Simplified IT and sales processes key to accelerating AI adoption in imaging — DeepHeath’s Niccolo Stefani
US-based outpatient diagnostic imaging services provider RadNet has recently acquired French radiology AI company Gleamer in an all-cash transaction valued at €230 million (~$270 million).
Gleamer will be integrated into RadNet’s wholly-owned subsidiary DeepHealth, making it the largest provider of radiology clinical AI solutions worldwide.
RadNet sees a sizeable opportunity in what it calls the “AI-enabled health informatics market,” which comprises clinical AI solutions and radiology informatics. The company estimates this market to be worth around $5.1 billion in 2024, growing towards $7.7 billion by 2028, at an 11% growth rate. Within this market, clinical AI solutions is “by far the fastest-growing segment” with a 26% growth rate, while radiology informatics is growing at 5%.
HBI spoke to Niccolo Stefani, Business Leader Population Health & Clinical AI, RadNet DeepHealth, about what this acquisition means for the AI platform and the overall adoption of AI in imaging.
HBI: What was the strategic rationale behind acquiring Gleamer, and how does it strengthen DeepHealth’s AI platform?
Niccolo Stefani: We have been active in what we call the population health space, which is focused on screening. With Gleamer, we’re expanding into routine imaging, including both acute care, such as trauma, as well as diagnostics.
Niccolo Stefani, Business Leader Population Health & Clinical AI, RadNet DeepHealth
The acquisition also accelerates our commercial and global reach. Additionally, the Gleamer portfolio further drives operational efficiency at RadNet. With 2.8 million X-rays performed annually, even a 5% improvement in speed could yield significant financial benefits.
We are headed in a direction where AI can be seen more as a commodity, and only the full end-to-end AI-enabled workflow will have an impact to carry the right value and drive the expected outcome at scale
HBI: What capabilities or geographies are you prioritising next? Are further European deals on the horizon?
Stefani: Europe is a key market for us, with particular focus on Germany, where the National Lung Cancer Screening Program is launching. We have experience in the UK with lung screening and the prostate trial, and we aim to apply that in Germany.
We are now across multiple European countries — Spain, Germany, the UK, France, Italy, and Eastern Europe.
And in terms of capabilities, we recently presented at the European Congress of Radiology (ECR). For lung screening, our AI can now detect lung nodules, perform calcium scoring, and assess emphysema — all from a single scan. These features will be integrated into a suite that will allow the clinician to use the data in what we call a semi- automated pre-fill report. The goal is to reduce non-clinical tasks and to streamline the workflow.
HBI: Screening models and reimbursement differ widely across Europe. In which markets does the business case for AI in breast imaging already work, and where is it still difficult?
Stefani: Each market is different. Reimbursement approaches vary across France, the UK, and Germany. In general, reimbursement tends to focus on AI in genomics rather than clinical detection in radiology, which is where our main differentiator lies.
The business case for AI often depends on workflow and resource pressures. For example, the UK has around 4,000-5,000 radiologists for the entire country, and the volume of scans per radiologist is high. It is a market where efficiency makes a lot of sense, because you don’t have enough radiologists.
In a market like Italy, it’s less about the shortage of radiologists, because you have around 16,000 radiologists in Italy for the same population as in the UK. It’s more about providing access to patients to the same level of quality. In France, there’s a lot of cost pressure.
Overall, our main advantage and ROI come from workflow improvements, that is, reducing reading time and potentially lowering biopsy rates, which translates to cost savings for health systems. Reimbursement is a supplementary benefit rather than the primary driver.
HBI: AI in imaging is still fragmented and unevenly adopted. What are your expectations for the adoption in 2026?
Stefani: The use of AI in imaging remains quite fragmented. From a commercial perspective, the market’s fragmentation makes it difficult for buyers to see adoption as sustainable. Adoption will really accelerate only when IT, contracting, and sales are simplified and integrated. While most people agree that AI can be valuable, implementing multiple disconnected solutions is costly. That’s why we are investing in a unified IT platform that provides access to all our AI solutions in one place, streamlining deployment and reducing the overall cost of ownership.
HBI: For investors looking at imaging, is AI now a prerequisite for value creation, or is it still too early to be decisive?
Stefani: AI is increasingly becoming a prerequisite. There may still be a lack of understanding in terms of how sustainable it is to deploy. But across healthcare providers in general, AI is considered essential for optimising workflows, managing resources, and expanding patient access.
The market is large. Adoption will likely be phased over five, ten, or fifteen years, but I envision a future where nearly every single examination would have an AI component.
We would welcome your thoughts on this story. Email your views to Hemani Vipul Sheth or call 0207 183 3779.
