Quantifying intratumor heterogeneity through MRI-based habitat analysis to distinguish HER2-low and HER2-zero triple-negative breast cancer

Background To evaluate whether magnetic resonance imaging (MRI)-derived intratumoral heterogeneity (ITH), quantified using habitat analysis, can distinguish human epidermal growth factor receptor 2 (HER2)-low from HER2-zero triple-negative breast cancer (TNBC), and to develop a combined model incorporating ITH, MRI features, and clinicopathological characteristics. Methods This retrospective study included patients with surgically confirmed TNBC from two medical centers between April 2019 and September 2024. Center 1 provided the training cohort and center 2 provided the external validation cohort. Preoperative dynamic contrast-enhanced MRI (DCE-MRI) and clinicopathological data were collected. Habitat maps were generated by K-means clustering of peak-enhancement-phase images, and an ITH score was calculated from the spatial distribution of habitat regions. Logistic regression was used to develop an ITH-score model, a clinical-imaging model, and a combined model. Discrimination, calibration, and potential net benefit were evaluated. Results The study included 134 patients: 96 in the training cohort and 38 in the external validation cohort. The combined model achieved an area under the receiver operating characteristic curve (AUC) of 0.827 (95% confidence interval [CI]: 0.746–0.907) in the training cohort and 0.800 (95% CI: 0.660–0.940) in the validation cohort. Corresponding AUCs were 0.786 (95% CI: 0.695–0.876) and 0.769 (95% CI: 0.612–0.927) for the ITH-score model, and 0.712 (95% CI: 0.613–0.812) and 0.747 (95% CI: 0.595–0.900) for the clinical-imaging model. The combined model showed a numerically higher validation AUC, but the between-model differences were not statistically significant. Decision curve analysis suggested potential net benefit within the evaluated threshold ranges. Conclusions The combined model showed promising discrimination between HER2-low and HER2-zero TNBC. The small external validation cohort and wide confidence interval warrant cautious interpretation. Larger prospective studies are needed before the model can inform clinical decision-making.

Authors

Institutions

Publication Details

Journal
Frontiers in Oncology
Published
2026-11-18
DOI
https://doi.org/10.3389/fonc.2026.1978489
Primary Topic
MRI in cancer diagnosis
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Quantifying intratumor heterogeneity through MRI-based habitat analysis to distinguish HER2-low and HER2-zero triple-negative breast cancer

Ao Wu, Jiayu Nan, Wei Zeng, Jiale Hu et al.
7 links
Frontiers in Oncology
MRI in cancer diagnosis
article

Quantifying intratumor heterogeneity through MRI-based habitat analysis to distinguish HER2-low and HER2-zero triple-negative breast cancer

Ao Wu, Jiayu Nan, Wei Zeng, Jiale Hu, Yingyi Luo, Lan Liu, Jiahui Jiang, Hao Li
article en

Abstract

Background To evaluate whether magnetic resonance imaging (MRI)-derived intratumoral heterogeneity (ITH), quantified using habitat analysis, can distinguish human epidermal growth factor receptor 2 (HER2)-low from HER2-zero triple-negative breast cancer (TNBC), and to develop a combined model incorporating ITH, MRI features, and clinicopathological characteristics. Methods This retrospective study included patients with surgically confirmed TNBC from two medical centers between April 2019 and September 2024. Center 1 provided the training cohort and center 2 provided the external validation cohort. Preoperative dynamic contrast-enhanced MRI (DCE-MRI) and clinicopathological data were collected. Habitat maps were generated by K-means clustering of peak-enhancement-phase images, and an ITH score was calculated from the spatial distribution of habitat regions. Logistic regression was used to develop an ITH-score model, a clinical-imaging model, and a combined model. Discrimination, calibration, and potential net benefit were evaluated. Results The study included 134 patients: 96 in the training cohort and 38 in the external validation cohort. The combined model achieved an area under the receiver operating characteristic curve (AUC) of 0.827 (95% confidence interval [CI]: 0.746–0.907) in the training cohort and 0.800 (95% CI: 0.660–0.940) in the validation cohort. Corresponding AUCs were 0.786 (95% CI: 0.695–0.876) and 0.769 (95% CI: 0.612–0.927) for the ITH-score model, and 0.712 (95% CI: 0.613–0.812) and 0.747 (95% CI: 0.595–0.900) for the clinical-imaging model. The combined model showed a numerically higher validation AUC, but the between-model differences were not statistically significant. Decision curve analysis suggested potential net benefit within the evaluated threshold ranges. Conclusions The combined model showed promising discrimination between HER2-low and HER2-zero TNBC. The small external validation cohort and wide confidence interval warrant cautious interpretation. Larger prospective studies are needed before the model can inform clinical decision-making.

Frontiers in OncologyVol. 16
Jiangxi Provincial Cancer Hospital (CN), Nanchang Medical College (CN)
Good health and well-being
MRI in cancer diagnosis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon