Share:
Share this content in WeChat
X
• Clinical Article •
Predicting tumor shrinkage patterns after neoadjuvant therapy in breast cancer using a stacking model integrating MRI habitat analysis and intratumoral microbiota
ZHOU Xiaoping  YANG Wei  ZHANG Ningmei  ZHANG Chaolin  YIN Qingyun  LIU Zhaodi 

DOI:10.12015/issn.1674-8034.2026.11.010.


[Abstract] Objective To investigate the value of a multimodal stacking machine-learning model integrating MRI tumor habitat, peritumoral radiomics, intratumoral microbiota, and clinicopathological features for predicting the tumor shrinkage pattern (TSP) after neoadjuvant therapy (NAT) in breast cancer.Materials and Methods This multicenter ambispective cohort study included 285 patients with breast cancer, who were divided into a training set (n = 178), an internal validation set (n = 77), and an external validation set (n = 30). Using postoperative pathology as the reference standard, TSP was classified as concentric or non-concentric. Tumor radiomic features, including inter-subregion interaction features [volume ratio (VR); spatial dispersion (SD); transition-zone heterogeneity (TZH)], were extracted after habitat subregion partitioning. Together with peritumoral radiomics, clinicopathological features, and four derived microbial features, they were used to construct four base learners, which were integrated by stacking with logistic regression as the meta-learner. The incremental value of the microbial module was assessed using the DeLong test, net reclassification improvement (NRI), and integrated discrimination improvement (IDI), and habitat-microbiota associations were analyzed by Spearman correlation.Results The meta-model achieved AUCs of 0.872 in the training set, 0.846 in the internal validation set, and 0.839 in the external validation set, outperforming all individual base learners. After incorporation of the microbial module, the AUC increased by 0.061 (P < 0.001), 0.054 (P = 0.004), and 0.058 (P = 0.003) in the training, internal validation, and external validation sets, respectively. A higher normalized bacterial burden (NBB) was independently associated with concentric shrinkage (OR = 2.68, P < 0.001), whereas an imbalanced intratumoral bacterial distribution accompanied by endotoxin accumulation was associated with non-concentric shrinkage. The positive predictive value of concentric shrinkage for pathological complete response (pCR) remained stable at 76.5% to 78.4%.Conclusions This model enables noninvasive prediction of the NAT shrinkage pattern; the intratumoral microbiota provides an independent and reproducible incremental value across populations, offering a basis for individualized breast-conserving decision-making.
[Keywords] breast cancer;neoadjuvant therapy;tumor shrinkage pattern;prediction;habitat analysis;intratumoral microbiota;machine learning;magnetic resonance imaging

ZHOU Xiaoping1   YANG Wei2*   ZHANG Ningmei3   ZHANG Chaolin4   YIN Qingyun5   LIU Zhaodi6  

1 The First Clinical Medical College, Ningxia Medical University, Yinchuan 750004, China

2 Department of Radiology, General Hospital of Ningxia Medical University, Yinchuan 750004, China

3 Department of Pathology, General Hospital of Ningxia Medical University, Yinchuan 750004, China

4 Department of Surgical Oncology, General Hospital of Ningxia Medical University, Yinchuan 750004, China

5 Department of Medical Oncology, General Hospital of Ningxia Medical University, Yinchuan 750004, China

6 Department of Radiology, Shizuishan First People's Hospital, Shizuishan 753000, China

Corresponding author: YANG W, E-mail: yangwei_0521@163.com

Conflicts of interest   None.

Received  2026-07-17
Accepted  2026-09-24
DOI: 10.12015/issn.1674-8034.2026.11.010
DOI:10.12015/issn.1674-8034.2026.11.010.

[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]
ZHANG Q, CHEN J M, WU L L, et al. Value of DCE-MRI-based tumor heterogeneity quantification and deep learning in predicting neoadjuvant chemotherapy response in breast cancer[J]. Chin J Magn Reson Imaging, 2026, 17(1): 42-50. DOI: 10.12015/issn.1674-8034.2026.01.007.
[3]
FUKADA I, ARAKI K, KOBAYASHI K, et al. Pattern of tumor shrinkage during neoadjuvant chemotherapy is associated with prognosis in low-grade luminal early breast cancer[J]. Radiology, 2018, 286(1): 49-57. DOI: 10.1148/radiol.2017161548.
[4]
HUANG Y H, SONG X Y, CHEN Y L, et al. Intratumoral microbiome-related MRI model for predicting breast cancer shrinkage pattern following neoadjuvant therapy[J/OL]. Radiology, 2025, 316(2): e243545 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/40892452/. DOI: 10.1148/radiol.243545.
[5]
WANG J B, LIU Y C, LIU T H, et al. Tumor shrinkage patterns and optimal timing of response assessment during neoadjuvant therapy for breast cancer: a study based on multiparametric MRI[J/OL]. Cancer Imaging, 2025, 26(1): 2 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/41345717/. DOI: 10.1186/s40644-025-00971-0.
[6]
The Society of Breast Cancer China AntiCancer Association;Breast Oncology 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 Oncol, 2023, 33(12): 1092-1187. DOI: 10.19401/j.cnki.1007-3639.2023.12.004.
[7]
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.
[8]
LO GULLO R, SUROVITSKY M, KIM D, et al. Abbreviated MRI for the evaluation of treatment response in patients undergoing neoadjuvant chemotherapy for breast cancer[J/OL]. Radiology, 2026, 319: e252413 [2026-07-16]. https://pubs.rsna.org/doi/10.1148/radiol.252413. DOI: 10.1148/radiol.252413.
[9]
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 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/40232145/. DOI: 10.1148/radiol.241805.
[10]
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 [2026-07-16]. https://link.springer.com/article/10.1186/s40644-024-00758-9. DOI: 10.1186/s40644-024-00758-9.
[11]
ZHU H J, ZHANG B Y, PENG D N, et al. MRI-based habitat analysis for predicting pathologic complete response to neoadjuvant chemotherapy in breast cancer: a multicenter study[J]. Magnetic Resonance Imaging, 2026, 63(6): 1769-1779. DOI: 10.1002/jmri.70285.
[12]
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 Imaging, 2023, 34(12): 892-896. DOI: 10.12117/jccmi.2023.12.012.
[13]
SEPICH-POORE G D, ZITVOGEL L, STRAUSSMAN R, et al. The microbiome and human cancer[J/OL]. Science, 2021, 371(6536): eabc4552 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/33766858/. DOI: 10.1126/science.abc4552.
[14]
NEJMAN D, LIVYATAN I, FUKS G, et al. The human tumor microbiome is composed of tumor type-specific intracellular bacteria[J]. Science, 2020, 368(6494): 973-980. DOI: 10.1126/science.aay9189.
[15]
FU A K, YAO B Q, DONG T T, et al. Tumor-resident intracellular microbiota promotes metastatic colonization in breast cancer[J/OL]. Cell, 2022, 185(8): 1356-1372.e26 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/35395179/. DOI: 10.1016/j.cell.2022.02.027.
[16]
RYSPAYEVA D, SEYHAN A A, MACDONALD W J, et al. Signaling pathway dysregulation in breast cancer[J/OL]. Oncotarget, 2025, 16: 168-201 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/40080721/. DOI: 10.18632/oncotarget.28701.
[17]
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 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/40190416/. DOI: 10.2147/CMAR.S511796.
[18]
WANG L J, WANG Y C, YANG L, et al. Using baseline MRI radiomics to predict the tumor shrinkage patterns in HR-Positive, HER2-Negative Breast Cancer[J/OL]. Front Oncol, 2025, 15: 1539644 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/40809032/. DOI: 10.3389/fonc.2025.1539644.
[19]
EISENHAUER E A, THERASSE P, BOGAERTS J, et al. New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1)[J]. Eur J Cancer, 2009, 45(2): 228-247. DOI: 10.1016/j.ejca.2008.10.026.
[20]
YAU C, OSDOIT M, VAN DER NOORDAA M, et al. Residual cancer burden after neoadjuvant chemotherapy and long-term survival outcomes in breast cancer: a multicentre pooled analysis of 5161 patients[J]. Lancet Oncol, 2022, 23(1): 149-160. DOI: 10.1016/S1470-2045(21)00589-1.
[21]
LANDIS J R, KOCH G G. The measurement of observer agreement for categorical data[J/OL]. Biometrics, 1977, 33(1): 159 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/843571/. DOI: 10.2307/2529310.
[22]
ZWANENBURG A, VALLIÈRES M, ABDALAH M A, et al. The image biomarker standardization initiative: standardized quantitative radiomics for high-throughput image-based phenotyping[J]. Radiology, 2020, 295(2): 328-338. DOI: 10.1148/radiol.2020191145.
[23]
DING J W, QIU Z K, PENG H B, et al. Peritumoral habitat radiomics predicts axillary lymph node metastasis in breast cancer[J/OL]. iScience, 2026, 29(4): 115365 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/41993696/. DOI: 10.1016/j.isci.2026.115365.
[24]
DECOUX A, DURON L, HABERT P, et al. Comparative performances of machine learning algorithms in radiomics and impacting factors[J/OL]. Sci Rep, 2023, 13(1): 14069 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/37640728/. DOI: 10.1038/s41598-023-39738-7.
[25]
PIRRACCHIO R, PETERSEN M L, CARONE M, et al. Mortality prediction in intensive care units with the Super ICU Learner Algorithm (SICULA): a population-based study[J]. Lancet Respir Med, 2015, 3(1): 42-52. DOI: 10.1016/S2213-2600(14)70239-5.
[26]
BROWN M M, KUHLE S, SMITH B, et al. Development and validation of super learner models to predict small and large for gestational age in the second generation[J/OL]. Sci Rep, 2025, 15(1): 33212 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/41006777/. DOI: 10.1038/s41598-025-18466-0.
[27]
YANG Y F, WEI L, HU Y, et al. Classification of Parkinson's disease based on multi-modal features and stacking ensemble learning[J/OL]. J Neurosci Methods, 2021, 350: 109019 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/33321153/. DOI: 10.1016/j.jneumeth.2020.109019.
[28]
ZHANG S H, WANG X L, YANG Z, et al. Intra- and peritumoral radiomics model based on early DCE-MRI for preoperative prediction of molecular subtypes in invasive ductal breast carcinoma: a multitask machine learning study[J/OL]. Front Oncol, 2022, 12: 905551 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/35814460/. DOI: 10.3389/fonc.2022.905551.
[29]
ZHAO Q M, ZHANG H, XING W. Integrating peritumoral and intratumoral radiomics with deep learning for preoperative prediction of lymphovascular invasion in invasive breast cancer using DCE-MRI[J/OL]. Technol Cancer Res Treat, 2025, 24: 15330338251374945 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/40899931/. DOI: 10.1177/15330338251374945.
[30]
ZENG Q, LIU L, HE C W, et al. Early and noninvasive prediction of response to neoadjuvant therapy for breast cancer via longitudinal ultrasound and MR deep learning: a multicentre study[J]. Acad Radiol, 2025, 32(3): 1264-1273. DOI: 10.1016/j.acra.2024.10.033.
[31]
LIU J H, LENG X L, YUAN Z F, et al. Predicting breast cancer response to neoadjuvant chemotherapy with ultrasound-based deep learning radiomics models: dual-center study[J/OL]. BMC Cancer, 2025, 25(1): 1737 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/41214584/. DOI: 10.1186/s12885-025-15148-y.
[32]
LUNDBERG S M, ERION G, CHEN H, et al. From local explanations to global understanding with explainable AI for trees[J]. Nat Mach Intell, 2020, 2(1): 56-67. DOI: 10.1038/s42256-019-0138-9.
[33]
VALENZA C, TRAPANI D, LOIBL S, et al. Optimizing postneoadjuvant treatment in patients with early breast cancer achieving pathologic complete response[J]. J Clin Oncol, 2024, 42(20): 2372-2376. DOI: 10.1200/JCO.23.01935.
[34]
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.
[35]
GONG X Q, WANG X X, WANG L, et al. Comparing multi–b-Value diffusion MRI models for predicting pathologic complete response to neoadjuvant chemotherapy in breast cancer[J/OL]. Radiology, 2025, 316: e242969 [2026-07-16]. https://pubs.rsna.org/doi/10.1148/radiol.242969. DOI: 10.1148/radiol.242969.
[36]
GALEANO NIÑO J L, WU H R, LACOURSE K D, et al. Effect of the intratumoral microbiota on spatial and cellular heterogeneity in cancer[J]. Nature, 2022, 611(7937): 810-817. DOI: 10.1038/s41586-022-05435-0.
[37]
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.
[38]
GELLER L T, BARZILY-ROKNI M, DANINO T, et al. Potential role of intratumor bacteria in mediating tumor resistance to the chemotherapeutic drug gemcitabine[J]. Science, 2017, 357(6356): 1156-1160. DOI: 10.1126/science.aah5043.
[39]
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 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/35359387/. DOI: 10.3389/fonc.2022.846775.
[40]
MONGAN J, MOY L, KAHN C E. Checklist for artificial intelligence in medical imaging (CLAIM): a guide for authors and reviewers[J/OL]. Radiol Artif Intell, 2020, 2(2): e200029 [2026-07-16]. https://pubmed.ncbi.nlm.nih.gov/33937821/. DOI: 10.1148/ryai.2020200029.

PREV Application value of APT imaging combined with T2-FLAIR in non-contrast-enhanced diffuse gliomas
NEXT The value of DWI-based virtual magnetic resonance elastography combined with the magnetic resonance severity index for early risk stratification in acute pancreatitis
  



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