본문으로 건너뛰기
← 뒤로

MRI-Based Habitat Analysis for Predicting Pathologic Complete Response to Neoadjuvant Chemotherapy in Breast Cancer: A Multicenter Study.

Journal of magnetic resonance imaging : JMRI 2026

Zhu H, Zhang B, Peng D, Song S, Huang L, Zhang Y, Wu X

📝 환자 설명용 한 줄

[BACKGROUND] Habitat imaging has been widely used to assess tumor treatment response; however, the role of MRI-based habitat analysis in identifying pathological complete response (pCR) after neoadjuv

🔬 핵심 임상 통계 (초록에서 자동 추출 — 원문 검증 권장)
  • 표본수 (n) 249
  • p-value p < 0.05

이 논문을 인용하기

BibTeX ↓ RIS ↓
APA Zhu H, Zhang B, et al. (2026). MRI-Based Habitat Analysis for Predicting Pathologic Complete Response to Neoadjuvant Chemotherapy in Breast Cancer: A Multicenter Study.. Journal of magnetic resonance imaging : JMRI. https://doi.org/10.1002/jmri.70285
MLA Zhu H, et al.. "MRI-Based Habitat Analysis for Predicting Pathologic Complete Response to Neoadjuvant Chemotherapy in Breast Cancer: A Multicenter Study.." Journal of magnetic resonance imaging : JMRI, 2026.
PMID 41965129
DOI 10.1002/jmri.70285

Abstract

[BACKGROUND] Habitat imaging has been widely used to assess tumor treatment response; however, the role of MRI-based habitat analysis in identifying pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) in breast cancer remains an unresolved issue.

[OBJECTIVES] To evaluate the utility of dynamic contrast-enhanced MRI (DCE-MRI)-based habitat imaging in identifying pCR after NAC in breast cancer patients.

[STUDY TYPE] Retrospective.

[FIELD STRENGTH/SEQUENCE] 1.5 T or 3.0 T, DCE-MRI (Gradient echo).

[SUBJECTS] Three hundred and sixty-three women with biopsy-confirmed breast cancer from Center A (n = 249, training set) and Center B and Center C (n = 114, external validation set).

[ASSESSMENT] DCE-MRI peak-enhancement images were used to generate habitat maps via supervoxel segmentation and K-means clustering. Two intratumoral heterogeneity (ITH) metrics (Volume Entropy and Intensity Entropy) were extracted to quantify the structural and signal complexity of tumors. Three discriminative models were developed: a clinical model based on clinicopathologic variables, an ITH model incorporating Volume Entropy and Intensity Entropy, and an integrated nomogram combining both feature sets.

[STATISTICAL TESTS] Student's t test, Wilcoxon U test, χ, Fisher exact test, and receiver operating characteristic curve analysis. Significance was set at p < 0.05.

[RESULTS] Volume Entropy and Intensity Entropy were significantly lower in pCR versus non-pCR groups. HR status, HER2 status, and both ITH features were independent indicators of pCR. The nomogram showed superior performance (AUC = 0.849 in the training set and 0.825 in the validation set), outperforming the clinical model (DeLong test). Subgroup analysis across four molecular subtypes showed AUCs ranging from 0.762 to 0.890. An interactive online tool was developed for clinical application.

[DATA CONCLUSION] MRI-based habitat analysis offers a simple, interpretable, and clinically applicable approach for noninvasive identification of pCR to NAC in breast cancer.

[TECHNICAL EFFICACY] Stage 3.

같은 제1저자의 인용 많은 논문 (5)