Department of Gastrointestinal Surgery, China-Japan Friendship Hospital, Beijing, China.
General Surgery Department, The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital, Jinan, China.
Abdom Radiol (NY). 2024 Jan;49(1):3-10. doi: 10.1007/s00261-023-04029-2. Epub 2023 Oct 3.
Our study aimed to determine whether radiomics models based on contrast-enhanced computed tomography (CECT) have considerable ability to predict serosal involvement in gallbladder cancer (GBC) patients.
A total of 152 patients diagnosed with GBC were retrospectively enrolled and divided into the serosal involvement group and no serosal involvement group according to paraffin pathology results. The regions of interest (ROIs) in the lesion on all CT images were drawn by two radiologists using ITK-SNAP software (version 3.8.0). A total of 412 features were extracted from the CT images of each patient. The Mann‒Whitney U test was applied to identify features with significant differences between groups. Seven machine learning algorithms and a deep learning model based on fully connected neural networks (f-CNNs) were used for radiomics model construction. The prediction efficacy of the models was evaluated using receiver operating characteristic (ROC) curve analysis.
Through the Mann‒Whitney U test, 75 of the 412 features extracted from the CT images of patients were significantly different between groups (P < 0.05). Among all the algorithms, logistic regression achieved the highest performance with an area under the curve (AUC) of 0.944 (sensitivity 0.889, specificity 0.8); the f-CNN deep learning model had an AUC of 0.916, and the model showed high predictive power for serosal involvement, with a sensitivity of 0.733 and a specificity of 0.801.
Radiomics models based on features derived from CECT showed convincing performances in predicting serosal involvement in GBC.
本研究旨在确定基于增强 CT（CECT）的影像组学模型是否能够准确预测胆囊癌（GBC）患者的浆膜侵犯情况。
回顾性纳入了 152 名经病理证实的 GBC 患者，根据石蜡病理结果将其分为浆膜侵犯组和无浆膜侵犯组。由两位放射科医生使用 ITK-SNAP 软件（版本 3.8.0）在所有 CT 图像上勾画病灶的感兴趣区（ROI）。从每位患者的 CT 图像中提取了 412 个特征。采用 Mann-Whitney U 检验筛选组间差异有统计学意义的特征。采用 7 种机器学习算法和基于全连接神经网络（f-CNN）的深度学习模型进行影像组学模型构建。采用受试者工作特征（ROC）曲线分析评估模型的预测效能。
通过 Mann-Whitney U 检验，从患者 CT 图像中提取的 412 个特征中有 75 个在组间差异有统计学意义（P < 0.05）。在所有算法中，逻辑回归的性能最佳，曲线下面积（AUC）为 0.944（敏感度 0.889，特异度 0.8）；f-CNN 深度学习模型的 AUC 为 0.916，对浆膜侵犯具有较高的预测能力，敏感度为 0.733，特异度为 0.801。
基于 CECT 特征的影像组学模型在预测 GBC 的浆膜侵犯方面具有令人信服的表现。
