AAPM-SIIM Symposium on Machine Intelligence in Medical Imaging: A Medical Physics Perspective
This Joint Symposium consists of two talks and aims to convey the valuable medical physics perspective on AI/ML/DL methods for image analysis with real clinical impact – one from Imaging and one from Radiation Therapy, In the last 3 years, this symposium has helped to expose new, synergistic relationships between medical physics and imaging informatics.
Towards Robustness in Medical Imaging AI: Exploring Strategies for Reproducible and Generalizable Results Horizons: An In-depth Exploration of Generative AI in Medical Imaging
In this talk, Dr. Drukker will discuss how characteristics and volume of data influence the achievable complexity of AI models used in medical imaging, as well as share strategies to optimize model performance based on available data aimed at maintaining generalizability. Then, she will address common challenges and pitfalls encountered during medical imaging AI model training and evaluation, and identify effective techniques to overcome them, promoting more reliable and robust results. Lastly, Dr. Drukker will discuss the significance of large-scale imaging repositories and associated meta-data in benchmarking AI algorithms, promoting reproducibility and enhancing the generalizability of research findings.
Objectives
Underline the influence of data on model complexity
Identify and mitigate model training pitfalls
Discuss the role of public imaging repositories and other benchmarking datasets
The Role of Machine Intelligence in Image-Guided Radiation Therapy
Image-guided radiation therapy (IGRT) employs images obtained during radiotherapy treatment sessions to accurately direct the administration of high-energy radiation to the intended target while minimizing radiation exposure to healthy organs. The emergence of cutting-edge machine intelligence techniques has facilitated the acquisition of high-quality images and their effective utilization in IGRT. In this presentation Dr. Jia will provide an overview of the pivotal role played by machine intelligence, delve into the challenges encountered in this field, and explore potential solutions to overcome them.
Objectives
Underline Underline the role of machine intelligence in image-guided radiation therapy (IGRT), including its applications and benefits in guiding the precise delivery of high-energy radiation
Identify and analyze the challenges faced in implementing machine intelligence techniques in IGRT
Explore potential solutions to address these challenges
Session Chair
Jeffrey H. Siewerdsen, PhD, FAAPM, FAIMBE
Professor, Department of Imaging Physics
Director, Surgical Data Science Program, Institute for Data Science in Oncology, The University of Texas MD Anderson Cancer Center
Conference Co-Chair
Speakers
Karen Drukker, PhD, MBA, FAAPM, FSPIE
Research Associate Professor of Radiology
University of Chicago
Xun Jia, PhD, DABR, FAAPM
Professor and Chief of the Medical Physics Division
Department of Radiation Oncology and Molecular Radiation Sciences
Johns Hopkins University School of Medicine
