Radiological analysis of the brain is complicated, time-consuming, and requires years of experience. In brain disorders like stroke, every minute counts. Having access to a comprehensive solution with high-precision oxygen availability maps and fast, accurate decision support, can mean life or death for a patient.
Cercare Medical is a software company founded in 2013 as a spin-out of the CFIN and MINDlab research centers at Aarhus University that have more than 20 years of research in neuroimaging and artificial intelligence, led by Professor Leif Østergaard and Professor Kim Mouridsen. Its focus is making state-of-the-art decision support systems available to hospitals and physicians, to accelerate and optimize clinical decision-making. The current product portfolio is collected in Cercare Medical Neurosuite. Besides modules for the automated assessment of brain perfusion and oxygenation, which is critical in a wide range of conditions including stroke, cancer and dementia, we provide profound decision support targeted for acute stroke using artificial intelligence technology to increase the efficiency of treatment decisions.
Creating the base for AI with top-of-the-line brain perfusion imaging
The success of an AI solution highly depends on the input data that was used to train the algorithm. It was important to gain high-precision brain oxygen availability maps.
In-depth understanding of brain tissue oxygenation is a key to the accurate assessment of the brain in a wide range of conditions. Research has shown that assessing capillary transit heterogeneity (CTH), besides traditional cerebral blood flow, can help significantly improve the accuracy of brain perfusion images and thus help healthcare professionals gain a better understanding of a patient’s brain condition.
The patented technology led to Capillary Function, Cercare Medical Neurosuite’s module providing a detailed picture of oxygen availability on CT and MR images, with automated post-processing. By showing an elaborate colour visualization map of brain tissue oxygenation, including microvessel integrity and function, radiologists get an in-depth insight into tissue perfusion necessary to evaluate and monitor disease progression.
The high-precision images seen with the Capillary Function module became a solid base to train AI algorithms and help radiologists increase treatment decision efficiency. “We believe that AI technology will eventually become a standard element in patient management,” says Professor Kim Mouridsen, Managing Director of Cercare Medical, “Machine learning will likely become a key to improving patient outcomes and quality of life, even though work is still needed to ensure ‘Artificial Intelligence’ are not just buzzwords but represent solutions that bring significant value to the patients.”
Incorporating AI for precision medicine in stroke
With stroke patients, faster diagnosis means accelerated time-to-treatment and reduced brain damage. When triaging patients suspected of having an acute ischemic stroke, physicians must determine the risk of hemorrhagic transformation, assess the extent of already-damaged tissue, and estimate the volume of salvageable tissue. Virtual Expert, our AI-driven module of Cercare Medical Neurosuite, accurately detects and provides such volumetric estimation using for better and faster patient care.
However, eliminating manual tasks in stroke is only one way of using AI for increasing efficiency. To provide radiologists with meaningful decision support, Cercare Medical continued building on past research and discoveries, to further apply AI in advancing imaging analysis.
Precision Medicine, Cercare Medical Neurosuite’s third module, takes advantage of large databases of previous stroke patients and displays the most likely stroke outcomes (infarcted and salvageable tissue). It incorporates the individual voxel level and provides healthcare professionals with high-resolution personalized risk maps for different treatment scenarios for each individual patient.
A longstanding challenge in predicting tissue outcome is that stroke progression is heterogeneous. Traditional models implicitly assume a fixed progression pattern for all patients, which is not accurate. This approach was changed, with AI technology learning to apply historic expert image reading of a patient’s actual tissue outcome to new cases, making individual treatment stratification possible.
“This is a great step towards choosing an optimal treatment for each patient. As no person’s stroke or reaction is exactly the same as another person’s, an individual approach is needed,” says Mouridsen. “There are still limitations, of course. Our solution cannot tell yet what was the reason for stroke, for example, which is important for post-stroke workup and further recovery planning. However, an ability to view most likely stroke progression for a given patient can help significantly speed up the decision-making process which is critical when time is brain.”
Making AI part of the workflow
It is important that AI technology can be easily integrated into the daily workflow of healthcare professionals to become commonly used. Cercare Medical Neurosuite is a vendor-neutral solution that is compatible with different scanners, PACS and medical software platforms.
Last November Cercare Medical announced its partnership with Siemens Healthineers. The goal of the partnership is to launch its state-of-the-art AI-based decision support and perfusion imaging on syngo.via OpenApps.
Meet Cercare Medical at ECR 2019, booth AI-14
We look forward to introducing Cercare Medical Neurosuite at ECR 2019. We believe that AI has the power to increase patient’s chances for survival, recovery and even prevention of health disorders, allowing patients to live a fulfilling life. Our mission is to bring these technologies to life.
Disclaimer: the solutions presented here might not be commercially available in your region. Please contact Cercare Medical at [email protected] for more information.
* Contents provided by Cercare Medical and adapted by the ESR office