Impact of AI-based Real Time Image Quality Feedback for Chest Radiographs in the Clinical Routine

Purpose To implement a tool for real time image quality feedback for chest radiographs into the clinical routine and to evaluate the effect of the system on the image quality of the acquired radiographs. Materials and Methods A real time Artificial Intelligence (AI) image quality feedback tool is developed that analyzes chest PA x-rays right after the completion of the examination at the x-ray system and provides visual feedback to the system operator with respect to adherence to desired standards of collimation, patient rotation and inspiration. In order to track image quality changes over time, results were compared to image quality assessment for images, acquired prior to system implementation. Results Compared to the image quality prior to the installation of the real time image quality feedback solution, it is shown that a relative increase of images with optimal image quality with respect to collimation, patient rotation and inspiration is achieved by 30% (p<0.01). A relative improvement of 28% (p<0.01) is observed for the increase of images with optimal collimation, followed by a relative increase of 4% (p<0.01) of images with optimal inspiration. Finally, a detailed analysis is presented that shows that the average unnecessarily exposed area is reduced by 34% (p<0.01). Discussion This study shows that it is possible to significantly improve image quality using a real time AI-based image quality feedback tool. The developed tool not only provides objective and impartial criticism and helps x-ray operators identify areas for improvement, but also gives positive feedback.