| 张丝媛,彭雨硕,石晨,等.乳腺癌新辅助治疗后内乳淋巴结临床完全缓解患者的预测模型[J].中华放射医学与防护杂志,2026,46(4):359-366.Zhang Siyuan,Peng Yushuo,Shi Chen,et al.A predictive model for internal mammary node clinical complete response after neoadjuvant therapy for breast cancer[J].Chin J Radiol Med Prot,2026,46(4):359-366 |
| 乳腺癌新辅助治疗后内乳淋巴结临床完全缓解患者的预测模型 |
| A predictive model for internal mammary node clinical complete response after neoadjuvant therapy for breast cancer |
| 投稿时间：2025-12-08 |
| DOI：10.3760/cma.j.cn112271-20251208-00424 |
| 中文关键词: 乳腺癌|内乳淋巴结|新辅助治疗|列线图|治疗反应 |
| 英文关键词:Breast cancer|Internal mammary node|Neoadjuvant therapy|Nomogram|Treatment response |
| 基金项目: |
|
| 摘要点击次数: 1570 |
| 全文下载次数: 31 |
| 中文摘要: |
| 目的 开发并验证一个整合了基线临床病理与影像学特征的列线图模型,用于预测乳腺癌患者新辅助治疗后的内乳淋巴结临床完全缓解(icCR)状态。方法 回顾性纳入了2018 年至2024 年接受治疗的208 例基线内乳淋巴结阳性的乳腺癌患者。通过十折交叉验证的最小绝对收缩和选择算子(LASSO)回归方法,从33 个候选变量中筛选预测因子,构建多因素logistic回归模型,以列线图形式呈现。采用基于1 000 次重抽样的Bootstrap法进行内部验证,并从区分度、校准度和临床实用性三个维度综合评估模型性能。结果 多因素logistic回归显示,临床N分期(cN2~3 vs. cN0~1: OR = 0.31, P = 0.001)、分子分型(管腔型 vs. HER-2富集型: OR = 0.31, P = 0.005;三阴性 vs. HER-2富集型: OR = 0.24, P = 0.005)及阳性内乳淋巴结数量(多于1 个 vs. 1 个: OR = 2.09, P = 0.033)是icCR的独立预测因素。最终模型纳入了7 个预测因素,曲线下面积(AUC) = 0.762;Bootstrap校正后AUC = 0.728。决策曲线分析证实在17%~93%的风险阈值范围内具有临床净获益。结论 构建并验证了一个个体化预测新辅助治疗后内乳淋巴结临床完全缓解的列线图模型。该模型可以为新辅助治疗后缺乏影像评估的患者提供内乳淋巴结降阶梯放疗的决策依据。 |
| 英文摘要: |
| Objective To develop and validate a nomogram model that integrates baseline clinicopathological and imaging characteristics to predict internal mammary node (IMN) clinical complete response (icCR) after neoadjuvant therapy for breast cancer patients. Methods A retrospective analysis was conducted on 208 breast cancer patients with baseline positive IMNs who were treated between 2018 and 2024. Predictors were selected from 33 candidate variables using the least absolute shrinkage and selection operator (LASSO) regression method with 10-fold cross-validation. Then, based on the selected predictors, a multivariable logistic regression model in the form of a nomogram was constructed. Internal validity was conducted for the model using 1 000 bootstrap resamples, followed by a comprehensive performance evaluation from three dimensions: discrimination, calibration, and clinical utility. Results Multivariable logistic regression analysis indicated the independent predictors of icCR included clinical N stage (cN2-3 vs. cN0-1: OR = 0.31, P = 0.001), molecular subtype (luminal vs. HER2-enriched: OR = 0.31, P = 0.005; triple-negative vs. HER2-enriched: OR = 0.24, P = 0.005), and the number of positive IMNs (> 1 vs. 1: OR = 2.09, P = 0.033) The final model incorporated seven predictors, yielding an area under the curve (AUC) of 0.762 and a bootstrap-corrected AUC of 0.728. Decision curve analysis (DCA) confirms that the occurrence of clinical net benefits across a wide risk threshold range of 17%-93%. Conclusions The nomogram model developed and validated in this study for the individualized icCR prediction following neoadjuvant therapy can provide a basis for decision-making in IMN de-escalated radiotherapy for patients for whom post-treatment imaging assessment is unavailable. |
| HTML 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|