Share：
Share this content in WeChat
X
[Chinese] [PDF] 9 1
• Clinical Article •
Value of DCE-MRI-based tumor heterogeneity quantification and deep learning in predicting neoadjuvant chemotherapy response in breast cancer
ZHANG Qing  CHEN Jiming  WU Lili  DING Jun  YE Hui  XIA Yi  JIANG Xuan  JIAO Nanlin 

DOI:10.12015/issn.1674-8034.2026.01.007.


[Abstract] Objective To explore the value of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI)-based tumor heterogeneity quantification integrated with deep learning (DL) in predicting the pathological complete response of neoadjuvant chemotherapy (NAC) for breast cancer.Materials and Methods The clinical and imaging data of 179 patients with pathologically confirmed breast cancer at the First Affiliated Hospital of Wannan Medical College from January 2019 to January 2025 were retrospectively collected. Among them, 58 patients achieved pathological complete response (pCR) after NAC, and 121 patients achieved non-pathological complete response (non-pCR). The patients were randomly divided into a training group (n = 125) and a validation group (n = 54) at a ratio of 7∶3. All patients underwent MRI examination before NAC. The ITK-SNAP software was used to manually delineate the region of interest (ROI) layer by layer and perform three-dimensional fusion. The Gaussian mixture model (GMM) was used for cluster analysis, and the Bayesian information criterion (BIC) was used to determine the sub-regions of the tumor lesions. The intratumoral heterogeneity score (ITH-score) was calculated, and a habitat imaging model was established. The PyRadiomics package in Python software was used to extract the traditional radiomics features of the whole tumor, and the ViT deep learning model was used to extract the deep learning features. The minimum redundancy maximum relevance (mRMR) and least absolute shrinkage and selection operator (LASSO) regression methods were used for feature dimensionality reduction and screening. A traditional radiomics model and a deep learning model were constructed respectively, and the quantitative score of each patient was calculated according to the feature weights in the models. Multivariate logistic regression analysis was used to construct a clinical model and a combined model. Receiver operating characteristic (ROC) curves were drawn to evaluate the predictive efficacy of each model. The DeLong test was used to compare the efficacy of each model, and decision curve analysis (DCA) was used to analyze the clinical benefits of the models. The SHAP method was used to analyze the importance of each feature in the combined model.Results The AUC [95% (confidence interval, CI)] values of the clinical model, traditional radiomics model, deep learning model, habitat imaging model, and combined model in predicting pCR after NAC in the training group were 0.864 (0.832 to 0.895), 0.776 (0.745 to 0.807), 0.728 (0.703 to 0.752), 0.823 (0.785 to 0.881), and 0.943 (0.903 to 0.983) respectively, and in the validation group were 0.732 (0.684 to 0.781), 0.634 (0.589 to 0.679), 0.757 (0.720 to 0.791), 0.750 (0.690 to 0.840), and 0.875 (0.821 to 0.929) respectively. The combined model had the best predictive performance. The DCA results showed that the clinical benefit of the combined model was higher than that of the clinical model and other radiomics models. In the SHAP method, the importance of the ITH-score was higher than that of the molecular subtype. The larger the SHAP value, the more the prediction result tended to pCR.Conclusions The combined model based on DCE-MRI heterogeneity quantitative analysis and deep learning demonstrates superior predictive performance for pCR in breast cancer patients after NAC, which holds clinical application value for early prediction of pCR after NAC and contributes to clinical diagnosis and treatment management of breast cancer.
[Keywords] breast cancer；neoadjuvant chemotherapy；magnetic resonance imaging；habitat imaging；deep learning；radiomics

ZHANG Qing1   CHEN Jiming1*   WU Lili1   DING Jun1   YE Hui1   XIA Yi1   JIANG Xuan1   JIAO Nanlin2  

1 Medical Imaging Center, the First Affiliated Hospital of Wannan Medical College, Wuhu 241001, China

2 Department of Pathology, the First Affiliated Hospital of Wannan Medical College, Wuhu 241001, China

Corresponding author: CHEN J M, E-mail: yjsyycjm@126.com

Conflicts of interest   None.

Received  2025-11-11
Accepted  2026-01-04
DOI: 10.12015/issn.1674-8034.2026.01.007
DOI:10.12015/issn.1674-8034.2026.01.007.

[1]
KORDE L A, SOMERFIELD M R, CAREY L A, et al. Neoadjuvant chemotherapy, endocrine therapy, and targeted therapy for breast cancer: ASCO guideline[J]. J Clin Oncol, 2021, 39(13): 1485-1505. DOI: 10.1200/JCO.20.03399.
[2]
HAQUE W, VERMA V, HATCH S, et al. Response rates and pathologic complete response by breast cancer molecular subtype following neoadjuvant chemotherapy[J]. Breast Cancer Res Treat, 2018, 170(3): 559-567. DOI: 10.1007/s10549-018-4801-3.
[3]
FAYANJU O M, REN Y, THOMAS S M, et al. The clinical significance of breast-only and node-only pathologic complete response (pCR) after neoadjuvant chemotherapy (NACT): a review of 20, 000 breast cancer patients in the national cancer data base (NCDB)[J]. Ann Surg, 2018, 268(4): 591-601. DOI: 10.1097/SLA.0000000000002953.
[4]
DAGOGO-JACK I, SHAW A T. Tumour heterogeneity and resistance to cancer therapies[J]. Nat Rev Clin Oncol, 2018, 15(2): 81-94. DOI: 10.1038/nrclinonc.2017.166.
[5]
QIAN F S, MAO Y J, DONG J, et al. Development of a multiparametric model for predicting the response to neoadjuvant chemotherapy in breast cancer[J]. Transl Cancer Res, 2024, 13(2): 558-568. DOI: 10.21037/tcr-23-770.
[6]
THERASSE P, ARBUCK S G, EISENHAUER E A, et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada[J]. J Natl Cancer Inst, 2000, 92(3): 205-216. DOI: 10.1093/jnci/92.3.205.
[7]
LIU Z Y, LI Z L, QU J R, et al. Radiomics of multiparametric MRI for pretreatment prediction of pathologic complete response to neoadjuvant chemotherapy in breast cancer: a multicenter study[J]. Clin Cancer Res, 2019, 25(12): 3538-3547. DOI: 10.1158/1078-0432.CCR-18-3190.
[8]
MA X L, XIA L M, CHEN J, et al. Development and validation of a deep learning signature for predicting lymph node metastasis in lung adenocarcinoma: comparison with radiomics signature and clinical-semantic model[J]. Eur Radiol, 2023, 33(3): 1949-1962. DOI: 10.1007/s00330-022-09153-z.
[9]
ZHANG Z, LUO T C, YAN M, et al. Voxel-level radiomics and deep learning for predicting pathologic complete response in esophageal squamous cell carcinoma after neoadjuvant immunotherapy and chemotherapy[J/OL]. J Immunother Cancer, 2025, 13(3): e011149 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/40090670/. DOI: 10.1136/jitc-2024-011149.
[10]
YIN S, MING J G, CHEN H, et al. Integrating deep learning and radiomics for preoperative glioma grading using multi-center MRI data[J/OL]. Sci Rep, 2025, 15(1): 36756 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/41120521/. DOI: 10.1038/s41598-025-20711-5.
[11]
ZHANG X, SHEN Y Y, SU G H, et al. A dynamic contrast-enhanced MRI-based vision transformer model for distinguishing HER2-zero, -low, and-positive expression in breast cancer and exploring model interpretability[J/OL]. Adv Sci, 2025, 12(33): e03925 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/40488332/. DOI: 10.1002/advs.202503925.
[12]
HUANG Y, WANG X X, CAO Y, et al. Nomogram for predicting neoadjuvant chemotherapy response in breast cancer using MRI-based intratumoral heterogeneity quantification[J/OL]. Radiology, 2025, 315(1): e241805 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/40232145/. DOI: 10.1148/radiol.241805.
[13]
LIU H F, WANG M, LU Y J, et al. CEMRI-based quantification of intratumoral heterogeneity for predicting aggressive characteristics of hepatocellular carcinoma using habitat analysis: comparison and combination of deep learning[J]. Acad Radiol, 2024, 31(6): 2346-2355. DOI: 10.1016/j.acra.2023.11.024.
[14]
Breast Cancer Professional Committee of the Chinese Anti-Cancer Association and Breast Tumor Group of the Oncology Branch of the Chinese Medical Association. Guidelines for breast cancer diagnosis and treatment by China Anti-cancer Association (2024 edition)[J]. China J Oncol, 2023, 33(12): 1092-1186. DOI: 10.19401/j.cnki.1007-3639.2023.12.004.
[15]
OGSTON K N, MILLER I D, PAYNE S, et al. A new histological grading system to assess response of breast cancers to primary chemotherapy: prognostic significance and survival[J]. Breast, 2003, 12(5): 320-327. DOI: 10.1016/s0960-9776(03)00106-1.
[16]
LIU W X, CHENG Y L, LIU Z Y, et al. Preoperative prediction of ki-67 status in breast cancer with multiparametric MRI using transfer learning[J/OL]. Acad Radiol, 2021, 28(2): e44-e53 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/32278690/. DOI: 10.1016/j.acra.2020.02.006.
[17]
ZHANG Y Y, HONG D, MCCLEMENT D, et al. Grad-CAM helps interpret the deep learning models trained to classify multiple sclerosis types using clinical brain magnetic resonance imaging[J/OL]. J Neurosci Methods, 2021, 353: 109098 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/33582174/. DOI: 10.1016/j.jneumeth.2021.109098.
[18]
LI J Q, QIU Z B, ZHANG C, et al. ITHscore: comprehensive quantification of intra-tumor heterogeneity in NSCLC by multi-scale radiomic features[J]. Eur Radiol, 2023, 33(2): 893-903. DOI: 10.1007/s00330-022-09055-0.
[19]
VON MINCKWITZ G, UNTCH M, BLOHMER J U, et al. Definition and impact of pathologic complete response on prognosis after neoadjuvant chemotherapy in various intrinsic breast cancer subtypes[J]. J Clin Oncol, 2012, 30(15): 1796-1804. DOI: 10.1200/JCO.2011.38.8595.
[20]
VON MINCKWITZ G, UNTCH M, NÜESCH E, et al. Impact of treatment characteristics on response of different breast cancer phenotypes: pooled analysis of the German neo-adjuvant chemotherapy trials[J]. Breast Cancer Res Treat, 2011, 125(1): 145-156. DOI: 10.1007/s10549-010-1228-x.
[21]
CORTAZAR P, ZHANG L J, UNTCH M, et al. Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis[J]. Lancet, 2014, 384(9938): 164-172. DOI: 10.1016/S0140-6736(13)62422-8.
[22]
YANG C, LIU H, FENG X, et al. Research hotspots and frontiers of neoadjuvant therapy in triple-negative breast cancer: a bibliometric analysis of publications between 2002 and 2023[J]. Int J Surg, 2024, 110(8): 4976-4992. DOI: 10.1097/JS9.0000000000001586.
[23]
RYSPAYEVA D, SEYHAN A A, MACDONALD W J, et al. Signaling pathway dysregulation in breast cancer[J]. Oncotarget, 2025, 16(1): 168-201. DOI: 10.18632/oncotarget.28701.
[24]
CAI Y B, ZENG S F, TU M T. Relationship between DTI parameters and cell density and pathological grade in patients with breast cancer[J]. Chin J CT MRI, 2021, 19(7): 103-105. DOI: 10.3969/j.issn.1672-5131.2021.07.032.
[25]
HOTTAT N A, BADR D A, LECOMTE S, et al. Value of diffusion-weighted MRI in predicting early response to neoadjuvant chemotherapy of breast cancer: comparison between ROI-ADC and whole-lesion-ADC measurements[J]. Eur Radiol, 2022, 32(6): 4067-4078. DOI: 10.1007/s00330-021-08462-z.
[26]
LI X S, FENG R, WANG D, et al. The comparison of the value of mono-exponential mode and diffusion kurtosis imaging mode in predicting the response to neoadjuvant chemotherapy for locally advanced breast carcinoma using diffusion-weighted imaging[J]. Chin J Radiol, 2019, 53(1): 26-32. DOI: 10.3760/cma.j.issn.1005-1201.2019.01.007.
[27]
KREIPE H, HARBECK N, CHRISTGEN M. Clinical validity and clinical utility of Ki67 in early breast cancer[J/OL]. Ther Adv Med Oncol, 2022, 14: 17588359221122725 [2025-11-10]. https://sage.cnpereading.com/paragraph/article/?doi=10.1177/17588359221122725. DOI: 10.1177/17588359221122725.
[28]
WANG H T, YU T, XU S. Research progress on the application of MRI radiomics in neoadjuvant chemotherapy for breast cancer[J]. J China Clin Med Imag, 2023, 34(12): 892-896. DOI: 10.12117/jccmi.2023.12.012.
[29]
LIU X Y, YANG A R, CAO M T, et al. Research on predicting pathological complete response in breast cancer following neoadjuvant chemotherapy using a multiparametric MRI radiomics model[J/OL]. J Radiat Res Appl Sci, 2025, 18(3): 101769 [2025-11-10]. https://www.sciencedirect.com/science/article/pii/S1687850725004819?via%3Dihub. DOI: 10.1016/j.jrras.2025.101769.
[30]
LITJENS G, KOOI T, BEJNORDI B E, et al. A survey on deep learning in medical image analysis[J/OL]. Med Image Anal, 2017, 42: 60-88 [2025-11-10]. https://www.sciencedirect.com/science/article/pii/S1361841517301135?via%3Dihub. DOI: 10.1016/j.media.2017.07.005.
[31]
QIU J W, MITRA J, GHOSE S, et al. A multichannel CT and radiomics-guided CNN-ViT (RadCT-CNNViT) ensemble network for diagnosis of pulmonary sarcoidosis[J/OL]. Diagnostics, 2024, 14(10): 1049 [2025-11-10]. https://www.mdpi.com/2075-4418/14/10/1049. DOI: 10.3390/diagnostics14101049.
[32]
PENG Y S, CHENG Z L, GONG C, et al. Pretreatment DCE-MRI-based deep learning outperforms radiomics analysis in predicting pathologic complete response to neoadjuvant chemotherapy in breast cancer[J/OL]. Front Oncol, 2022, 12: 846775 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/35359387/. DOI: 10.3389/fonc.2022.846775.
[33]
LI S L, DAI Y M, CHEN J Y, et al. MRI-based habitat imaging in cancer treatment: current technology, applications, and challenges[J/OL]. Cancer Imaging, 2024, 24(1): 107 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/39148139/. DOI: 10.1186/s40644-024-00758-9.
[34]
WU L X, DING N, JI Y D, et al. Habitat analysis in tumor imaging: advancing precision medicine through radiomic subregion segmentation[J/OL]. Cancer Manag Res, 2025, 17: 731-741 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/40190416/. DOI: 10.2147/CMAR.S511796.
[35]
HINOHARA K, WU H J, VIGNEAU S, et al. KDM5 histone demethylase activity links cellular transcriptomic heterogeneity to therapeutic resistance[J/OL]. Cancer Cell, 2018, 34(6): 939-953.e9 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/30472020/. DOI: 10.1016/j.ccell.2018.10.014.
[36]
HAMMERL D, MARTENS J W M, TIMMERMANS M, et al. Spatial immunophenotypes predict response to anti-PD1 treatment and capture distinct paths of T cell evasion in triple negative breast cancer[J/OL]. Nat Commun, 2021, 12(1): 5668 [2025-11-10]. https://pubmed.ncbi.nlm.nih.gov/34580291/. DOI: 10.1038/s41467-021-25962-0.
[37]
MROZ E A, ROCCO J W. MATH, a novel measure of intratumor genetic heterogeneity, is high in poor-outcome classes of head and neck squamous cell carcinoma[J]. Oral Oncol, 2013, 49(3): 211-215. DOI: 10.1016/j.oraloncology.2012.09.007.
[38]
JIMÉNEZ-SÁNCHEZ A, CYBULSKA P, MAGER K L, et al. Unraveling tumor-immune heterogeneity in advanced ovarian cancer uncovers immunogenic effect of chemotherapy[J]. Nat Genet, 2020, 52(6): 582-593. DOI: 10.1038/s41588-020-0630-5.

PREV Correlation between arterial recanalization and the susceptibility vessel sign following endovascular therapy in patients with acute ischemic stroke
NEXT The advantage of gadoxetate disodium-enhanced MRI in differentiating focal chronic inflammatory lesions of the liver from solitary hepatic metastases
  


LINK


Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn