What we do
As modern medical imaging systems become increasingly more quantitative, studies suggest that Artificial Intelligence (A.I.) will lead to personalized imaging and precision medicine. The Woo Center’s Division of A.I. in Medical Imaging conducts operational research to solve highly complex clinical problems using machine learning, computer vision, and other A.I. techniques. We take a multi-disciplinary approach to close the gap between A.I., imaging science, and clinical domain knowledge. The approaches we use (i.e., computer vison, deep learning, quantitative image processing, radiation physics, digital pathology, etc.) are brought together in a unique fashion to improve the prevention, diagnosis, treatment, and prognosis of human disease.
Radiomics and Computational Biomarker Discovery. Radiomics is the process of transforming radiological images into mineable data, from which computational biomarkers can be developed. High-throughput radiomics datasets can be mapped to biological phenomena otherwise unappreciated by the human eye. Our team has identified a number of radiomic biomarkers for various clinical applications, including: (a) virtual tumor biopsies, (b) prediction of disease response to modern therapies, and (c) characterization of physiological processes, such as pulmonary function and cell metabolism. This work includes collaborations with the Department of Radiation Oncology and the Department of Radiology.
Pathomics and Precision Renal Medicine. In digital pathology, the advent of whole slide image acquisition has resulted in new opportunities for A.I. applications. Of particular interest, the characterization of donor kidney biopsies is essential to renal allograft transplantation. Our team is implementing state-of-the-art techniques in deep learning and computer vision to identify and characterize histological primitives on pre-transplantation kidney biopsies. This work includes collaborations with the Department of Pathology and the Department of Medicine’s Division of Nephrology.
Precision Prevention and Surgical Risk Stratification. Surgical treatment options for patients with thoracic cancers is based on many competing factors, including the increasing use of lung cancer screening images. Our team is developing AI-guided precision prevention techniques to (a) provide clinical decision support and surgical quality control, (b) quantify the risk of locoregional recurrence, and (c) predict the utility of adjuvant systemic therapies. This work includes collaborations with the Department of Surgery and the Department of Radiology.
Automation and Optimization of Radiation Therapy. The goal of radiation therapy is to deliver a high dose of ionizing radiation to cancer cells, while sparing the surrounding healthy tissue. This process is inherently based on imaging, and is essentially a complex optimization problem. Our team is developing knowledge-based techniques using multi-omics data (i.e., radiomics, genomics, pathomics, dosiomics, etc.) to automate different components of clinical radiotherapy workflows, including: (a) feature extraction and organ segmentation, (b) optimization of radiation beams and intensities, (c) improvement of on-board image quality with deep learning, and (d) verification and assessment of treatment delivery. This work is a collaboration with the Department of Radiation Oncology.
How to participate
The Woo Center works to improve health care through educational experiences, research projects, and entrepreneurial opportunities for Duke faculty and students in collaboration with clinical and industry partners worldwide.
The following are ways in which you can participate in this work:
Students interested in gaining valuable experience in this Division are encouraged to forward your CV directly to the Lead Data Scientist. If there is an opportunity that matches your interests, you will be contacted for an interview.
Clinical staff within the Duke Health system with a potential project should complete the Contact form below, and include your name, department, contact details, and a description of your proposed project. The Woo Center is unfortunately unable to accept all proposals due to limited resources, but will select projects based on their greatest impact on health care and health care processes.
Meet the team
Fang-Fang Yin | firstname.lastname@example.org
Dr. Fang-Fang Yin is a Professor of Radiation Oncology and the Director of Radiation Physics in the Department of Radiation Oncology. Dr. Yin is an established investigator with over 35 years of experience in quantitative medical image analysis, image-guided treatment assessment, image reconstruction, feature extraction, and computer aided diagnosis. Dr. Yin is a member of the Duke Cancer Institute, and has published over 280 peer-reviewed publications.
Kyle Lafata | email@example.com
Dr. Kyle Lafata is an Assistant Professor of Radiology and Radiation Oncology, as well as a Fellow in the Big Data Scientist Training Enhancement Program at the U.S. Department of Veterans Affairs. Dr. Lafata is a physicist, data scientist, and medical imaging researcher working on quantitative imaging, feature engineering, and computational biomarker development. His primary research interests are in applications of Artificial Intelligence to biomedical imaging to improve the prevention, diagnosis, and prognosis of human disease.
Ouwen Huang | firstname.lastname@example.org
Ouwen is currently an M.D, Ph.D. Candidate at Duke with a passion for automation, machine learning, and computer vision. He works on Gradient Health which partners with international PACS to bridge the gap between academic imaging algorithms and the clinic. Prior, he was Co-founder and CTO of a YCombinator image analytics company which was acquired by Syngenta for $5MM. In his spare time, he enjoys contributing to open source repos such as TensorFlow.