Kang Wenyan, Ye Xiaoqin, Deng Wenming, Zhong Yihong, Li Xiaojun, Luo Dehong
Department of Radiology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital & Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shenzhen, China.
Medical Affairs Department, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital & Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shenzhen, China.
Transl Cancer Res. 2025 Jun 30;14(6):3400-3409. doi: 10.21037/tcr-24-2084. Epub 2025 Jun 24.
In colorectal cancer, the gene mutations are associated with adverse clinical outcomes and therapeutic resistance. Radiomics, a quantitative imaging analysis approach, enables high-throughput extraction of tumor features from computed tomography (CT) scans to develop predictive models. This study aimed to explore the predictive value of CT imaging model for gene mutation in patients with colorectal cancer and provide a reference for clinical practice.
Patients with colorectal cancer treated in National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital & Shenzhen Hospital from May 2022 to December 2023 were divided into mutation group and non-mutation group according to the occurrence of gene mutation. Univariate analysis of the influencing factors of gene mutation and multivariate logistics regression analysis of the influencing factors of gene mutation were used to construct a CT-based radiomics nomogram. Receiver operating characteristic (ROC) curve analysis of the predictive value of CT-based imaging model for gene mutation in patients with colorectal cancer was conducted.
In total, 100 patients were selected. Univariate analysis and multivariate logistics regression analysis showed that lymph node metastasis, tumor differentiation, tumor invasion depth and tumor size were independent influencing factors of gene mutation. Based on the influencing factors, the CT-based radiomics nomogram was constructed. The ROC analysis results showed that the area under the curve (AUC) of training and verification set of the CT-based radiomics model for predicting gene mutations in patients with colorectal cancer was 0.826 and 0.670.
In this study, the CT-based radiomics nomogram was constructed and its value was analyzed. The imaging model of CT is of high value in predicting gene mutation in patients with colorectal cancer.
在结直肠癌中，[具体基因]突变与不良临床结局和治疗耐药性相关。放射组学作为一种定量成像分析方法，能够从计算机断层扫描（CT）图像中高通量提取肿瘤特征以建立预测模型。本研究旨在探讨CT成像模型对结直肠癌患者[具体基因]突变的预测价值，为临床实践提供参考。
选取2022年5月至2023年12月在国家癌症中心/国家癌症临床研究中心/中国医学科学院肿瘤医院深圳医院接受治疗的结直肠癌患者，根据[具体基因]突变情况分为突变组和非突变组。采用单因素分析[具体基因]突变的影响因素，并进行多因素logistic回归分析[具体基因]突变的影响因素，构建基于CT的放射组学列线图。对基于CT的成像模型预测结直肠癌患者[具体基因]突变的价值进行受试者操作特征（ROC）曲线分析。
共纳入100例患者。单因素分析和多因素logistic回归分析显示，淋巴结转移、肿瘤分化程度、肿瘤浸润深度和肿瘤大小是[具体基因]突变的独立影响因素。基于这些影响因素，构建了基于CT的放射组学列线图。ROC分析结果显示，基于CT的放射组学模型预测结直肠癌患者[具体基因]突变的训练集和验证集曲线下面积（AUC）分别为0.826和0.670。
本研究构建了基于CT的放射组学列线图并分析了其价值。CT成像模型在预测结直肠癌患者[具体基因]突变方面具有较高价值。
