Overview
My research is at the intersection of computer vision, machine learning, and medical imaging, with a dual focus on mammography and computed tomography (CT). Together with our industry partner, we developed deep learning algorithms for breast cancer screening with 2D/3D mammography, and that product is now undergoing FDA approval with anticipated rollout to clinics worldwide. We also pioneer the creation of "digital twin" anatomical models from patient imaging data, using these models to forge new paths in CT scan analysis through virtual readers and deep learning techniques. Additionally, we're developing a computer-aided triage system for detecting diseases across multiple organs in body CT scans, leveraging hospital-scale datasets and integrating natural language processing with deep learning for comprehensive disease classification.
Current Duke Appointments & Affiliations
Recent Scholarly Works
Breast area affects the performance of a commercial artificial intelligence algorithm assessment of negative digital breast tomosynthesis exams.
Journal article Eur J Radiol · July 2026 OBJECTIVE: To understand whether cancer-neutral image attributes (breast area and number of slices) impact an AI algorithm assessment of negative digital breast tomosynthesis (DBT) screening exams. METHODS: This retrospective cohort study included women fr ... Full text Link to item CiteAAPM Task Group Report 336: Quality assurance for 3D printing in medical imaging and radiation therapy applications
Journal article Medical Physics · May 1, 2026 Additive manufacturing (AM), commonly referred to as 3D printing, has been broadly used in the field of medical imaging and radiotherapy. In medical imaging applications, patient-specific anatomic models have been used to aid in preoperative planning, pati ... Full text CiteFine Annotation Loss and Top-k Analysis in Interpretable Models for Breast Cancer Prediction
Conference Progress in Biomedical Optics and Imaging Proceedings of SPIE · April 3, 2026 Mammograms provide critical information to radiologists, aiding in the early detection of cancer. We would like to use neural network models to assist radiologists in this challenging and important task, however, these algorithms are “black box” – unable t ... Full text CiteRecent Grants
Comparing an Operation to Monitoring, with or without Endocrine Therapy (COMET), for low-risk DCIS - clinical follow-up
ResearchCo Investigator · Awarded by Department of Defense · 2026 - 2030SCH: Interpretable Machine Learning and Discovery in Medical Images
ResearchCo-Principal Investigator · Awarded by National Institutes of Health · 2025 - 2029Dynamic imaging and tissue biomarker models to delineate indolent from aggressive breast calcifications
ResearchCo Investigator · Awarded by National Cancer Institute · 2022 - 2027View All Grants