Scholarlink

<p>Breast cancer, a heterogeneous malignancy, complicates predicting neoadjuvant chemotherapy (NAC) response, crucial for personalized treatment. Multiparametric magnetic resonance imaging (mpMRI)-based radiomics, combined with machine learning (ML), uses quantitative imaging features to enhance prediction accuracy non-invasively.&#xa0;This systematic review and meta-analysis evaluates mpMRI-based radiomics and ML for predicting NAC response in breast cancer, focusing on predictive performance, radiomic features, ML algorithms, and methodological limitations.&#xa0;Registered on PROSPERO (CRD420251109403) and following PRISMA guidelines, we searched PubMed, Embase, Cochrane Library, Scopus, Web of Science, and IEEE Xplore for studies using mpMRI-based radiomics and ML to predict NAC response in adult breast cancer patients. Eligible studies included mpMRI sequences (e.g., T1-weighted, T2-weighted, DWI, DCE) and ML techniques. We extracted data on study characteristics, radiomic features, ML algorithms, and predictive performance (e.g., AUC, sensitivity, specificity). Quality was assessed using QUADAS-2 and Radiomics Quality Score. A random-effects meta-analysis pooled AUCs for pathological complete response (pCR) prediction.&#xa0;Eight studies (six retrospective, two prospective, <i>n</i> = 22–328) were included. mpMRI sequences involved DCE, DWI, and T2-weighted imaging, with radiomic features like texture (entropy, kurtosis), ADC, and morphological parameters. ML algorithms, from logistic regression to deep neural networks, yielded AUCs of 0.609–0.94. The pooled AUC was 0.803 (95% CI: 0.682–0.925). Limitations included small sample sizes, lack of external validation, and variable imaging protocols.&#xa0;mpMRI-based radiomics with ML shows potential for predicting NAC response, but methodological inconsistencies necessitate standardized, multicenter studies for clinical translation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting Neoadjuvant Chemotherapy Response in Breast Cancer Using Multiparametric MRI-Based Radiomics and Machine Learning: A Systematic Review and Meta-Analysis of 1,469 Patients

  • 期刊论文

  • 发表日期：

    2025年7月1日

查看全文
  • Mohammadreza Elhaie,
  • Abolfazl Koozari,
  • Marzieh Sheikh,
  • Iraj Abedi

摘要

Breast cancer, a heterogeneous malignancy, complicates predicting neoadjuvant chemotherapy (NAC) response, crucial for personalized treatment. Multiparametric magnetic resonance imaging (mpMRI)-based radiomics, combined with machine learning (ML), uses quantitative imaging features to enhance prediction accuracy non-invasively. This systematic review and meta-analysis evaluates mpMRI-based radiomics and ML for predicting NAC response in breast cancer, focusing on predictive performance, radiomic features, ML algorithms, and methodological limitations. Registered on PROSPERO (CRD420251109403) and following PRISMA guidelines, we searched PubMed, Embase, Cochrane Library, Scopus, Web of Science, and IEEE Xplore for studies using mpMRI-based radiomics and ML to predict NAC response in adult breast cancer patients. Eligible studies included mpMRI sequences (e.g., T1-weighted, T2-weighted, DWI, DCE) and ML techniques. We extracted data on study characteristics, radiomic features, ML algorithms, and predictive performance (e.g., AUC, sensitivity, specificity). Quality was assessed using QUADAS-2 and Radiomics Quality Score. A random-effects meta-analysis pooled AUCs for pathological complete response (pCR) prediction. Eight studies (six retrospective, two prospective, n = 22–328) were included. mpMRI sequences involved DCE, DWI, and T2-weighted imaging, with radiomic features like texture (entropy, kurtosis), ADC, and morphological parameters. ML algorithms, from logistic regression to deep neural networks, yielded AUCs of 0.609–0.94. The pooled AUC was 0.803 (95% CI: 0.682–0.925). Limitations included small sample sizes, lack of external validation, and variable imaging protocols. mpMRI-based radiomics with ML shows potential for predicting NAC response, but methodological inconsistencies necessitate standardized, multicenter studies for clinical translation.