The joint analysis of imaging data and genetic data for early tumor detection, prevention, diagnosis and treatment
University of North Carolina at Chapel Hill
Professor Hongtu Zhu
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
We intend to explore different imaging and genetic characteristics to work on automatic early tumor detection, benign and malignant tumor discrimination, and prediction of tumor progression. These are critical in human cancer research because in our understanding, existing automatic cancer detection approaches using imaging modalities are rare, while by only recognizing possible physical warning signs of cancer for early diagnosis is not accurate enough. Furthermore, investigation into genetic factor and other risk factors including body mass index, medications, smoking and dietary factors, etc., will hopefully lead to the prevention and improvement of the cancer tumor symptoms. Our methods directly addresses the aim of UK Biobank for tumor detection, prevention, diagnosis and treatment. We will make use of the wealth of UK Biobank data to provide important new insights into the cancer research. And it will lead to an improved understanding of genetic mechanism in tumor progression and tumor imaging characteristics which has the potential to inspire new and urgently needed approaches to prevention, diagnosis, and treatment of cancer. Initially we implement machine learning techniques to automatically classify the tumor area for all imaging modalities. Imaging characteristics are derived after preprocessment of the raw imaging data to discriminate different tumor types. Second, we use different statistical approaches to select oncogenes and tumor suppressor genes. A genetic-imaging interactive analysis will be conducted for predictions of tumor progression and treatment. We will develop different novel statistical tools, find those to handle the dataset more efficiently, verify efficacy of the newly developed statistical tools using simulations and existing studies, and finally get the clinical conclusions and develop companion software. The full cohort data if applicable.