Development of a Deep Learning Model to Identify Lymph Node Metastasis on Magnetic Resonance Imaging in Patients With Cervical Cancer
Development of a Deep Learning Model to Identify Lymph Node Metastasis on Magnetic Resonance Imaging in Patients With Cervical Cancer
复制标题
开发深度学习模型以通过磁共振成像识别宫颈癌患者的淋巴结转移
DOI：
10.1001/jamanetworkopen.2020.11625
复制
发表时间：
2020-07-24
影响因子：
13.8
通讯作者：
Tian, Jie
中科院分区：
文献类型：
--
作者：
Wu, Qingxia;Wang, Shuo;Tian, Jie
关键问题：深度学习能否识别宫颈癌的术前非侵袭性淋巴转移诊断？结果这项诊断性研究包括479名患者，建立了一个深度学习模型，用于术前在磁共振成像上无创识别淋巴结转移，在独立验证队列中实现了接收者操作特征曲线下面积为0.933。预测的淋巴结转移概率与宫颈癌的预后密切相关。这项研究的结果表明，深度学习可以作为一种术前非侵入性工具用于诊断宫颈癌的淋巴转移。
Key Points Question Can deep learning identify preoperative noninvasive lymph node metastasis diagnosis in cervical cancer? Findings This diagnostic study including a total of 479 patients developed a deep learning model to preoperatively and noninvasively identify lymph node metastasis on magnetic resonance imaging, achieving an area under the receiver operating characteristic curve of 0.933 in the independent validation cohort. The predicted lymph node metastasis probability was significantly associated with prognosis of cervical cancer. Meaning Findings from this study suggest that deep learning can be used as a preoperative noninvasive tool for diagnosing lymph node metastasis in cervical cancer.