收录论文29篇,中文论文26篇,负责国家级大学生科研课题26项,省级25项。受邀参加国内、国际胸心外科学术年会口头发言18次(篇),壁报展出22次(篇),参与翻译、审校专著7部,已申请国家专利5项,斩获中国国际大学生创新大赛(2023)总决赛银奖等国家级、省级各级赛事10余项,成功建立西南地区第一个标准的小动物肺移植实验室(《四川日报-川观新闻》报道)。团队成员绝大部分本科毕业后继续攻读硕士研究生(北京协和医学院、北京大学、四川大学、中南大学、中国医科大学、南京医科大学、南方医科大学、广州医科大学、温州医科大学等国内知名院校),并有8名队员本科毕业后被日本名校(京都大学、长崎大学、東北大学、名古屋大学、顺天堂大学、熊本大学、筑波大学、庆应义塾大学)直博录取。

      在团队导师的指导下,团队已在多个研究领域取得了丰硕的成果。回顾过去,展望未来,胸“星”外科学术团队会继续秉持“Never Stop Making Chanlleges”医学科研精神,为提高医学本科生科研水平,完善医学本科生早期科研训练模式,推动医学高等教育改革而不懈努力!


胸外学术时间
胸外学术时间系胸“星”外科学术团队(EATTS)下特色平台,主要共享胸外科领域最前沿的医学论文及科普。此外还分享医学论文写作的基本技能,旨在为医学专业人士(尤其是胸外专业)或大众提供网络共享平台,让大家一起 EAT 每份胸外盛宴。
 最新文章