Road to RSNA 2020: Artificial IntelligenceBy Erik L. Ridley, AuntMinnie.com staff writer
November 9, 2020
As has been the case in recent years, artificial intelligence (AI) will have a dominant presence at the meeting, permeating throughout the meeting program. In our preview, we look at topics and issues in the informatics track, as well as provide coverage of AI-related developments in other subspecialty areas.
The growing maturity of AI technology in radiology will be evident at RSNA 2020, with studies increasingly focusing less on model performance in narrow tasks and more on how it will be integrated into radiology workflow. Other notable, hot areas in AI include non-image-based AI, federated learning, and using AI to reduce the amount of time needed for image acquisition.
Among the various subspecialties, look for research on utilizing AI for rib fracture detection and treatment outcome prediction, as well as in a wide range of neuroradiology applications.
AI is also set for a starring role in breast imaging presentations at the meeting, featuring studies assessing its utility in mammography, as well as breast ultrasound and MRI. And even though the abstract submission deadline for RSNA 2020 came early in the pandemic, there will still be a number of talks describing how AI can help in the fight against COVID-19, including, for example, predicting patient prognosis based on quantified changes in the lung.