The first phase of the AIFI (AI for Imaging) project has been completed, according to the VZVZ (Association of Healthcare Providers for Healthcare Communication). The AIFI project is investigating how AI technology can be applied practically and safely in radiology. According to VZVZ, this first phase has demonstrated that a national AI infrastructure for radiology is technically and organizationally feasible and is valued by participating hospitals.Author: Journalism Team | ICT &amp; HealthThe insights gained now form the basis for the next step: determining how to proceed with a follow-up to AIFI, says Tim Smits, product manager for the Twiin portal at VZVZ. “AIFI has demonstrated that by centrally providing AI in healthcare—both technically and organizationally—the burden on hospitals is reduced and the adoption of AI can be strengthened. However, this remains a feasibility study. We are currently investigating whether and how a follow-up to AIFI can be implemented. By the end of this year, this will be reviewed with the clients, and a decision will be made regarding the next steps.&quot;<h1>National Project</h1>AIFI is a national project aimed at increasing the adoption of AI by making desirable and specific AI algorithms available to hospitals through a national infrastructure. This is intended to make (AI) software for medical diagnostics more easily accessible. By eliminating the “overhead” associated with local implementations, the project aims to reduce the burden on hospitals. It also provides a basis for joint procurement, which should lead to lower costs. In addition, AI can offer a partial solution to the growing workload in radiology, caused by a widening gap between the demand for imaging exams and the number of available specialists.The AIFI project and the pilot were announced in late 2023; the pilot itself launched in February 2025 with five hospitals (Radboudumc, Rivierenland Hospital, Gelre Hospitals, Twente Hospital Group, and Catharina Hospital). A deliberate choice was made to include a mix of academic centers and regional hospitals, each with different IT systems. Normally, AI pilots take place primarily in university medical centers (UMCs), whereas small and medium-sized hospitals can also benefit greatly from AI applications. The collaboration is also intended to prevent the wheel from being reinvented in multiple places.<h1>Three-Pronged Focus</h1>The pilot focused on three AI-supported applications: fracture detection, identification of incidental pulmonary embolisms, and skeletal age assessment. Zorgverzekeraars Nederland (ZN) funded the pilot, which was co-facilitated by the Dutch Society of Radiology (NVvR) and VZVZ. The AI algorithms are made available to participating hospitals via the existing national infrastructure of the Twiin Portal. All patient data is processed in a pseudonymized manner, ensuring that analyses can be conducted securely and without the risk of identification.Within the AIFI project, the Twiin Portal is being used differently than before. For example, images are now sent by Hospital A to an AI algorithm, after which Hospital A receives a result, rather than the images being sent from Hospital A to Hospital B. This made it possible to integrate AI into existing healthcare processes without many additional technical steps, says Smits. “We were able to effectively leverage the technology of an existing infrastructure to test AI in an accessible way.”<h1>Various Challenges</h1>According to Smits, the focus wasn’t limited to the technical challenges of deploying AI applications via a national infrastructure. “In a hospital, you approach issues primarily from the perspective of daily practice, but in a project, entirely different issues suddenly come into play.”Examples include legal responsibilities, governance, and scalability—areas where AIFI provides support. Smits: “These are all issues where you want to relieve the burden on hospitals as much as possible. I think it’s great that we’re trying to take those difficult issues off the hospital’s plate.”The legal aspect of the first AIFI phase turned out to be much more complicated than the technical aspect, partly because agreements and responsibilities had to be formalized. Issues such as: who will be the client; what role does VZVZ play; and how do you sufficiently involve the hospitals in the procurement and contracting process. Previously, the project team had hoped to build on existing contract models, but those turned out to be legally insufficient.Source: <a href="https://icthealth.nl/nieuws/eerste-fase-ai-samenwerkingsproject-voor-radiologie-afgerond" target="_blank" rel="noopener noreferrer nofollow">First Phase of AI Collaboration Project for Radiology Completed | ICT&amp;health</a>