联合常规磁共振成像参数与临床病理数据区分乳腺癌三层次人表皮生长因子受体2状态
回顾性纳入211例浸润性乳腺癌患者,按HER2状态分为零表达、低表达和阳性三组,利用常规动态增强乳腺MRI参数联合临床病理数据建立模型,区分HER2零表达/低表达与阳性(任务1)及零表达与低表达(任务2)。联合模型在训练集任务1和任务2的AUC分别为0.836和0.874,测试集分别为0.845和0.805,提示联合模型可能有助于区分乳腺癌HER2表达状态。
要点
- ADC和Ki-67状态是区分HER2零表达/低表达与阳性及零表达与低表达的独立相关因素。
- 联合常规MRI参数与临床病理数据的模型在测试集区分HER2零表达/低表达与阳性的AUC为0.845。
- 联合模型在测试集区分HER2零表达与低表达的AUC为0.805。
核心结论
- 联合常规MRI参数与临床病理数据的模型可能有助于区分乳腺癌HER2表达状态,可能辅助在无荧光原位杂交结果的患者中选择HER2靶向治疗对象。
研究设计与样本
- 回顾性评估211例浸润性乳腺癌患者,按HER2状态分为零表达、低表达和阳性组,患者接受常规动态增强乳腺MRI,按7:3随机分为训练集和测试集,采用单因素和多因素logistic回归筛选预测因子,ROC曲线评估模型区分性能。
诊断标准与关键数值
- 任务1(区分HER2零表达/低表达[n=129]与HER2阳性[n=82])中ADC和Ki-67状态的OR分别为41.22和5.68;任务2(区分HER2零表达[n=90]与HER2低表达[n=57])中ADC和Ki-67状态的OR分别为0.02和0.29;联合模型训练集任务1和任务2的AUC分别为0.836和0.874,测试集任务1和任务2的AUC分别为0.845和0.805。
临床应用与阅片要点
- 作者认为联合常规MRI参数与临床病理数据的模型可能有助于区分乳腺癌HER2表达状态,可能辅助在无荧光原位杂交结果的患者中选择HER2靶向治疗对象。
局限性
- 原文未报告局限性
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
AIM: To assess the three-tiered human epidermal growth factor receptor 2 (HER2) classification of breast cancer (BC) patients based on conventional magnetic resonance imaging (MRI) parameters combined with clinicopathologic data.
MATERIALS AND METHODS: 211 patients with invasive BC were retrospectively evaluated and divided into the HER2-zero, HER2-low, and HER2-positive BC groups. Patients underwent conventional dynamic contrast-enhanced breast MRI. Radiologists assessed clinicopathologic features and measured the apparent diffusion coefficient (ADC) and haemodynamic parameters to differentiate HER2-zero/-low (n=129) from HER2-positive (n=82) BC (task 1) and then HER2-zero (n=90) from HER2-low (n=57) BC (task 2). Patients were randomly assigned to the training and test sets at a ratio of 7:3. Univariate and multivariate logistic regression analyses were applied to select the most useful predictors. Receiver operating characteristic curve analysis was applied to evaluate the discriminative performance of the models.
RESULTS: The ADC and Ki-67 status were independently associated factors both for task 1 (OR: 41.22, 5.68) and task 2 (OR: 0.02, 0.29). The models established combining conventional MRI parameters with clinicopathologic data in the training set for task 1 and task 2 yielded an area under the curve (AUC) of 0.836 and 0.874, respectively, and demonstrated effective prediction in the test set, with the AUC of 0.845 for task 1 and an AUC of 0.805 for task 2, respectively.
CONCLUSION: Models combining conventional magnetic resonance imaging (MRI) parameters and clinicopathologic data could be valuable for differentiating BC HER2 expression, which may aid in selecting patients for HER2-targeted therapies in those without fluorescence in situ hybridisation results.
引用
- Liu W, Liu C, Yang Y, et al. Combining conventional magnetic resonance imaging (MRI) parameters with clinicopathologic data for differentiation of the three-tiered human epidermal growth factor receptor 2 (HER2) status in breast cancer. Clin Radiol. 2025. PMID: 40479855. doi:10.1016/j.crad.2025.106955
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。