【摘要】 目的 评估基于Kaiser评分的MRI影像特征列线图模型术前预测肿块型乳腺癌脉管侵犯的价值。方法 回顾性分析经手术病理证实的345例肿块型浸润性乳腺癌患者临床、病理、影像学和Kaiser评分资料，按照7∶3随机分为训练集（n=242）和验证集（n=103）。应用单因素和多因素Logistic回归模型分析肿块型乳腺癌脉管侵犯的独立危险因素并构建列线图预测模型，通过受试者工作特征（receiver operating characteristic,ROC）曲线、校准曲线和临床决策曲线评估模型效能。结果 单因素Logistic回归分析发现，肿瘤最大直径、Kaiser评分、扩散加权成像信号、形状和相关受侵征象与肿块型乳腺癌脉管侵犯相关（均P<0.05）；进一步的多因素Logistic回归分析显示，Kaiser评分≥6分、扩散加权成像高信号、形状不规则和存在相关受侵征象是肿块型乳腺癌脉管侵犯的独立危险因素（均P<0.05）。Kaiser评分联合扩散加权成像信号、形状和相关侵犯征象构建的肿块型乳腺癌脉管侵犯列线图预测模型在训练集和验证集中的ROC曲线下面积（area under the ROC curve,AUC）分别为0.899(95%CI:0.859～0.939)和0.827(95%CI:0.744～0.909)；训练集中特异性为0.845，敏感性为0.840；验证集中的特异性为0.787，敏感性为0.750；校准曲线和Hosmer-Lemeshow拟合优度检验结果表明列线图模型一致性较好；临床决策曲线结果显示列线图预测肿块型乳腺癌脉管侵犯可获得较高收益。结论 本研究构建的基于Kaiser评分的MRI影像特征列线图模型有助于术前预测肿块型乳腺癌脉管侵犯，并且该模型具有较高的预测效能，可为临床术前评估肿块型乳腺癌脉管侵犯提供参考依据。更多还原

【Abstract】 Objective To evaluate the value of Kaiser score-based MRI feature nomogram model in preoperative prediction of vascular invasion in mass breast cancer.MethodsThe data of clinical, pathological, imaging and Kaiser score of 345 patients with mass invasive breast cancer confirmed by surgical pathology were retrospectively analyzed. The patients were randomly divided into training set(n=242)and validation set(n=103) according to a ratio of 7∶3. Univariable and multivariable Logistic regression models were used to analyze the independent risk factors of vascular invasion in mass breast cancer, and to construct the nomogram prediction model. The efficacy of the model was evaluated by receiver operating characteristic(ROC) curve, calibration curve and clinical decision curve.ResultsUnivariable Logistic regression analysis showed that the maximum diameter, Kaiser score, diffusion-weighted imaging signal, shape and related invasion signs were all associated with vascular invasion in mass breast cancer(all P<0.05). Further multivariable Logistic regression analysis showed that Kaiser score≥6 scores, hyperintensity on diffusion-weighted imaging, irregular shape, and presence of related invasion signs were independent risk factors for vascular invasion in mass breast cancer(all P<0.05). The area under the ROC curve(AUC)of the nomogram prediction model for mass breast cancer constructed by Kaiser score combined with diffusion-weighted imaging signals,shape and related invasion signs were 0.899(95%CI: 0.859-0.939) and 0.827(95%CI: 0.744-0.909) in the training set and validation set, respectively. The specificity and sensitivity were 0.845 and 0.840, respectively, in the training set, and 0.787 and 0.750, respectively,in the validation set. The calibration curve and Hosmer-Lemeshow test showed that the nomogram model had good consistency. Clinical decision curve results showed that the nomogram could predict the vascular invasion of mass breast cancer with higher benefit.Conclusions The nomogram model of MRI image features based on Kaiser score constructed in this study is helpful for preoperative prediction of vascular invasion of mass breast cancer, and the model has high predictive efficiency, which provides a reference for the clinical preoperative prediction of vascular invasion of mass breast cancer.更多还原