【摘要】 目的：本研究旨在探讨利用多层螺旋CT图像上的特征信息以及临床病理指标，构建预测术前三阴性乳腺癌（triple-negative breast cancer,TNBC）患者腋窝淋巴结转移（axillary lymph node metastasis,ALNM）的列线图模型。方法：回顾性分析2020年11月至2024年10月就诊于哈尔滨医科大学附属肿瘤医院经病理证实的265例TNBC女性患者的CT图像及病理资料，以6∶4的结局比例分配为训练集（161例）和验证集（104例）。使用最小绝对收缩和选择算子（least absolute shrinkage and selection operator,LASSO）回归进行变量选择，并进行10倍交叉验证，对训练集进行Logistic回归分析，筛选ALNM的独立危险因素，构建预测TNBC患者ALNM的列线图模型。使用受试者工作特征（receiver operating characteristic,ROC）曲线、校准曲线和临床决策曲线分析（decision curve analysis,DCA）评估模型的性能。结果：经多因素Logistic回归最终确定了包括临床N分期（OR=6.789,95%CI:2.203～22.20,P=0.001）、淋巴结的CT短轴直径（OR=1.686,95%CI:1.349～2.257,P<0.001）及皮质厚度（OR=6.296,95%CI:2.170～19.31,P=0.001）在内的3个重要独立预测因子，以此构建列线图预测模型。最终训练集的和验证集的ROC曲线下面积分别为0.918(95%CI:0.860～0.977)、0.885(95%CI:0.809～0.962)。训练集和验证集的HL检验分别为P=0.609和P=0.694。校准曲线显示预测概率与实际概率基本一致。决策曲线显示训练集和验证集在0.02～0.96、0.03～0.87时，具有临床实用价值。结论：本研究基于多层螺旋CT联合临床病理特征的列线图预测模型具有良好的预测效能，为TNBC患者的术前个体化评估及临床治疗提供参考。更多还原

【Abstract】 Objective: We aimed to develop a nomogram in corporating multidetector computed tomography(MDCT) imaging features and clinicopathological indicators for the preoperative prediction of axillary lymph node metastasis(ALNM) in patients with triple-negative breast cancer(TNBC). Methods: We retrospectively analyzed data from 265 female patients with pathologically confirmed TNBC treated at Harbin Medical University Cancer Hospital between November 2020 and October 2024. Patients were randomly assigned into a training cohort(n =161) and a validation cohort(n = 104) in a 6:4 ratio. Feature selection was performed using least absolute shrinkage and selection operator(LASSO) regression with 10-fold cross-validation. Independent predictors of ALNM were identified by multivariate Logistic regression analysis,and a nomogram was constructed accordingly. Model performance was assessed using receiver operating characteristic(ROC) curves, calibration plots, and decision curve analysis(DCA). Results: Three independent predictors of ALNM were identified: clinical N-stage(odds ratio[OR] = 6.789; 95% confidence interval [CI]: 2.203-22.20; P = 0.001), short-axis diameter of lymph nodes on CT(OR = 1.686; 95% CI: 1.349-2.257; P< 0.001), and cortical thickness(OR=6.296; 95% CI: 2.170-19.310; P=0.001). The nomogram showed strong discrimination, with areas under the ROC curve(AUC) of 0.918(95% CI: 0.860-0.977) in the training cohort and 0.885(95% CI: 0.809-0.962) in the validation cohort.Calibration was confirmed by Hosmer–Lemeshow tests(P=0.609 and P=0.694 for training and validation cohorts, respectively). DCA demonstrated clinical utility across probability thresholds of 0.02-0.96 and 0.03-0.87 in the training and validation cohorts, respectively. Conclusions: This nomogram, integrating MDCT imaging features and clinical indicators, provides a practical tool for individualized preoperative risk assessment and may aid clinical decision-making in patients with TNBC.更多还原