Radiogenomics: Bridging AI and IR to Inform Clinical Decision-Making Strategies

To review technical details and methodology of radiogenomics and its potential for enhancing interventional radiology (IR) therapy outcomes.



Background: Radiogenomics refers to the application of combining distinct image characteristics with individual genotypic data to construct models that predict tumor behavior{1,2}. Over recent years, radiogenomics has rapidly garnered interest in IR, especially within the interventional oncology domain, to predict tumor response or non-invasively determine molecular features to drive therapy strategies{3,4}. As methods combining artificial intelligence (AI) and machine learning algorithms with tumor gene characteristics, there is potential to further improve the precision and efficacy of current IR approaches.

Clinical Findings/Procedure Details: This education exhibit will: (1) provide an overview of radiogenomics and describe current methods for building data models and IR-tailored algorithms, (2) review recent data utilizing radiogenomics in IR with a special emphasis on interventional oncology, (3) present an evidence-based review of radiogenomic applications and their emerging role in IR therapies for hepatocellular carcinoma, renal cell carcinoma, and non-small cell lung cancer, among other tumors, (4) discuss challenges and limitations of radiogenomic techniques, and (5) review future directions of this data-driven tool for enhancing patient outcomes in interventional oncology.

Conclusion and/or Teaching Points: After reviewing this exhibit, viewers will become familiar with the diagnostic and clinical applications of radiogenomics, understand a general framework of how AI-based models are constructed using imaging and genetic data, and advantages and barriers of radiogenomics use in interventional treatments particularly in the oncology domain.