‘Scaling Smart’ — HBI webinar explores how diagnostic providers are applying AI and other technologies to address growing service demand
This week saw the latest event in HBI’s webinar series, “Scaling smart: how providers are leveraging new technologies for better diagnostic delivery”.
The background to the event
The purpose of the webinar was to explore the role that new digital tools, and artificial intelligence (AI) in particular, are playing in helping diagnostic providers to scale their capacity and improve quality, even in the face of converging demographic and funding challenges.
European healthcare providers face unprecedented pressure to scale their diagnostic imaging delivery while maintaining quality. With almost half of Germany’s radiographers, for example, expected to retire between 2030 and 2040 — a trend mirrored across Europe — and an increasingly elderly population requiring more complex care, the diagnostic imaging sector is in desperate need of capacity building and scaling tools that can offset the relative decline in practitioners.
Against this backdrop, artificial intelligence (AI) is emerging as a crucial tool in addressing the growing gap between radiologist capacity and patient needs.
At the webinar, Dr Gaspard d’Assignies, Chief Medical Officer and co-founder of Incepto Medical, a Paris-based AI platform serving 350 medical imaging sites across Europe, presented compelling evidence that AI implementation is now delivering measurable returns in both clinical outcomes and productivity.
Drawing on his dual perspective as a practicing radiologist and co-founder of an AI platform, d’Assignies highlighted how three major prospective studies in 2023 have demonstrated AI’s ability to maintain or improve cancer detection rates while significantly reducing radiologist workload.
AI impact is moving from the theoretical to the substantive
The largest of these, the Mammography Screening with Artificial Intelligence (MASAI) study involving 80,000 patients, showed that replacing second human readings with AI in breast cancer screening could reduce workload by 44% without compromising detection rates.
“This level of evidence is unprecedented in radiology,” d’Assignies explained. “We are seeing the kind of robust, prospective randomised controlled trials typically associated with pharmaceutical research.”
This evidence is particularly timely as radiologists face what d’Assignies described as “an avalanche of data” from accelerating image production speeds over the past two decades.
The benefits extend beyond breast cancer screening. In lung nodule detection, AI solutions have demonstrated a 10% improvement in sensitivity, while similar gains have been documented in bone fracture detection. According to d’Assignies, AI tools can dramatically improve risk assessment in breast cancer screening, identifying cases where risk might be as high as one in five, compared to baseline risks of one in 6,000.
Productivity gains are equally significant. Studies across various applications show time savings ranging from 30% to 77% in specific tasks. Real-world implementation data from one Swiss group demonstrated a 40% reduction in MRI reporting time “from discovering the image to the click of validation of the report.” The next frontier, d’Assignies suggested, lies in report generation, where generative AI could deliver productivity gains of up to 80%.
However, successful implementation requires more than just technological capability. Drawing on Incepto’s experience processing 350,000 exams monthly, d’Assignies emphasised the importance of a curated portfolio approach over a pure marketplace model:
“You can’t just bunch everything you can find and say to the radiologist ‘choose what you want’,” he explained. Success requires careful onboarding of stakeholders and clear KPI establishment from the outset.
An improving business case, and increasing ROI
The financial impact is becoming clearer. Implementation at one large Swiss healthcare group demonstrated a 5-7% impact on EBITDA through productivity gains, though d’Assignies noted in his response that this raises strategic questions about reinvestment priorities.
Looking ahead, d’Assignies sees the market for AI in medical imaging growing from its current size of approximately ~€2 billion to potentially ~€7 billion within the next decade.
While some larger healthcare groups are developing internal capabilities, he advocates for partnerships between healthcare providers and technology specialists, combining clinical expertise with technical capabilities.
One of the challenges providers face, from a strategic perspective, is deciding on their overall approach to competence and capacity development — should they Buy, Build, or Both? d’Assignies explained, in response to a question, that there is no one right method for any company, and that each requires their own bespoke approach informed by the needs of their patients, their business, the regulatory environment which they are navigating, and other dependencies. This again points to the need for strong partnerships and cross-sector collaboration.
With mounting evidence of both clinical and economic benefits, d’Assignies suggests the focus is shifting from whether to implement AI to how best to do so.
The video of the webinar, which delves much deeper into the themes of this article, is available to members of HBI’s Swapcard ‘Diagnostics Community’, which is free and available to all. The HBI Webinar Series continues in 2025, with the next event focussing on the Dental and Oral Health sector.
