多中心评估一种用于 MRI 直肠癌淋巴结诊断的弱监督深度学习模型。
Multicenter Evaluation of a Weakly Supervised Deep Learning Model for Lymph Node Diagnosis in Rectal Cancer at MRI.
机构信息
From the Department of Medical Imaging, Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences, Suzhou, China (W.X., J.J., R.Z., X.G.); Department of Radiology, Shanxi Province Cancer Hospital/Shanxi Hospital Affiliated to Cancer Hospital, Chinese Academy of Medical Sciences/Cancer Hospital Affiliated to Shanxi Medical University, Taiyuan 030013, China (D.L., J.Z., R.S., X.Y., X.G., Y.C.); Department of Radiology, the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China (W.H.); Department of Radiology, University of Wisconsin School of Medicine and Public Health, E3/311 Clinical Science Center, Madison, Wis (P.J.P.); Department of Radiology, Fudan University Shanghai Cancer Center, Shanghai, China (T.T.); Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, China (T.T.); and Guangdong Provincial Key Laboratory of Artificial Intelligence in Medical Image Analysis and Application, Guangzhou, China (Y.C.).
出版信息
Radiol Artif Intell. 2024 Mar;6(2):e230152. doi: 10.1148/ryai.230152.
Purpose To develop a Weakly supervISed model DevelOpment fraMework (WISDOM) model to construct a lymph node (LN) diagnosis model for patients with rectal cancer (RC) that uses preoperative MRI data coupled with postoperative patient-level pathologic information. Materials and Methods In this retrospective study, the WISDOM model was built using MRI (T2-weighted and diffusion-weighted imaging) and patient-level pathologic information (the number of postoperatively confirmed metastatic LNs and resected LNs) based on the data of patients with RC between January 2016 and November 2017. The incremental value of the model in assisting radiologists was investigated. The performances in binary and ternary N staging were evaluated using area under the receiver operating characteristic curve (AUC) and the concordance index (C index), respectively. Results A total of 1014 patients (median age, 62 years; IQR, 54-68 years; 590 male) were analyzed, including the training cohort ( = 589) and internal test cohort ( = 146) from center 1 and two external test cohorts (cohort 1: 117; cohort 2: 162) from centers 2 and 3. The WISDOM model yielded an overall AUC of 0.81 and C index of 0.765, significantly outperforming junior radiologists (AUC = 0.69, < .001; C index = 0.689, < .001) and performing comparably with senior radiologists (AUC = 0.79, = .21; C index = 0.788, = .22). Moreover, the model significantly improved the performance of junior radiologists (AUC = 0.80, < .001; C index = 0.798, < .001) and senior radiologists (AUC = 0.88, < .001; C index = 0.869, < .001). Conclusion This study demonstrates the potential of WISDOM as a useful LN diagnosis method using routine rectal MRI data. The improved radiologist performance observed with model assistance highlights the potential clinical utility of WISDOM in practice. MR Imaging, Abdomen/GI, Rectum, Computer Applications-Detection/Diagnosis . Published under a CC BY 4.0 license.
目的
开发一种弱监督模型开发框架（WISDOM）模型，以构建一种基于术前 MRI 数据和术后患者病理信息的直肠肿瘤患者淋巴结（LN）诊断模型。
材料与方法
本回顾性研究使用 MRI（T2 加权和弥散加权成像）和患者水平的病理信息（术后证实的转移性 LN 数量和切除的 LN 数量），基于 2016 年 1 月至 2017 年 11 月直肠肿瘤患者的数据，构建了 WISDOM 模型。研究了模型在辅助放射科医生方面的增量价值。通过接受者操作特征曲线下面积（AUC）和一致性指数（C 指数）评估模型在二元和三元 N 分期中的性能。
结果
共分析了 1014 例患者（中位年龄为 62 岁；IQR：54-68 岁；590 例男性），包括中心 1 的训练队列（n=589）和内部测试队列（n=146），以及中心 2 和 3 的两个外部测试队列（队列 1：n=117；队列 2：n=162）。WISDOM 模型的总体 AUC 为 0.81，C 指数为 0.765，显著优于初级放射科医生（AUC=0.69，<.001；C 指数=0.689，<.001），与高级放射科医生的表现相当（AUC=0.79，=.21；C 指数=0.788，=.22）。此外，该模型显著提高了初级放射科医生（AUC=0.80，<.001；C 指数=0.798，<.001）和高级放射科医生（AUC=0.88，<.001；C 指数=0.869，<.001）的性能。
结论
本研究证明了 WISDOM 作为一种使用常规直肠 MRI 数据进行 LN 诊断的有用方法的潜力。观察到模型辅助提高了放射科医生的表现，突出了 WISDOM 在实践中的潜在临床应用价值。
磁共振成像，腹部/胃肠道，直肠，计算机应用-检测/诊断。根据 CC BY 4.0 许可发布。
Zotero 插件