Authors' objectives: Artificial intelligence (AI) applications in diagnostic imaging are increasingly used to support image interpretation, prioritise examinations, and optimise clinical workflows. Against the backdrop of rising imaging volumes, limited human resources, and increasing diagnostic complexity, AI-supported digital health technologies (DHTs) are discussed as potential tools to alleviate pressures on radiology services. This assessment aimed to identify and prioritise AI-enabled DHTs relevant for Austrian hospitals and to evaluate the clinical effectiveness, organisational implications, and resource considerations of one prioritised application.

Authors' results and conclusions: Following a stakeholder-based prioritisation process, AI-supported analysis of chest X-ray images for suspected lung cancer was selected for detailed assessment. Three HTAs and two additional primary studies were included, comprising 15 unique primary studies evaluating 14 AI software products. Most studies were retrospective and conducted in single-centre settings. No study reported patient-relevant clinical outcomes such as mortality, morbidity, health-related quality of life, or safety. In some studies, AI-assisted image interpretation increased sensitivity without consistent improvements in specificity. Stand-alone AI systems demonstrated high negative predictive values but frequently low positive predictive values. Evidence on organisational effects and costs was limited. The currently available evidence is insufficient to demonstrate an added value of the selected AI-supported DHT for chest X-ray interpretation for lung cancer detection in Austrian hospitals. Substantial uncertainties remain regarding diagnostic accuracy, workflow integration, and resource implications.

Authors' recommendations: Prospective studies under real-world care conditions are required before routine implementation can be recommended.

Authors' methods: Relevant AI applications in diagnostic imaging were first identified and prioritised through expert consultation and structured shortlisting. A systematic review was then conducted according to a predefined protocol. The literature search followed a two-step approach, first identifying high-quality HTAs and systematic reviews, followed by an update search for primary studies. Evidence was synthesised narratively.

Project Status: Completed

URL for project: https://aihta.at/page/kuenstliche-intelligenz-im-gesundheitswesen-bewertung-der-klinischen-und-organisatorischen-effekte-ausgewaehlter-ki-anwendungen-in-krankenhaeusern/en

Year Published: 2026

URL for published report: https://eprints.aihta.at/1598/1/HTA-Projektbericht_Nr.171b.pdf

URL for additional information: https://eprints.aihta.at/1598/

English language abstract: An English language summary is available

Publication Type: Full HTA

Country: Austria

Organisation Name: Austrian Institute for Health Technology Assessment

Contact Address: Josefstaedter Strasse 39, A-1080 Vienna, Austria

Contact Name: office@aihta.at

Contact Email: office@aihta.at