SUMMARY
Hao Gong, Ph.D., is a clinical imaging researcher whose research vision focuses on developing and applying clinical imaging and imaging analytics using disease-driven, quantitative artificial intelligence (AI) and deep-learning (DL) techniques. Dr. Gong strives to use these advanced techniques to address unmet needs in radiology practice, spanning multiple critical areas of patient care.
His research includes:
- Task-driven quality assessment and clinical protocol optimization.
- Physics-informed deep learning.
- Uncertainty and bias quantification in AI and DL models.
- Image reconstruction and processing in quantitative spectral CT.
- Quantitative image biomarkers.
- Diagnostic image quality enhancement.
Dr. Gong's research has been supported by funding from the National Institute of Biomedical Imaging and Bioengineering, the National Institute of Arthritis and Musculoskeletal and Skin Diseases, the Minnesota Partnership for Biotechnology and Medical Genomics, and other organizations.
Focus areas
- AI-assisted virtual clinical imaging trials. Dr. Gong develops neural network models to generate realistic patient CT images with target radiopathological features and to accurately predict diagnostic performance. He uses these AI models to conduct virtual clinical imaging trials. These models facilitate task-driven optimization of visualization, detection, and assessment of critical pathological features, such as low-contrast lung nodules and liver metastases. Dr. Gong also uses these models to optimize radiation dose and refine clinical protocols across different CT systems and commercial reconstruction and postprocessing algorithms.
- Physics-informed AI-assisted quantitative spectral CT. Dr. Gong develops physics-informed neural network models for image processing and quantitative assessment in spectral CT imaging tasks. These models integrate CT physics with neural network architectures and mathematical regularization to enhance interpretability, robustness, and generalizability. He uses these models to visualize and quantify mass densities and spatial distribution of contrast media and biological materials, reflecting the underlying physiopathologic processes. Dr. Gong also uses these models to derive high-quality supplemental CT image types, such as virtual noncalcium images, virtual noncontrast images, and virtual monoenergetic images. These images are used in a variety of clinical applications, such as bone marrow pathology assessment, stenosis quantification, and metastatic lesion detection.
- Quantitative image biomarkers. Dr. Gong investigates new quantitative image biomarkers that are closely linked to different conditions, such as osteoarthritis and multiple myeloma. He also uses these biomarkers in machine learning models for early detection, staging, and treatment response assessment.
- Uncertainty and bias in AI models. Dr. Gong develops new methods to quantify the uncertainty and bias of AI models used in clinical CT image processing, including noise reduction, to ensure model reliability and safety. Dr. Gong studies methods to leverage the information of uncertainty and bias to advance AI model performance in clinical tasks.
Significance to patient care
Dr. Gong's research aims to make diagnosis and treatment planning more accurate, efficient, and tailored for each patient. His research provides new clinical imaging capabilities and quantitative information to better understand and treat a wide range of medical conditions. Dr. Gong's work helps improve various aspects of care, from the initial steps before collecting raw data to the later stages after clinical interpretation.
Professional highlights
- National Institutes of Health:
- Principal investigator, R01, AI-Assisted Quantitative Photon-Counting-Detector CT Imaging for Cytogenetic Risk Prediction and Treatment Response in Multiple Myeloma, funded by the National Institute of Biomedical Imaging and Bioengineering, 2025-present.
- Principal investigator, R01, Diagnostic Performance Assessment and Dose Optimization Using Patient CT Images: Application to Deep-Learning CT Reconstruction and Denoising Technologies, funded by the National Institute of Biomedical Imaging and Bioengineering, 2024-present.
- Principal investigator, R21, One-Stop-Shop Early Diagnosis of Joint Instability Using AI-Assisted 6D Computed Tomography, funded by the National Institute of Arthritis and Musculoskeletal and Skin Diseases, 2024-present.
- American Association of Physicists in Medicine (AAPM):
- Voting member, Working Group on Generative Artificial Intelligence, 2025-present.
- Voting member, associate editor, Board of Associate Editors, Medical Physics, 2024-present.
- Voting member, Working Group on Journal-Sourced Educational Content, 2023-2024.
- Featured cover paper and Editor's Choice, First author, Medical Physics, 2022.
- Best-in-physics, first author, diagnostic imaging track, Annual Meeting & Exhibition, 2018, 2021.
- Principal investigator, Enhanced Assessment of Bone Marrow Pathology Using a Deep-Learning-Based Virtual Non-Calcium Technique in Multi-Energy Computed Tomography, Translational Product Development Fund, funded by Mayo Clinic, University of Minnesota and the Minnesota Partnership for Biotechnology and Medical Genomics, 2023-2025.
- Invited grant proposal reviewer, KWF Dutch Cancer Society, 2024.
- Invited topic editor, Frontiers in Radiology, 2022-2023.
- Best paper, First author, North American Chinese Medical Physicists Association, 2022.