NCI Convenes Workshop on Artificial Intelligence (AI) in Radiation Oncology

On April 4-5, 2019, DCTD’s Radiation Research Program convened the “NCI Workshop on Artificial Intelligence in Radiation Oncology” (agenda), which addressed the application of AI to radiation oncology. This recent, cutting-edge technology could help facilitate cancer diagnosis, assess radiation treatment response, and data-mine Big Data for clinical and imaging databases from NCI clinical trials and publicly accessible archives, such as the Cancer Imaging Archive (TCIA). Using AI may also lead to the correlation of radiomics and genomics to gain insight into possible imaging biomarkers to predict patient outcomes and improve future cancer treatments.

Meeting Goals

• Discuss the latest capabilities of AI and Machine Learning and how these new tools and algorithms can address the clinical needs of radiation oncology
• Discuss current issues of AI applications to medicine, namely causality, interpretability, scalability, and data de-identification

Highlights from the Scientific Sessions and Meeting Outcomes

Overview of AI in Medicine and Oncology

• Current capabilities of various machine learning techniques and algorithms
• Requirements of AI informatics infrastructure

Current Research Topics in AI/Machine Learning and Challenges for Medical Applications

• Specific issues in the clinical implementation of AI
• Applications to data mining and retrospective studies of clinical trials databases

AI-specific Issues in Radiation Oncology and Medical Imaging

• Adversarial attacks on neural nets, data compression and heterogeneity

Action Items and White Papers

• Working Groups on specific issues of Radiomics
• Outcomes analysis, predictive patient response, and training and education in AI for clinicians and related professionals

For more information on this conference, contact Ceferino Obcemea, PhD — ceferino.obcemea@nih.gov