基于列线图的临床病理生物标志物联合多参数MRI对乳腺癌肿瘤浸润淋巴细胞表达的预测价值
该研究纳入171例浸润性导管癌患者(2023-2025年),术前接受多参数MRI,分析临床病理特征、常规MRI特征及定量参数,用多因素logistic回归筛选高/低TIL水平的独立预测因子并构建列线图。列线图在训练队列C-index为0.944、验证队列为0.964,AUC在训练队列为0.954(敏感度85.1%、特异度91.1%、准确率87.4%),验证队列为0.974(敏感度96.7%、特异度92.1%、准确率92.6%),均显著高于相应队列的单个模型。结论认为联合临床病理特征与多参数MRI参数可显著提高对乳腺癌TIL水平的预测准确性。
要点
- 列线图预测乳腺癌高/低TIL水平:训练队列AUC 0.954,验证队列AUC 0.974。
- 列线图C-index训练队列0.944、验证队列0.964。
- 独立预测因子包括组织学分级、D、D*、Ktrans和Kep。
核心结论
- 联合临床病理特征与多参数MRI参数可显著提高对乳腺癌TIL水平的预测准确性。
研究设计与样本
- 研究纳入171例经诊断的浸润性导管癌患者,术前接受多参数MRI检查(2023-2025年),分析临床病理特征、常规MRI特征及一系列定量参数。
诊断标准与关键数值
- 列线图模型在训练队列AUC为0.954,敏感度85.1%,特异度91.1%,准确率87.4%。
- 列线图模型在验证队列AUC为0.974,敏感度96.7%,特异度92.1%,准确率92.6%。
- 列线图C-index在训练队列为0.944,在验证队列为0.964。
- 列线图模型AUC显著高于相应队列的单个模型(训练队列Z=3.018-6.653,均P<0.05;验证队列Z=2.546-5.668,均P<0.05)。
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
RATIONALE AND OBJECTIVES: To investigate the value of clinicopathological features and multiparametric magnetic resonance imaging (MRI) in predicting tumour-infiltrating lymphocyte (TIL) levels in breast cancer.
MATERIALS AND METHODS: A total of 171 patients diagnosed with invasive ductal carcinoma who underwent preoperative MRI (2023-2025) were included. The analysis focused on the clinicopathological characteristics alongside conventional MRI features and a range of quantitative parameters. Multiple logistic regression analysis identified independent predictors of high and low TIL levels. A nomogram was constructed based on the multivariable logistic regression model results.
RESULTS: Logistic regression analysis identified histological grade, D, D*, Ktrans, and Kep as independent factors in the training cohort. The nomogram's C-index was 0.944 in the training cohort and 0.964 in the validation cohort. The area under the curve (AUC) of the nomogram model was 0.954 (85.1% sensitivity, 91.1% specificity, and 87.4% accuracy) in the training cohort and 0.974 (96.7% sensitivity, 92.1% specificity, and 92.6% accuracy) in the validation cohort, both significantly higher than those of the individual models in the corresponding cohorts (Z=3.018-6.653, all P<0.05 and Z=2.546-5.668, all P<0.05).
CONCLUSION: Combining clinicopathological characteristics with multiparametric MRI parameters significantly improves prediction accuracy for TIL levels in breast cancer. This integrated model holds considerable clinical potential, providing robust support for personalised treatment strategies.
引用
- Zhou Z, Sun Z, Zhao F, et al. Predictive Value of Nomogram-Based Clinicopathological Biomarkers Combined with Multiparametric MRI for Tumour-Infiltrating Lymphocyte Expression in Breast Cancer. Acad Radiol. 2025. PMID: 41046212. doi:10.1016/j.acra.2025.09.017
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。