Chen Weicui, Zheng Kaiyi, Yuan Wenjing, Jia Ziqi, Wu Yuankui, Duan Xiaohui, Yang Wei, Wen Zhibo, Zhong Liming, Liu Xian
Radiology Department, The Second Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou, 510120, Guangdong, China.
School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, Guangdong, China.
Radiol Med. 2025 Feb;130(2):214-225. doi: 10.1007/s11547-024-01909-5. Epub 2024 Nov 26.
To develop and validate deep learning (DL) models using preoperative contrast-enhanced CT images for tumor auto-segmentation and microsatellite instability (MSI) prediction in colorectal cancer (CRC).
Patients with CRC who underwent surgery or biopsy between January 2018 and April 2023 were retrospectively enrolled. Mismatch repair protein expression was determined via immunohistochemistry or fluorescence multiplex polymerase chain reaction-capillary electrophoresis. Manually delineated tumor contours using arterial and venous phase CT images by three abdominal radiologists are served as ground truth. Tumor auto-segmentation used nnU-Net. MSI prediction employed ViT or convolutional neural networks models, trained and validated with arterial and venous phase images (image model) or combined clinical-pathological factors (combined model). The segmentation model was evaluated using patch coverage ratio, Dice coefficient, recall, precision, and F1-score. The predictive models' efficacy was assessed using areas under the curves and decision curve analysis.
Overall, 2180 patients (median age: 61 years ± 17 [SD]; 1285 males) were divided into training (n = 1159), validation (n = 289), and independent external test (n = 732) groups. High-level MSI status was present in 435 patients (20%). In the external test set, the segmentation model performed well in the arterial phase, with patch coverage ratio, Dice coefficient, recall, precision, and F1-score values of 0.87, 0.71, 0.72, 0.74, and 0.71, respectively. For MSI prediction, the combined models outperformed the clinical model (AUC = 0.83 and 0.82 vs 0.67, p < 0.001) and two image models (AUC = 0.75 and 0.77, p < 0.001). Decision curve analysis confirmed the higher net benefit of the combined model compared to the other models across probability thresholds ranging from 0.1 to 0.45.
DL enhances tumor segmentation efficiency and, when integrated with contrast-enhanced CT and clinicopathological factors, exhibits good diagnostic performance in predicting MSI in CRC.
利用术前对比增强CT图像开发并验证深度学习（DL）模型，用于结直肠癌（CRC）的肿瘤自动分割和微卫星不稳定性（MSI）预测。
回顾性纳入2018年1月至2023年4月期间接受手术或活检的CRC患者。通过免疫组织化学或荧光多重聚合酶链反应 - 毛细管电泳测定错配修复蛋白表达。由三位腹部放射科医生使用动脉期和静脉期CT图像手动勾勒肿瘤轮廓作为金标准。肿瘤自动分割采用nnU-Net。MSI预测采用ViT或卷积神经网络模型，使用动脉期和静脉期图像（图像模型）或联合临床病理因素（联合模型）进行训练和验证。分割模型使用补丁覆盖率、Dice系数、召回率、精确率和F1分数进行评估。预测模型的效能使用曲线下面积和决策曲线分析进行评估。
总体而言，2180例患者（中位年龄：61岁±17[标准差]；1285例男性）被分为训练组（n = 1159）、验证组（n = 289）和独立外部测试组（n = 732）。435例患者（20%）存在高水平MSI状态。在外部测试集中，分割模型在动脉期表现良好，补丁覆盖率、Dice系数、召回率、精确率和F1分数分别为0.87、0.71、0.72、0.74和0.71。对于MSI预测，联合模型优于临床模型（曲线下面积分别为0.83和0.82对0.67，p < 0.001）和两个图像模型（曲线下面积分别为0.75和0.77，p < 0.001）。决策曲线分析证实，在概率阈值从0.1到0.45的范围内，联合模型比其他模型具有更高的净效益。
DL提高了肿瘤分割效率，并且与对比增强CT和临床病理因素相结合时，在预测CRC的MSI方面表现出良好的诊断性能。
