Artificial intelligence integration into the CT system with a focus on CT planning and scanning-review

Thomas Flohr, Michael Sühling
Br J Radiol. 2025 Jun 16:tqaf129. doi: 10.1093/bjr/tqaf129. Online ahead of print.
Abstract
For many years, the development of computed tomography (CT) has been driven by technological advances in CT scanners to improve their performance, reduce radiation dose and expand their clinical capabilities. Meanwhile, another focus of CT development has emerged-on improving and automating the scanning workflow and on approaches to better display the clinical information of CT images, supported by artificial intelligence (AI). In this review, we provide an overview of AI-driven functionalities integrated into the CT system. We discuss 3D camera-based patient positioning, patient characterization, and scan guidance, and automated scan range definition based on the localizer scan. We present approaches to automatically predict patient-specific individual scan delays for CT angiographic scans, and we discuss algorithms for AI-supported initial image quality checks. Finally, we present modern approaches that enable automatic calculation of anatomically aligned reconstructions and advanced visualizations directly on the scanner. We focus on clinically established methods, but we also present some prototype algorithms as a glimpse into the future.