A Dutch Hospital Scanned 2,000 Extra Patients in 2025 — Without Adding a Single Scanner

2026-05-31 bio

Zutphen, Sunday, 31 May 2026.
Gelre Ziekenhuizen used AI-powered MRI software to scan nearly 2,000 additional patients in 2025, cutting wait times without expanding physical capacity.

A Healthtech Breakthrough in the Heart of the Netherlands

This story sits squarely in the domain of healthtech — the application of advanced software and artificial intelligence to improve clinical outcomes and operational efficiency in healthcare settings. The protagonist is Gelre Ziekenhuizen, a regional hospital group operating across Apeldoorn and Zutphen, two cities in the Stedendriehoek region of the eastern Netherlands [1][2]. The innovation in question is Deep Resolve, an AI-powered MRI image reconstruction software developed by Siemens Healthineers [1][2][3]. Deployed across Gelre’s MRI scanners in 2024, Deep Resolve produced its most measurable impact in 2025 — a year in which nearly 2,000 additional patients received MRI scans compared to the prior year, using virtually the same physical scanning infrastructure [1][2][3]. In a healthcare system where waiting lists for diagnostic imaging are a persistent and well-documented problem across Dutch hospitals [7], that figure represents a meaningful structural improvement, achieved not by pouring concrete or installing new machines, but by making existing equipment significantly smarter.

How Deep Resolve Works: Intelligence Applied to the Imaging Process

Understanding why this result is significant requires a brief look at how MRI scanning functions and where AI can intervene to create efficiency. Conventional MRI scanners capture raw signal data from the body and reconstruct that data into diagnostic images — a process that historically demanded long acquisition times to ensure sufficient image sharpness and resolution [GPT]. The longer each scan takes, the fewer patients a machine can process per day, and the longer waiting lists grow. Deep Resolve, developed by Siemens Healthineers, uses artificial intelligence to sharpen MRI images during the reconstruction phase, allowing radiologists to acquire diagnostic-quality scans in a shorter time without sacrificing — and in many cases actively improving — image quality [1][2][3]. The system works by applying trained neural network models to the raw imaging data, effectively learning to distinguish meaningful anatomical signal from noise and enhancing image clarity even when the acquisition time is compressed [GPT]. The clinical evaluation at Gelre Ziekenhuizen was conducted by clinical technologist Suzan Everink and clinical physicist Djim Kanters, who analyzed data from all three of the hospital group’s MRI scanners [1][3]. Their findings confirmed that the deployment of AI not only enabled more efficient use of MRI capacity, but also delivered improvements in image quality across a wide range of scan types [1].

The Numbers That Make the Case

The quantitative outcome of the Deep Resolve deployment is striking in its simplicity. In 2024, Gelre Ziekenhuizen introduced the software onto its MRI scanners in Apeldoorn and Zutphen [1][2]. By 2025, the hospital group was able to scan nearly 2,000 additional patients compared to 2024, operating with virtually unchanged scanner capacity [1][2][3]. To put that in perspective, the gain was achieved entirely through software — no additional capital investment in hardware, no new scanner suites, no additional physical footprint. The evaluation of all three MRI scanners within the Gelre network provided the data foundation for these conclusions [1][3]. Clinical physicist Djim Kanters summarized the impact directly: “This is a fine example of how technological innovation, collaboration with Siemens Healthineers, and the support of an engaged fund directly contribute to better care for our patients. The use of AI also seamlessly aligns with our digitalization vision” [1][2]. The practical result for patients is unambiguous: shorter waiting times for MRI diagnostics, a bottleneck that, as broader Dutch healthcare data shows, has been a systemic challenge across the sector [7].

Foundation Funding as a Model for Regional Healthtech Adoption

The financing mechanism behind this deployment deserves as much attention as the technology itself. The purchase of the Deep Resolve software was made possible through a grant from Stichting Oude en Nieuwe Gasthuis Zutphen, a local healthcare foundation based in Zutphen [1][2][3]. This is not a government subsidy program, a venture capital-backed pilot, or a national digital health initiative — it is a regionally rooted philanthropic foundation directly enabling clinical innovation at the ward level. For healthcare innovation professionals and policymakers, the Gelre case offers a replicable framework: targeted foundation funding, applied to a commercially available and clinically validated software solution, can de-risk adoption and generate measurable patient-benefit outcomes within a single calendar year. The broader Dutch healthcare landscape is actively grappling with questions of how to scale digital health solutions efficiently. Reporting from May 2026 highlights that initiatives such as digital monitoring apps for chronic bowel disease, developed in part by Gelre Ziekenhuizen in collaboration with UMCG and MUMC+, have struggled to scale nationally because hospitals and insurers fear revenue losses [7]. Against that backdrop, the Deep Resolve rollout — funded philanthropically and evaluated rigorously by in-house clinical staff — stands as a counterexample of what focused, outcome-driven adoption can achieve when financial misalignment is removed from the equation.

AI in Dutch Medical Imaging: A Growing but Uneven Landscape

The Gelre deployment does not exist in isolation. Across the Netherlands, hospitals are at various stages of integrating artificial intelligence into diagnostic imaging workflows, with results and readiness levels that vary considerably. In the field of breast cancer detection, for instance, radiologist Sofie De Vuysere noted in May 2026 that Computer Aided Detection (CAD) AI tools carry genuine added value but still require radiologist interpretation and are not yet fully reliable as standalone diagnostic instruments [4]. Separately, a collaboration between Philips and Disney — reported on 30 May 2026 by Eindhovens Dagblad — is addressing the patient experience dimension of MRI scanning, using immersive technology to make the scan environment less distressing for children and adults alike, with Catharina Ziekenhuis in Eindhoven already equipped with the relevant technology [5]. These parallel developments illustrate that AI’s integration into medical imaging is proceeding along multiple axes simultaneously: efficiency (as in the Gelre case), diagnostic accuracy (as in CAD-assisted breast imaging), and patient comfort (as in the Philips-Disney initiative). The Gelre Ziekenhuizen story is notable precisely because it demonstrates a completed deployment with verified, quantified results — not a pilot, not a plan, but a live clinical system that scanned nearly 2,000 more patients in 2025 than in 2024 [1][2][3]. In a sector where proof of concept can take years to translate into structural practice, that distinction matters considerably.

Bronnen

healthtech AI diagnostics