(Monday, 7/15/2019) 7:30 AM - 9:30 AM
Room: Stars at Night Ballroom 2-3
Artificial intelligence (AI) has demonstrated promising results in medical imaging applications. In this session, research will be presented demonstrating the use of AI for predicting treatment outcomes in four complimentary areas. The tumor has demonstrated regional variations in genotypes and phenotypes due to the underlying clonal evolution. Since the radiological scan provides a global view of the tumor as well as its surrounding tumor, image-based intratumoral partitioning can reveal intrinsic subregions, illustrating heterogeneity, sometimes referred to as habitats. The imaging patterns of such heterogeneity maps have demonstrated clinical value for determining prognosis and treatment response in breast, lung, as well as head and neck cancer. In hepatobiliary and pancreatic cancers, it is often challenging to determine imaging-based response metrics due to a lack of change in the size of the tumors after cytotoxic therapies. Efforts in qualitative and quantitative metrics of response as well as future directions in the application of artificial intelligence in response prediction for these cancers will be discussed. In breast cancer, both human-engineered radiomics/machine learning methods and deep learning methods will be discussed for use in MRI-based assessment of cancer progression, response to therapy, and risk of recurrence.
Learning Objectives:
1. Describe the use of artificial intelligence to predict treatment outcomes
2. Illustrate examples of predicting response in liver, breast, lung, and head and neck cancer treatment
3. Explain how imaging characteristics can be quantified for characterization of treatment response
Handouts
Keywords
Not Applicable / None Entered.
Taxonomy
Contact Email
