AI cuts time for radiologists reporting fractures on x-rays

Burnout in radiology is a well-documented syndrome of emotional exhaustion, depersonalization, and diminished professional efficacy. Although the sources of stressors may differ, they all relate to increased workload. During any given shift, many radiologists may interpret several hundred examinations, which, depending on the number of images per study, could amount to thousands of images. Night shifts and on-call duties only compound this experience. Hospitals now demand time-sensitive reporting, with expectations for prompt turnaround and clinical disposition, as part of the workload.

Darmstadt – „Hier sehen wir eine nicht verschobene Fraktur im Bereich der Fußwurzel“, sagt Dr. Sabine Geck und zeigt auf ihren Bildschirm in der Notaufnahme des AGAPLESION ELISABETHENSTIFT DARMSTADT. Auf der Röntgenaufnahme ist ein viereckiges Kästchen rot eingezeichnet, „hoher Verdacht auf eine Fraktur“, steht auf dem Bildschirm. Die Besonderheit: Die Anmerkung hat dort kein Arzt hinterlassen, sondern das KI-Programm „AZmed“.

Radiology departments worldwide are facing increased numbers of imaging studies than ever. On the other hand, they are also facing reduced staff, increased caseloads, and increased risks of making mistakes. Therefore, it is no surprise that many departments are looking to install artificial intelligence in radiology to assist with speeding up diagnosis. But while speed is crucial, it should not be at the cost of diagnostic precision. The question most medical professionals are asking is straightforward: Can AI enable radiologists to speed up diagnosis yet maintain identical high levels of patient care?