Tumor shrinkage patterns and optimal timing of response assessment during neoadjuvant therapy for breast cancer: a study based on multiparametric MRI

乳腺癌新辅助治疗期间肿瘤缩小模式及疗效评估的最佳时机:一项基于多参数磁共振成像的研究

阅读:2

Abstract

OBJECTIVE: To investigate the laws of tumor shrinkage and to determine the optimal time window for predicting the treatment efficacy (defined as rapid tumor shrinkage) during neoadjuvant systemic therapy (NST) in breast cancer using multiparametric MRI (mp-MRI). METHODS: This retrospective study included 215 breast cancer patients (221 breasts) who underwent complete NST and received mp-MRI at baseline and after 2, 4, and 6 treatment cycles. Tumors were categorized into four shrinkage patterns using latent class trajectory modeling (LCTM) based on changes in maximum tumor diameter on dynamic contrast-enhanced MRI (DCE-MRI). Clinical, pathological, and imaging features were analyzed to identify independent predictors of shrinkage rate using multivariate logistic regression. A predictive model was then constructed by combining the independent predictors. Receiver operating characteristic (ROC) curves were plotted to evaluate the model’s performance, and DeLong’s test was used to compare the area under the curve (AUC) between different models to assess predictive efficacy. RESULTS: Using LCTM, based on the change patterns of maximum lesion diameter on the first phase of DCE-MRI, all 221 breast cancer cases were classified into four distinct shrinkage types. Multivariate logistic regression analysis identified HER2 positivity, baseline peritumoral edema, DWI isointense signal at NST cycles 2/4/6, a lower early enhancement rate at cycle 6, and imaging features assessed at the fourth cycle—including non-mass enhancement, concentric shrinkage, and higher ADC values—as independent predictors of rapid tumor shrinkage. A combined model incorporating clinicopathological features and MRI characteristics at NST cycle 4 achieved the highest predictive performance (AUC = 0.814), which was comparable to that of the model based on cycle 6 imaging features (AUC = 0.759, P = 0.243). To determine the best-performing model, predictive models combining clinicopathological and baseline features with NST cycle 4 or cycle 6 imaging features were constructed. The AUCs were 0.835 and 0.796, respectively, with no significant difference (P = 0.319). CONCLUSION: Distinct laws of tumor shrinkage were revealed during NST; the fourth cycle was identified as the critical time window for accurate prediction of rapid shrinkage using multiparametric MRI. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s40644-025-00971-0.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。