CorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation
Abstract
Breast core needle biopsy (CNB) is central to breast cancer diagnosis yet remains challenging because limited tissue sampling, lesion heterogeneity, and subtle morphologic overlap can obscure subtype distinctions. We developed CorePath, a breast-specialized multimodal pathology foundation model fine-tuned from PRISM using 7901 paired CNB whole-slide images and diagnostic reports from two centers. Evaluated across six CNB cohorts and two public breast pathology benchmarks without task-specific retraining, CorePath consistently outperformed PRISM across cancer detection, invasion assessment, and histological subtyping. It achieved weighted area under the receiver operating characteristic curves (AUCs) of 0.9526-0.9735 for five-class CNB histological subtyping across private centers. On public benchmarks, CorePath outperformed leading pathology foundation models, achieving the highest weighted AUCs of 0.7780 for BCNB invasive carcinoma subtyping, 0.8178 for BRACS lesion stratification, and 0.8252 for BRACS fine-grained classification. In report generation, CorePath reduced the overall non-breast hallucinations from 30.1% to 2.8%, demonstrating improved domain fidelity after breast-specific adaptation. CorePath-CRG further combined conformal filtering of subtype and binary cancer status predictions with Learn-Then-Test-based threshold calibration to support selective narrative release, diagnostic fallback, and deferral. CorePath-CRG achieved zero non-breast hallucinations among released outputs and showed the strongest overall performance in pathologist-validated LLM-based Evaluation Scores and quantitative report-generation metrics across most centers. These results demonstrate that domain-specialized foundation models with statistical risk control offer a promising approach for accurate breast CNB diagnosis and reliable report generation.
Related
- Data generation for the logistic regression model under controlled conditions
- When surjectivity demands pathology
- Pathology-Free Real-Space Renormalization Group Theory on an Inverse Limit Space
- What masking geometry works best for EEG foundation models?
- Limits and promises of earth observation foundation models in predicting multi-trophic soil biodiversity
BibTeX
@misc{yin2026corepath,
title = {{CorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation}},
author = {Ting Yin and Danning Li and Chen Shu and Xiaoxia Yao and Boyu Fu and Yujing Chang and Tianyu Shi and Mengna Feng and Jie Chen and Jing Fu and Xiuli Xiao and Tianlin Li and Mumin Shao and Jiaxin Bi and Wenchuan Zhang and Xiaoyan Wu and Xiao Han and Zhang Zhang and Yuhao Yi and Hong Bu},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2608.03079},
eprint = {2608.03079},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.03079},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - CorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation
AU - Ting Yin
AU - Danning Li
AU - Chen Shu
AU - Xiaoxia Yao
AU - Boyu Fu
AU - Yujing Chang
AU - Tianyu Shi
AU - Mengna Feng
AU - Jie Chen
AU - Jing Fu
AU - Xiuli Xiao
AU - Tianlin Li
AU - Mumin Shao
AU - Jiaxin Bi
AU - Wenchuan Zhang
AU - Xiaoyan Wu
AU - Xiao Han
AU - Zhang Zhang
AU - Yuhao Yi
AU - Hong Bu
PY - 2026
DA - 2026/09/22
PB - arXiv
DO - 10.48550/arxiv.2608.03079
UR - https://arxiv.org/abs/2608.03079
AB - Breast core needle biopsy (CNB) is central to breast cancer diagnosis yet remains challenging because limited tissue sampling, lesion heterogeneity, and subtle morphologic overlap can obscure subtype distinctions. We developed CorePath, a breast-specialized multimodal pathology foundation model fine-tuned from PRISM using 7901 paired CNB whole-slide images and diagnostic reports from two centers. Evaluated across six CNB cohorts and two public breast pathology benchmarks without task-specific retraining, CorePath consistently outperformed PRISM across cancer detection, invasion assessment, and histological subtyping. It achieved weighted area under the receiver operating characteristic curves (AUCs) of 0.9526-0.9735 for five-class CNB histological subtyping across private centers. On public benchmarks, CorePath outperformed leading pathology foundation models, achieving the highest weighted AUCs of 0.7780 for BCNB invasive carcinoma subtyping, 0.8178 for BRACS lesion stratification, and 0.8252 for BRACS fine-grained classification. In report generation, CorePath reduced the overall non-breast hallucinations from 30.1% to 2.8%, demonstrating improved domain fidelity after breast-specific adaptation. CorePath-CRG further combined conformal filtering of subtype and binary cancer status predictions with Learn-Then-Test-based threshold calibration to support selective narrative release, diagnostic fallback, and deferral. CorePath-CRG achieved zero non-breast hallucinations among released outputs and showed the strongest overall performance in pathologist-validated LLM-based Evaluation Scores and quantitative report-generation metrics across most centers. These results demonstrate that domain-specialized foundation models with statistical risk control offer a promising approach for accurate breast CNB diagnosis and reliable report generation.
ER -