Perfusion Study Group Virtual Meeting

Artificial intelligence is increasingly being used in perfusion imaging applications and has the potential to achieve huge improvements in image quality, scan speed and the accuracy of automated image processing. However, there are also many challenges, including: generalisability across scanners, protocols and patient cohorts; access to sufficient training data; hallucinations within reconstructed images and the validation of performance metrics. In this session, we will showcase some novel AI applications in both non-contrast and contrast-enhanced perfusion imaging in the brain and body, outline some common pitfalls in AI-based image reconstruction, and discuss future opportunities and challenges in this field with our expert panel.