Artificial intelligence (AI) has become an essential feature of many subspecialties of radiology, particularly neuroradiology, which heavily relies on magnetic resonance imaging (MRI) to help diagnose conditions that afflict the central nervous system.
“Our most meaningful goal is to improve patient care and outcomes. And we can do that by making better diagnoses with the help of AI tools,” says Matthew Kuhn, MD, FACR, a neuroradiologist with Radiology Partners in central Illinois, and a clinical professor of radiology at the University of Illinois College of Medicine in Peoria, Ill.
Dr. Kuhn is also the Chief Medical Officer for A.I. Analysis, Inc., the company that developed the Change Detector software system to compare serial imaging studies and eGad, a synthetic enhancement tool for brain, liver and breast gadolinium studies.
Neuroradiologists like Dr. Kuhn are increasingly relying on Rapid AI tools to evaluate stroke patients, including Rapid LVO, Rapid AI for stroke, and Rapid MRI, which automatically delivers quantified, color-coded MR diffusion and perfusion maps to help identify patients who may benefit from endovascular thrombectomy.
“Rapid AI for stroke helps us very quickly evaluate if the patient had a stroke, if it’s treatable, and how to treat it and follow up,” he says. “These AI tools take us from just looking at a medical image, to computer data mining the pixels and their relationships to each other, in order to find things that the human eye can’t see.”
AI-based technologies deliver accurate volumetric measurements to help assess how much healthy brain tissue can be saved from damage and then automatically populate that data into dictation platforms and other systems.
“AI is doing a lot of the repetitive and tedious work, but more than that – it’s taking us to a new level of patient care,” Dr. Kuhn says.
The benefits extend to worklist triage, where AI tools can identify critical findings, such as intracranial hemorrhage, even before the radiologist opens the image file. This helps prioritize cases according to urgency and severity.
“The AI tool can boost those studies to the top of the worklist for the neuroradiologist,” Dr. Kuhn says, thereby saving valuable time between scanning time and initiation of therapy.
Artificial intelligence can also help radiologists diagnose disease more accurately and monitor treatment more efficiently, particularly when it comes to brain tumors and demyelinating diseases.
“AI tools such as the Change Detector can often compare studies better than humans can, and they can measure the prior metastases compared to current metastases to give a volume load difference,” he says. This tool is also used for comparison of multiple sclerosis plaques over time.
Gadolinium-based contrast agents (GBCAs), widely used in MRI, are generally considered safe and effective for FDA-approved indications. However, in recent years, there have been reports of gadolinium deposition in the brain, bones and other tissues. AI-based tools can help reduce the amount of contrast needed to complete scans.
“AI tools can work very well to actually help us paradoxically improve our image quality even as we lower the gadolinium dose,” Dr. Kuhn explains, noting that a synthetic contrast enhancement algorithm, such as those used in eGad, can produce high-quality, synthetic, contrast-enhanced brain MR images, potentially helping to reduce gadolinium deposition concerns.
“Even when we apply a standard dose of gadolinium for a brain MR exam, the enhancing lesions appear much brighter and much clearer, so it looks as if a double or triple dose was given,” says Dr. Kuhn. “As we move forward, we have found that by using these AI applications we can boost both standard and low doses of contrast, allowing us to see more valuable information that can better direct patient care and follow-up.”
Smart injectors, he added, could work together with AI applications to reduce contrast dose while streamlining workflow for technologists.
“Smart injectors are an essential part of tailoring and administering the correct contrast media for the patient,” he says. “AI can make sure that the patient gets the correct contrast, at the safest dose and optimal pressure, and everything will be recorded as part of the medical record.”
While AI application development has been hindered by the COVID-19 pandemic, that hasn’t slowed down adoption of new technologies.
“We’re using more of these applications in our daily practices in neuroradiology. AI development is about networking, and people in the AI space have been working on making their products better, developing patents, and working with regulatory bodies during this down period for interpersonal relations,” he says.
One area in which Dr. Kuhn believes AI can make even greater contributions is the burgeoning field of radiogenomics, which pairs quantitative data from medical images with individual genomic phenotypes to construct a prediction model that guides treatment strategies and evaluates clinical outcomes. This supports the future of personalized medicine.
“Instead of just looking at a tumor, we can now look at its DNA. And by using AI tools, we can more accurately predict what the tumor’s genome will be, how sensitive it is to different therapies and therefore how best to treat it,” he explains. “For neuroradiology, we can use the T2-FLAIR mismatch tool to decide what type of mutation might be in gliomas in order to guide therapy. This is the future for medical imaging, and it is at our doorstep.”
Neuro MRI, he says, also has great potential to advance the understanding of mental health, and he would like to see neuroradiology expand beyond its present scope to include mental health concerns such as depression, psychosis and autism spectrum disorders.
“I’d really like to see evaluation and diagnosis of mental illness become a greater part of the repertoire of neuroradiology, and I hope that the neuroradiologists of the near future are actively involved in this area,” he says. “We are developing AI tools to assist with this.”
Back To Top
AI Takes Neuro MRI to the Next Level of Care. Appl Radiol.
McKenna Bryant is a freelance healthcare writer based in Nashotah, WI.