Related Experiment Video
Updated: Oct 10, 2026
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
MR Cytometry of Microstructural Changes in Breast Cancer: Association With Treatment Response and Prognosis
Xiaoxia Wang1, Yao Huang1, Ruicheng Ba2
1Department of Radiology, Chongqing University Cancer Hospital, Chongqing Key Laboratory for Intelligent Oncology in Breast Cancer (iCQBC), Chongqing, China.
Background:
Although MR cytometry can probe microstructural features, the longitudinal trajectories of these changes during neoadjuvant chemotherapy (NAC) and their association with treatment response and prognosis remain poorly defined.
Purpose:
To characterize longitudinal changes in NAC-associated cellular microstructural properties and evaluate microstructural parameters for predicting pathologic complete response (pCR) and disease-free survival (DFS) in breast cancer.
Study Type:
Prospective.
Population:
194 patients with invasive breast cancer underwent 713 MRI examinations.
Field Strength/Sequence:
3.0 T, oscillating gradient spin-echo (OGSE) and pulsed gradient spin-echo (PGSE) sequences.
Assessment:
MR cytometry was acquired at four time-points: pretreatment, early, mid-, and late treatment. Microstructural parameters were estimated using the IMPULSED (imaging microstructural parameters using limited spectrally edited diffusion) model. Microstructural parameters were compared with histopathologic measurements.
Statistical Tests:
Generalized estimating equations, logistic regression, bootstrap resampling, the DeLong test with Bonferroni correction, Cox proportional hazards regression, Kaplan-Meier analysis with the log-rank test, the C-index, and the Pearson correlation coefficient were performed. p < 0.05 was significant.
Results:
Four logistic regression models were developed for predicting pCR, based on molecular subtype alone or combined with diameter at Time 1, ADC50Hz at Time 2, or cellularity at Time 2. The clinicopathological-cellularity model achieved the best performance in predicting pCR (AUC = 0.89). For DFS, a Cox model incorporating ER status, HER2 status, cT stage, cN stage, and extracellular diffusivity at Time 1 yielded a C-index of 0.81; patients stratified by the median risk score into low- and high-risk groups differed significantly in DFS. Extracellular diffusion was positively correlated with pathologic stroma fraction (r = 0.56).
Data Conclusion:
MR cytometry demonstrated longitudinal microstructural alterations during NAC and shows potential for predicting pCR and stratifying patients by DFS risk in breast cancer patients.
Technical Efficacy:
Stage 2.