Cerebriu, a Danish radiology AI company, envisions a future with autonomous scanners as their AI software automates brain MRI workflows ensures timely diagnosis of critical findings, shortens scan time, and reduces patient recalls. Guido Gebhardt spoke with Cerebriu’s CEO, Robert Lauritzen, and CCO Thomas Juhl Olesen about their innovation.
Thomas Juhl Olesen, Cerebriu CCO and Robert Lauritzen, Cerebriu CEO:
“We have extensively examined the possibility of enhancing the workflow and operational efficiency in acquiring images. The main objective is to ensure that the right image is obtained on the first attempt and that any critical findings are promptly addressed to shorten intervention time and thus improve acute patient treatment and outcome.“
What was the idea behind Cerebriu’s AI solution for MRI scans?
Robert Lauritzen: Cerebriu aims to improve the workflow for MR imaging by ensuring accurate, first-time image capture and promptly addressing critical findings, thus minimizing the number of repeat examinations and enhancing workflow.
Thomas Juhl Olesen: This approach not only saves time and resources but also improves patient management and healthcare system efficiency. We quickly identify abnormalities during scans, reducing unnecessary procedures. For the patients, this means reduced anxiety by streamlining the MRI acquisition procedure and avoiding callbacks.
Please explain how your solution works.
RL: Based on the initial MRI scan, our solution identifies possible findings requiring closer examination and allows tailoring of the additional sequences needed when the patient is still on the table. This reduces unnecessary comprehensive scans but also avoids diagnostic gaps – you can say it enables first-time-right imaging. This AI co-pilot approach enables technologists to make autonomous adjustments, streamlining the workflow and ensuring the right information is available for the radiologist’s reading. Also, an AI second reader can lead to improved detection accuracy. A recently published study in the European Journal of Radiology demonstrated excellent detection performance in a clinical validation setting.
“I believe in AI-driven brain MRI. Cerebriu’s unique solutions ensure a first-time-right image acquisition, minimizing interruptions to radiologists, repetition of scanning, and patient recalls.“
Lawrence Tanenbaum, MD, FACR,
VP and CTO of RadNet
It sounds like you need to integrate deeply into the system and might have to work closely with the MRI manufacturers…
TJO: Indeed, working closely with MRI manufacturers like for example, Siemens Healthineers is key for us. Our technology, integrated directly into MR consoles, enhances clinical MRI with seamless operation for technologists. While MRI manufacturers can develop similar technology, it requires a blend of clinical expertise and machine learning, an area we excel in. Our partnerships with OEMs aim to create more efficient, personalized scanners. We’re proud to be the only provider with a brain solution embedded in scanners, a testament to our specialized capabilities.
Does your solution only work on the newest scanner models, or can it also be added to older ones?
RL: That’s a great question and touches the core of our offering. We’ve conducted tests and validated our results globally to ensure diversity across various demographic characteristics, such as age, gender, and ethnic background. We’ve tested on different machines and models, ranging from those ten years and younger to those over ten years old.
Your solutions seem to be at the forefront of MRI technology. What are you doing to maintain this lead??
TJO:Â We are partnering with MRI manufacturers. Our focus is on embedding our advanced software within their hardware. We’re working on making scanners more efficient with software, potentially on a global scale. As a software company, we’re committed to setting new industry standards and innovating to maintain our lead. With our clinical workflow focus and strong clinical cooperation network, we tackle real clinical needs and see high potential and synergy in partnering with OEMs.
“We look forward to testing MR-integrated Al technology from Cerebriu in our Siemens scanner to assist our technologist in the patient journey.”
Prof. Mikael Boesen, MD, PhD,
Copenhagen University Hospital Bispebjerg and Frederiksberg, Department of Radiology
Your software is deeply integrated into scanner technology, but are there plans to integrate it into automated reporting systems?
RL: In radiology, we focus on standardizing and accelerating report generation by partnering beyond OEMs. We’re adapting to emerging post-processing technologies in the market. Our analysis can be transferred into structured reports. We plan to collaborate with others to integrate our technology and insights into these reports, enhancing workflow efficiency. Despite our technological focus, there is a distinct difference between our company and OEMs in this regard.
One of the biggest problems for customers buying AI is the lack of reimbursement policies. Where do you see the European AI market going?
RL: Introducing new AI technology in healthcare, especially in Europe, involves reimbursement challenges. However, the technology’s value, especially in efficiency and quality, makes it a compelling choice. The European healthcare sector, particularly in Germany, faces staff shortages and a growing patient population. AI tools like ours can alleviate these pressures. While standardized reimbursement isn’t always available, there are budgets for such innovations.
TJO: We have already signed customers and set competitive price points based on product evaluation and value creation. The European market, including Denmark and other countries, occasionally sees reimbursement for specific solutions, but our workflow-oriented approach typically relies on different budget allocations within institutions, focusing on overall efficiency improvements.
“Rigshospitalet is excited to work with Cerebriu at the forefront of radiology Al research for MRI stroke care.“
Martin Lundsgaard, MD, PhD, Head of Radiology, Rigshospitalet, Copenhagen
Radiology is a very conservative market, and radiologists only buy when they trust the company and the solution. How do you deal with this situation?
TJO: The radiology market is shifting towards software-enabled scanner upgrades as AI solutions such as Cerebriu’s continuously improve scanner performance and thus value. This trend is creating a stronger business case for our technology. Although the scanners are initially expensive, the ROI is accelerated by AI software shortening scan times and optimizing subsequent workflows since this enables increased patient volume while not risking higher recall rates, quickly paying for themselves by increasing patient throughput. Our technology enhances efficiency by reducing scan times and optimizing follow-up procedures. Our strategy to gain market trust involves changing industry standards and building partnerships with healthcare institutions, enabling them to derive more value from their offerings.
Another big problem in Germany is the cloud. But since you’re deeply integrated into the scanner, does your software run on-premise?
RL: Our technology is versatile in deployment: it runs directly on MRI scanners without external data analysis, supports on-premise setups with local servers, and offers traditional private cloud setups for hospitals that prefer cloud computing. This flexibility in technology distribution allows us to cater to various preferences and requirements, including those in Germany, where cloud usage in healthcare can be a concern.
What developments can we expect from Cerebriu in the future?
RL: We’re enhancing our products for greater efficiency, aiming for scans under five minutes. Our focus extends beyond the brain to other body parts. This year, we plan to introduce on-table decision support for devices, especially in neurovascular contexts. Our work is crucial as clinical trials show extended windows for treating stroke patients, necessitating advanced imaging for patient selection. We’re integrating scanners with cutting-edge technology, transitioning from a subjective to an automated, standardized basis for clinical decision-making.
Looking a bit further down the road, we’re researching and developing towards automating MRI workflows and making MR imaging accessible to broader use. We’re involved in several collaborative research projects with exciting innovations in the pipeline.
Meet Cerebriu at ECR 2024: https://cerebriu.com/ecr-2024/
