基于CT增强图像的影像组学与临床特征融合 模型在直肠癌淋巴结转移中的诊断价值

The Diagnostic Value of a Combined Enhanced CT Image Radiomics and Clinical Feature Model for Lymph Node Metastasis in Rectal Cancer

摘要: 目的：本研究旨在开发基于CT影像组学与临床特征的融合模型，以精准预测直肠癌患者的淋巴结转移状态。方法：研究纳入220例患者，以7:3随机分为训练集和测试集，通过多级质量控制标注感兴趣区域，提取影像组学特征，并采用10种机器学习算法构建并整合模型，再利用泛堆叠化将模型融合，以AUC值验证模型的判别能力，以决策分析曲线验证模型的临床实用性。结果：联合模型在训练集和测试集中的AUC分别为0.935和0.875，稳定优于单一影像组学或临床模型，且决策曲线分析证实其具有更高的临床净获益。结论：影像组学与临床特征模型融合有效提升了淋巴结转移的术前判别能力，为临床决策提供了客观、定量的辅助工具。

Abstract: Objective: This study aimed to develop an integrated model based on CT radiomics and clinical features to accurately predict the lymph node metastasis (LNM) status in patients with rectal cancer. Methods: The study enrolled 220 patients, randomly divided into training and test sets in a 7:3 ratio. Multi-level quality control was applied to annotate the regions of interest (ROI). Radiomic features were extracted, and 10 different machine learning algorithms were used to build and subsequently integrate the models via super-stacking. The discriminative ability of the models was validated using the Area Under the Curve (AUC) value, and their clinical utility was assessed using Decision Curve Analysis (DCA). Results: The integrated model achieved AUCs of 0.935 in the training set and 0.875 in the test set, consistently outperforming both the standalone radiomics model and the clinical feature model. DCA further confirmed that the integrated model provided a higher net clinical benefit. Conclusion: The integration of radiomics and clinical features effectively enhanced the preoperative discriminative ability for LNM status, providing an objective and quantitative tool to aid clinical decision-making.