Hong-Yu Zhou
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Hong-Yu Zhou has a total of 287 co-authors. Here are the top co-authors by number of shared publications:
- Yizhou Yu
School of Computing and Data Science, The University of Hong Kong, Hong Kong SAR, China
20 shared publications - Liansheng Wang
Xiamen University
7 shared publications - Shanshan Wang
Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
6 shared publications - Tao Tan
The Netherlands Cancer Institute
6 shared publications - Tianyu Zhang
Department of Radiology, The Netherlands Cancer Institute, Amsterdam, The Netherlands
6 shared publications Hao Yang
Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
5 shared publications
Latest Publications
- Open Access
- 29 October 2026
- Published by Springer Nature in Nature Communications
The accuracy-safety trade-off of large reasoning models in clinical diagnostics
- 29 October 2026
- Published by Springer Nature in Nature Communications
Open AccessThe accuracy-safety trade-off of large reasoning models in clinical diagnostics
- Open Access
- 23 September 2026
- Published by Springer Nature in Research Square
A multi-model study of diagnostic faithfulness in AI-generated histopathology images
Translating pathological knowledge into images offers new ways to support medical education, illustrate rare disease patterns, and expand training data beyond the constraints of privacy regulation and annotation scarcity. Yet whether current generative AI can produce images diagnostically faithful to clinical descriptions remains unknown. We studied seven text-to-image models on an atlas-derived cohort spanning 6 organ systems and 19 tumor subcategories, combining tri-dimensional expert scoring with structured instruction-following verification across 9,612 pathologist ratings. Only the best-performing model approaches diagnostic interpretability; even so, it omits approximately one third of prompted features, with omission rates rising with case complexity. Automated metrics prove inadequate, as their scores do not reflect clinically meaningful quality differences. A model trained exclusively on breast histopathology is outperformed by general-purpose models within its own domain, suggesting that domain-specific training alone is insufficient without the underlying capacity to handle complex diagnostic descriptions. These findings reveal that visual plausibility is not a sufficient proxy for diagnostic utility, and that both generative models and evaluation standards must advance before synthetic pathology images can be responsibly adopted in oncology.
- 23 September 2026
- Published by Springer Nature in Research Square
Open AccessA multi-model study of diagnostic faithfulness in AI-generated histopathology images
Translating pathological knowledge into images offers new ways to support medical education, illustrate rare disease patterns, and expand training data beyond the constraints of privacy regulation and annotation scarcity. Yet whether current generative AI can produce images diagnostically faithful to clinical descriptions remains unknown. We studied seven text-to-image models on an atlas-derived cohort spanning 6 organ systems and 19 tumor subcategories, combining tri-dimensional expert scoring with structured instruction-following verification across 9,612 pathologist ratings. Only the best-performing model approaches diagnostic interpretability; even so, it omits approximately one third of prompted features, with omission rates rising with case complexity. Automated metrics prove inadequate, as their scores do not reflect clinically meaningful quality differences. A model trained exclusively on breast histopathology is outperformed by general-purpose models within its own domain, suggesting that domain-specific training alone is insufficient without the underlying capacity to handle complex diagnostic descriptions. These findings reveal that visual plausibility is not a sufficient proxy for diagnostic utility, and that both generative models and evaluation standards must advance before synthetic pathology images can be responsibly adopted in oncology.
- 25 June 2026
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 3 (7)
ReXGroundingCT: A 3D Chest CT Dataset for Segmentation of Findings from Free-Text Reports
- 25 June 2026
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 3 (7)
ReXGroundingCT: A 3D Chest CT Dataset for Segmentation of Findings from Free-Text Reports
- 23 June 2026
- Published by Springer Nature in Nature Biomedical Engineering
- Vol. 10 (8) , 1526-1535
Large reasoning models as thinking machines for medicine
Conventional artificial intelligence has achieved remarkable feats in identifying associations and predictive patterns, yet its limitations in causal reasoning present a hurdle for complex clinical challenges demanding deep professional expertise. Large reasoning models are creating opportunities to move beyond correlation towards emulating the analytical processes of humans. Applied to the practice of medicine, medical reasoning artificial intelligence (MRAI) envisions systems that can engage directly in patient care, draw on diverse clinical data and decision-support tools, and refine their reasoning by learning from clinician feedback and patient outcomes. Unlike traditional models that operate within fixed parameters, MRAI is expected to redefine clinical artificial intelligence as a thinking partner, enabling a more nuanced understanding of complex medical scenarios. We anticipate that MRAI may act as a collaborative aide, augmenting connection and decisions by managing complex evidence. This paradigm shift is expected to profoundly extend our understanding of medicine, free clinicians for more direct patient care, offer clearer insights and accelerate discovery.
- 23 June 2026
- Published by Springer Nature in Nature Biomedical Engineering
- Vol. 10 (8) , 1526-1535
Large reasoning models as thinking machines for medicine
Conventional artificial intelligence has achieved remarkable feats in identifying associations and predictive patterns, yet its limitations in causal reasoning present a hurdle for complex clinical challenges demanding deep professional expertise. Large reasoning models are creating opportunities to move beyond correlation towards emulating the analytical processes of humans. Applied to the practice of medicine, medical reasoning artificial intelligence (MRAI) envisions systems that can engage directly in patient care, draw on diverse clinical data and decision-support tools, and refine their reasoning by learning from clinician feedback and patient outcomes. Unlike traditional models that operate within fixed parameters, MRAI is expected to redefine clinical artificial intelligence as a thinking partner, enabling a more nuanced understanding of complex medical scenarios. We anticipate that MRAI may act as a collaborative aide, augmenting connection and decisions by managing complex evidence. This paradigm shift is expected to profoundly extend our understanding of medicine, free clinicians for more direct patient care, offer clearer insights and accelerate discovery.
- Top Cited Papers
- 26 March 2026
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 3 (4)
MedVersa: A Generalist Foundation Model for Diverse Medical Imaging Tasks
- 26 March 2026
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 3 (4)
Top Cited PapersMedVersa: A Generalist Foundation Model for Diverse Medical Imaging Tasks
- 16 February 2026
- Published by Elsevier in Medical Image Analysis
- Vol. 110, 103990
Incorporating global-local tissue changes to predict future breast cancer from longitudinal screening mammograms
- 16 February 2026
- Published by Elsevier in Medical Image Analysis
- Vol. 110, 103990
Incorporating global-local tissue changes to predict future breast cancer from longitudinal screening mammograms
- Top Cited Papers
- 6 January 2026
- Published by Springer Nature in Nature Biomedical Engineering
- Vol. 10 (8) , 1595-1609
A multimodal vision–language model for generalizable annotation-free pathology localization
Existing deep learning models for defining pathology from clinical imaging data rely on expert annotations and lack generalization capabilities in open clinical environments. Here we present a generalizable vision-language model for Annotation-Free pathology Localization (AFLoc). The core strength of AFLoc is extensive multilevel semantic structure-based contrastive learning, which comprehensively aligns multigranularity medical concepts with abundant image features to adapt to the diverse expressions of pathologies without the reliance on expert image annotations. We conducted primary experiments on a dataset of 220,000 pairs of image-report chest X-ray images and performed validation across 8 external datasets encompassing 34 types of chest pathology. The results demonstrate that AFLoc outperforms state-of-the-art methods in both annotation-free localization and classification tasks. In addition, we assessed the generalizability of AFLoc on other modalities, including histopathology and retinal fundus images. We show that AFLoc exhibits robust generalization capabilities, even surpassing human benchmarks in localizing five different types of pathological image. These results highlight the potential of AFLoc in reducing annotation requirements and its applicability in complex clinical environments.
- 6 January 2026
- Published by Springer Nature in Nature Biomedical Engineering
- Vol. 10 (8) , 1595-1609
Top Cited PapersA multimodal vision–language model for generalizable annotation-free pathology localization
Existing deep learning models for defining pathology from clinical imaging data rely on expert annotations and lack generalization capabilities in open clinical environments. Here we present a generalizable vision-language model for Annotation-Free pathology Localization (AFLoc). The core strength of AFLoc is extensive multilevel semantic structure-based contrastive learning, which comprehensively aligns multigranularity medical concepts with abundant image features to adapt to the diverse expressions of pathologies without the reliance on expert image annotations. We conducted primary experiments on a dataset of 220,000 pairs of image-report chest X-ray images and performed validation across 8 external datasets encompassing 34 types of chest pathology. The results demonstrate that AFLoc outperforms state-of-the-art methods in both annotation-free localization and classification tasks. In addition, we assessed the generalizability of AFLoc on other modalities, including histopathology and retinal fundus images. We show that AFLoc exhibits robust generalization capabilities, even surpassing human benchmarks in localizing five different types of pathological image. These results highlight the potential of AFLoc in reducing annotation requirements and its applicability in complex clinical environments.
- 1 January 2026
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 3 (1)
International Retrospective Observational Study of Continual Learning for AI on Endotracheal Tube Placement from Chest Radiographs
- 1 January 2026
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 3 (1)
International Retrospective Observational Study of Continual Learning for AI on Endotracheal Tube Placement from Chest Radiographs
- 1 January 2026
- Published by Radiological Society of North America (RSNA) in Radiology: Artificial Intelligence
- Vol. 8 (1) , e240646
Visualizing Radiologic Connections: An Explainable Coarse-to-Fine Foundation Model with Multiview Mammograms and Associated Reports
The proposed explainable pretraining framework for multiview mammograms and associated reports demonstrated high label efficiency in malignancy recognition, segmentation, and localization across four external public datasets.
- 1 January 2026
- Published by Radiological Society of North America (RSNA) in Radiology: Artificial Intelligence
- Vol. 8 (1) , e240646
Visualizing Radiologic Connections: An Explainable Coarse-to-Fine Foundation Model with Multiview Mammograms and Associated Reports
The proposed explainable pretraining framework for multiview mammograms and associated reports demonstrated high label efficiency in malignancy recognition, segmentation, and localization across four external public datasets.
- Open Access
- 8 December 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 10934
Mammo-AGE: deep learning estimation of breast age from mammograms
Biological age is an important indicator of organ functions and health. Although mammograms are widely used in breast cancer screening, the potential of mammogram-based biological age predictors remains underexplored. Here, we propose a deep learning model to estimate the biological age of the breast using healthy mammograms. The model is developed on three large datasets and externally validated on two additional datasets, encompassing 95,826 mammograms from 44,497 women aged 18 to 98 years. It demonstrates accurate age estimation (mean absolute error: 4.2 − 6.1 years) with strong correlation to chronological age. Predicted breast age stratifies breast cancer risk similarly to chronological age. Occlusion analysis, employed for model interpretation, reveals the aging-related pattern of the breast. The breast age gap (the difference between system-bias-corrected breast age and chronological age) may reflect breast health status. Breast cancer patients show higher breast age gaps than the healthy population. In two longitudinal datasets, larger breast age gaps are associated with increased future breast cancer risk, with hazard ratios of 1.013 − 1.022. Furthermore, we finetune the model specifically for downstream breast cancer diagnosis and risk prediction. Our approach outperforms other comparative methods, showing its potential for supporting both early detection and personalized screening strategies.
- 8 December 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 10934
Open AccessMammo-AGE: deep learning estimation of breast age from mammograms
Biological age is an important indicator of organ functions and health. Although mammograms are widely used in breast cancer screening, the potential of mammogram-based biological age predictors remains underexplored. Here, we propose a deep learning model to estimate the biological age of the breast using healthy mammograms. The model is developed on three large datasets and externally validated on two additional datasets, encompassing 95,826 mammograms from 44,497 women aged 18 to 98 years. It demonstrates accurate age estimation (mean absolute error: 4.2 − 6.1 years) with strong correlation to chronological age. Predicted breast age stratifies breast cancer risk similarly to chronological age. Occlusion analysis, employed for model interpretation, reveals the aging-related pattern of the breast. The breast age gap (the difference between system-bias-corrected breast age and chronological age) may reflect breast health status. Breast cancer patients show higher breast age gaps than the healthy population. In two longitudinal datasets, larger breast age gaps are associated with increased future breast cancer risk, with hazard ratios of 1.013 − 1.022. Furthermore, we finetune the model specifically for downstream breast cancer diagnosis and risk prediction. Our approach outperforms other comparative methods, showing its potential for supporting both early detection and personalized screening strategies.