【摘要】 目的 探究基于定量超声参数的列线图模型对乳腺癌的预测价值。方法 回顾性收集135例行超声检查且经穿刺或手术病理证实的乳腺肿瘤女性病人的临床及影像资料。将病人按7∶3比例随机分为建模集（95例）与内部验证集（40例）；另收集不同时期行超声检查且经穿刺或手术病理证实的30例乳腺肿瘤女性病人作为外部验证集。以病理结果作为“金标准”将建模集分为良性组（60例）和恶性组（35例）。采集彩色多普勒超声下的病人血流信号分级、血流参数，以及弹性成像模式下相关参数。采用多因素Logistic回归分析乳腺癌的独立危险因素，通过R软件构建乳腺癌列线图预测模型，采用受试者操作特征（ROC）曲线下面积（AUC）分析模型的预测效能。采用Hosmer-Lemeshow检验评估模型的拟合优度。结果 建模集超声诊断乳腺癌敏感度为94.29%，特异度为93.33%，准确度为93.68%。多因素Logistic回归分析显示最大血流速度（PSV）≥15.00 cm/s、最大弹性模量（Emax）≥86.4 kPa、弹性比值（Eratio）≥4.35及弹性模量标准差（Esd）≥17.0 kPa是乳腺癌的独立危险因素（均P<0.05）。ROC曲线显示建模集、内部验证集与外部验证集模型的预测效能均较好，AUC分别为0.900、0.926和0.820。校准曲线显示建模集、内部验证集与外部验证集均具有良好的预测结果，Hosmer-Lemeshow检验显示建模集、内部验证集与外部验证集的预测概率与实测概率均拟合良好（均P>0.05）。结论 基于PSV、Emax、Eratio及Esd定量超声参数构建的乳腺癌列线图预测模型具有较高的预测效能，可为乳腺癌诊断提供客观依据。更多还原

【Abstract】 Objective To explore the predictive value of a nomogram model based on quantitative ultrasound parameters for breast cancer. Methods Clinical and imaging data of 135 female patients with breast tumors confirmed by biopsy or surgical pathology who underwent ultrasound examination were retrospectively collected. Patients were randomly divided into a modeling set(95 cases) and an internal validation set(40 cases) at a ratio of 7∶3. Additionally, 30 patients with breast tumors confirmed by biopsy or surgical pathology who underwent ultrasound examination were collected as an external validation set. The pathological diagnosis results were used as the "gold standard" to divide the modeling set into a benign group(60 cases) and a malignant group(35 cases). Color doppler ultrasound data were collected, including blood flow grading and hemodynamic parameters, along with elasticity imaging parameters. Multivariate Logistic regression was used to identify independent risk factors for breast cancer. A nomogram prediction model of breast cancer was constructed using R software.The predictive performance of the model was assessed using the area under the receiver operating characteristic(ROC) curve(AUC). The goodness-of-fit of the model was evaluated using the Hosmer-Lemeshow test. Results In the modeling set, the sensitivity, specificity, and accuracy of ultrasound diagnosis for breast cancer were 94.29%, 93.33%, and 93.68% respectively.Multivariate Logistic regression analysis revealed that a peak systolic velocity(PSV) ≥15.00 cm/s, maximum elastic modulus(Emax) ≥86.4 kPa, elasticity ratio(Eratio) ≥4.35, and standard deviation of elastic modulus(Esd) ≥17.0 kPa were independent risk factors for breast cancer(all P<0.05). The ROC curve showed good predictive performance of the model in the modeling,internal validation, and external validation cohorts, with AUCs of 0.900, 0.926, and 0.820, respectively. Calibration curve demonstrated good agreement between predicted and actual outcomes in all three cohorts, and the Hosmer-Lemeshow test showed no significant deviation between predicted and observed probabilities(all P>0.05).Conclusion The breast cancer nomogram model based on PSV, Emax, Eratio and Esd quantitative ultrasound parameters has high prediction performance and can provide an objective basis for the diagnosis of breast cancer.更多还原