Computer Aided Diagnosis - Artificial Intelligence

Research Laboratory

Research Projects


Publications

Welcome

The Computer Aided Diagnosis-Artificial Intelligence (CAD-AI) Laboratory in the Department of Radiology was established in 1989 by Heang-Ping Chan, Ph.D. The laboratory conducts research in computer-aided diagnosis using machine learning and AI methods, aiming at providing decision support to clinicians in the various stages of patient care process, such as disease detection, classification of malignant and benign lesions, quantitative analysis of lesion characteristics and their changes, cancer risk assessment, staging, treatment response assessment, prognosis, and recurrence monitoring.

The researchers in the laboratory collaborate with clinicians and scientists within and outside the department/institution on various projects supported by research funding from federal and non-federal agencies. We train postdoctoral fellows and students in diagnostic imaging, medical image analysis, and CAD-AI research. We have been working on decision support tools for various diseases, including breast cancer, lung cancer, bladder and ureter cancer, cardiovascular disease, pulmonary embolism, head and neck cancer, and multiple myeloma. We explore radiomics as a phenotypic expression of other -omics or disease states, such as the association of radiomics and genomics biomarkers for differentiation of breast cancer subtypes, and the association of radiomics and pathomics biomarkers for invasive and indolent lung cancer. Machine learning methods, including deep convolutional neural networks, and data analytics are exploited to build the various CAD-AI tools. We study the impact of CAD-AI on clinicians’ decision making and workflow.

We also investigate methods for improving image quality and thus increasing diagnostic accuracy by clinicians or machine vision. We strive to foster multi-disciplinary collaborative research and work towards the goal of integrating AI-based decision support into clinical practice and improving health care.