[Possibilities for integrating artificial intelligence into the diagnostic process].

Authors

Abstract

The changing working conditions in routine radiological reporting require the use of new methods, such as the implementation of artificial intelligence (AI) applications. What practical possibilities currently exist to integrate AI applications into existing imaging reporting processes? Assuming that the systems involved in the current reporting process, such as image management systems and various information systems, are to collaborate with AI applications, internationally recognized integration interfaces are required. The data exchanged also require standardized definitions to ensure interoperability across systems. Integration profiles as provided by Integrating the Healthcare Enterprise (IHE) offer a viable approach. The IHE profiles, such as Scheduled Workflow.b (SWF.b) and AI Workflow for Imaging (AIW-I) enable the combination of traditional imaging processes from patient admission up to AI-supported image reporting. The AI Result Assessment (AIRA) profile ensures legally compliant usage of AI Results. The AIR profile defines interoperable AI result formats that can be distributed within a defined reporting context via the integrated reporting applications (IRA) profile. The Interactive Multimedia Report (IMR) profile supports the persistence of final reports. Additionally, various data format specifications are currently evolving toward the Fast Healthcare Interoperability Resources (FHIR) implementation guides for the Image Diagnostic Report and the EU Imaging Study Report. The incorporation and use of AI in imaging processes are practically achievable using IHE profiles; however, integration based on structured data remains challenging and is currently reserved for larger institutions with the necessary IT resources.

Tags

Topics