• OSF researchers developed an AI tool that creates a 3D model of the aorta and generates hundreds of precise measurements in seconds.
• More accurate and consistent measurements could help doctors detect aneurysm growth earlier and determine the safest time for surgery.
• The technology could eventually support automated diagnosis and improve care for patients with conditions such as Marfan syndrome and Turner syndrome.

A team at OSF HealthCare is developing an artificial intelligence-powered tool designed to improve how physicians measure and monitor the aorta, a breakthrough that could help doctors detect dangerous aneurysms earlier and make more informed decisions about when surgery is needed.

The project, led by pediatric cardiologist Mathew Bramlet, MD, at OSF HealthCare Children's Hospital of Illinois, combines expertise from OSF HealthCare, Jump Simulation and Education Center, the University of Illinois College of Medicine Peoria and University of Illinois engineering researchers. The technology creates a highly detailed three-dimensional "digital twin" of a patient's anatomy from MRI or CT scans, allowing software to automatically analyze the aorta in seconds instead of requiring lengthy manual measurements.

The aortic arch project, funded through a Jump ARCHES grant, is aimed at improving longitudinal surveillance and the diagnosis and the tracking of aortic aneurysms, so they don’t tear or rupture. The aorta is the main and largest artery in the human body, carrying oxygen-rich blood from the left ventricle of the heart to the rest of the body. An aneurysm is a large bulge in that artery.

Accurate measurements are critical

Dr. Bramlet says measurements that document any expansion of the aneurysm need to be obtained with clear imaging with an expert eye that is generating reproducible measurements. Since high-risk surgery can hinge on one measurement, the accuracy of that measurement at one point in time, or over years as the growth of the aneurysm is monitored, becomes critically important. Among the complications doctors trying to prevent is an aortic dissection which recently killed Senator Lindsey Graham of North Carolina.

Currently, physicians often compare scans taken over many years to determine whether an aneurysm has grown enough to justify preventative surgery — a decision that carries significant risks if performed too early or too late. Manually measuring the aorta can take about 30 minutes and depends heavily on specialist expertise like that of Dr. Bramlet.

The new software automates that process, which Dr. Bramlet explains can generate up to 300 measurements in about 30 seconds while improving consistency.

“If we can capture this aorta and generate those same measurements with fancy new math and machine learning programs, we can distribute the expertise that I have accumulated over years and years and years, essentially into a program that would allow this type of analysis to be distributed at a larger scale.”

The automated approach also could allow researchers to analyze thousands of patient scans, helping develop more personalized treatment guidelines for conditions such as Marfan syndrome and Turner syndrome.

Aorta library will be created

The project is also creating a library of 3D digital models of healthy and diseased aortas that could eventually support automated diagnosis.

"The model library that we will have, where we can compare a diseased aorta to a normal aorta and look for the differences, it's a much easier automated process to get toward automated diagnosis, automated analysis and this is really the mindset of the entire lab,” says Dr. Bramlet.

Dr. Bramlet directs the Advanced Imaging and Modeling Lab at OSF Jump Simulation.

Beyond aortic disease, the underlying technology is designed to create digital models of multiple organs, opening the door to broader AI-assisted imaging analysis across medicine.

For Dr. Bramlet, it is exciting to see how this will be able to help provide better guidance for diagnosis and surgery determinations. The team is continuing to validate the technology through research while working toward clinical deployment – something Dr. Bramlet says is unique to OSF HealthCare.

“To have a hospital institution assign operational resources towards a clinical tool but I believe it is the commitment to serve people and to actually see the fruits of that labor being deployed within this institution really makes it a special place.”

In research conducted, this tool showed that automated 3D measurements closely match those done by clinicians, paving the way for a shift from traditional 2D imaging to more advanced 3D analysis.

Dr. Mathew Bramlet is also an associate professor of Clinical Pediatrics at the University of Illinois College of Medicine in Peoria.

Video clips with Dr. Mathew Bramlet

B-roll-Aortic aneurysm detection

Harvey Castro, MD, known professionally as DR GPT™ highlighted the growing role of artificial intelligence in health care during an OSF Innovation Showcase at Jump Trading Simulation & Education Center, emphasizing its potential to shift medicine from a reactive system to one focused on prevention and predictive care.

During an interview ahead of his presentation, Castro said people don't come to the emergency department until something's wrong, but he thinks advances in AI-powered predictive analytics could give health care providers opportunities to intervene before diseases progress.

“Now we’re going to be able to do predictive analytics and know that you’re going to have a heart attack in five years, 10 years from now. Or, like I was saying, this conversation you could say, ‘You know what, the AI is detecting that Dr. Castro has prediabetes and I didn’t know that. And so now the AI is telling me. Now, I’m like wait I need to lose weight, I need to take care of myself, and now I don’t even get diabetes because I’m jumping on it years before the actual condition.”

AI will speed time to diagnosis

Dr. Castro suggests that today’s five-year-old could live to 150 because of the advancements in longevity that artificial intelligence will bring. He also highlighted the technology’s potential in pediatric medicine and rare disease diagnosis, where patients and families often wait months or years for answers. AI systems trained to recognize patterns in symptoms and medical histories could help physicians identify rare conditions faster and begin treatment sooner.

“But what if we could take that year and a half and crunch it down to a day or two, and now we have a diagnosis?” Castro said. “Now as a parent, you know what's wrong. It’s got to be so frustrating not to know, especially when it comes to your children.”

Among those at the OSF Innovation Showcase were physicians who are treating kids at OSF HealthCare Children’s Hospital of Illinois. His message hit home as Dr. Castro reminded the audience that 50% of rare diseases impact children.

During his remarks, Castro also discussed how health care systems, such as OSF HealthCare, are increasingly building their own AI models using local patient data. He said using region-specific information is critical to ensuring AI tools accurately reflect the populations they serve. He calls that the “gold standard.”

“Let’s say I'm in Texas and I come up here and I say, ‘Here's my data. Let's use this AI here,’ it's not representative of the population here.” Castro stressed, “Right there you can make the argument that it’s going to hallucinate versus yours which is not. I think that’s key.”

Patients are already leveraging AI

For physicians and other providers, Dr. Castro advised that they should leverage AI because patients are already doing it. However, he said patients should not rely on the technology alone for medical advice. Because AI-generated responses can sound authoritative even when inaccurate, he said physicians must remain actively involved in evaluating and confirming information.

“The problem with ChatGPT is it sounds so real and it sounds so correct and it personalizes to you and it knows what words to use so that if it’s making it up, it sounds so real, so the average person won’t know is this hallucinating or is it saying something wrong or is it right? He emphasized, “That's why I say, especially health care, we need that doctor in the loop.”

Dr. Castro serves on a panel of health care leaders advising OpenAI, which operates ChatGPT, on best practices.

He shared that OpenAI is rolling out a new, optional safety feature for ChatGPT called Trusted Contact. This feature is designed for adult users (18+) and acts as a mental health safety net, allowing users to designate a friend, family member or caregiver to be notified if ChatGPT's systems detect that the user is engaging in conversations that indicate a serious risk of self-harm.

Despite concerns that artificial intelligence could make health care feel less personal, Castro argued the technology may actually improve patient interactions by reducing administrative burdens on physicians. AI transcription tools, for example, can allow doctors to spend less time typing notes and more time focused on patients.

In looking to the future, Dr. Castro envisions a day when wearables will combine with AI-connected devices in the home that will analyze daily waste, weight and other personal data and send a text to alert someone about a medical concern. In the end, he says artificial intelligence will be better for humanity and allow health systems to better care for people.

Harvey Castro, MD, MBA, known as DR GPT™, is an emergency physician, Chief AI Officer, author, keynote speaker, and physician futurist. He speaks and advises on AI in healthcare, digital health, clinical workflow, space health, and the future of human-centered medicine.
 

Media clips with Dr. Harvey Castro

B-roll-OSF Innovation Showcase

A potentially life-saving pilot program is launching in Bloomington, Illinois to detect a little-known but surprisingly common genetic condition that can dramatically raise cholesterol levels—and the risk of heart attacks and strokes—even in children.

It’s called familial hypercholesterolemia, or FH. It’s estimated that 1 in 200 Americans has the genetic disorder where the body can’t effectively clear LDL or ‘bad’ cholesterol from the blood.

Darrel Gumm, MD, vice president of OSF HealthCare Cardiovascular Institute says the pilot will identify people who, due to their genetics, might not know they are at risk.

“It’s estimated about 1.3 million people in the United States have it [FH] and 90% of those patients don’t know they have it, Dr. Gumm stresses. “So, it’s a tremendous opportunity for us to step in and find those patients, identify those patients and provide therapies that may prevent terrible things from happening down the road.” 

OSF HealthCare and OSF Innovation are using artificial intelligence (AI) to conduct a pilot program for detecting and treating FH. To identify high-risk patients for this pilot, the Digital Innovation Development team used a two-step approach. First, an algorithm flagged patients over the past 18 months who had concerning cholesterol levels. Then, they applied an AI tool to scanned physician notes to detect clues—such as yellowing of the eyes or hardening of the Achilles tendon—signs a doctor may have observed but not flagged as FH-related.

People inherit FH from one or both parents, and it puts them at risk for cardiovascular events far earlier in life than normal, sometimes even in childhood.

Two Types of FH

According to Dr. Gumm, there are two kinds of FH:

• Heterozygous FH, inherited from one parent, which can lead to early heart attacks or strokes—before age 65 in women and 55 in men.
• Homozygous FH, a rarer and more severe form inherited from both parents, which can lead to strokes and heart disease in teenagers or even young children.

While most people begin cholesterol screenings around age 40, those with FH can silently develop sky-high levels from a young age.

"They can be a very healthy person, think they’re doing great, be a marathon runner and really they could have astronomically high cholesterol that’s not of their fault,” explains Ryan Loudermilk, a strategic program manager in Performance Improvement at OSF Innovation.

The new screening program was launched in Bloomington, starting with support from a strong primary care network involving 9 providers including Richard Ginetti, MD, who is a champion for the effort. It started with 200 patients found to potentially have FH. If testing confirms the diagnosis, individuals will be offered treatment, typically statins, which can reduce cardiovascular risk by up to 80%.

Dr. Gumm says every time someone is identified with FH, outreach will occur to encourage screening for parents, siblings, and children.

“One of the things I love to tell my patients is the best way to treat a heart attack is not have one. Same thing with a stroke. So, even if they’ve had a heart attack and survived it, now we’re going to be aggressive at not letting the second one occur. But here’s a chance to never let it occur in the first place. It’s primary prevention … really fantastic.” 

Education in a Short-Attention-Span World

Understanding FH—and the importance of early detection—isn’t easy in a world flooded with information. That’s why the pilot program teamed with OSF Innovation’s Medical Visualization team, to create for people identified at possible risk, mobile-friendly education designed to meet people where they are.

Loudermilk emphasizes it’s not just a brochure. The visuals use a unique, semi-animated shimmer effect that adds emotional weight without overwhelming the viewer.

“It’s something that can be easily read on their cell phone; they don’t have to log onto a computer. It’s something that’s eye-catching, something that’s straight to the point – what your risks are, why it’s important for you to get tested and checked … thinking about family members, so it speaks to that. It speaks to the ease of this. That’s the other big piece of this.” 

A Simple Test with a Big Impact

The testing process is easy. It’s just a cheek swab with no needles or blood draw. It takes a couple minutes.

The pilot is tracking how many people go on to be screened after the initial outreach with a goal of at least 30%. The initiative will follow not just how many patients test positive, but how many then go on to get family members tested, the true measure of ripple effect success. Additionally, the pilot includes having the Medical Visualization team provide insights about which content drove the most engagement.

The FH pilot is part of a broader effort at OSF HealthCare to advance care through personalized medicine that is specific to a person’s genetic make-up. Loudermilk sees a time in the future where genetic testing will be even easier.

“Where I could see this going – a patient doesn’t even have to come in. These kits could actually be mailed to them, to their house and they could do the check swab themself and get it mailed in. So, those who don’t want to deal with making an appointment and seeing their provider in person, they could do a virtual visit and have the kit sent to them. That’s the future of all of this.” 

If you receive primary care in Bloomington, talk to your provider if you think you would like to be screened for familial hypercholesterolemia.

Dr. Darrel Gumm

Ryan Loudermilk

B-roll-Familial Hypercholesterolemia (FH)

Imagine being a heart surgeon preparing for a complex case. Traditionally, you rely on stacks of 2D images from CT scans to build a mental picture of your patient’s anatomy. But the heart isn’t static – it twists and pumps blood in a rhythm unique to each person. Now, thanks to a breakthrough at OSF Innovation, surgeons can “see” their patient’s heart beating in 3D before ever stepping into the operating room.

This innovation is called the 4D Heart. Unlike standard 3D models, which capture a single snapshot in time, the 4D Heart recreates every phase of a patient’s cardiac cycle, allowing doctors to watch their heart beat virtually. And it isn’t just a generic animation. Rather, it is the creation of a digital twin.

“This is taking a CT scan of a patient, a retrospectively gated CT scan, and converting a patient’s own heart – in each phase of that cardiac cycle – and making it beat in a virtual environment,” says Matthew Bramlet, MD, who directs the Advanced Image and Modeling (AIM) Lab at OSF Innovation. The lab is focused on improving the understanding of complex anatomy for surgeons and diagnosticians. Dr. Bramlet also specializes in congenital cardiac MRI for OSF HealthCare Children’s Hospital of Illinois and is an associate professor of Clinical Pediatrics at the University of Illinois College of Medicine in Peoria and the University of Illinois Grainger College of Engineering in Urbana-Champaign.

Dr. Bramlet’s efforts began in 2018 with manually segmenting parts of the heart so it could be viewed in virtual reality. It continued with an infusion of grant money two years ago. Using an Innovation for Health (IFH) grant with Bradley University in Peoria, Illinois, a team used machine learning to convert standard medical images into a sequence of 3D models that, when played sequentially, create a 3D beating heart – the 4D heart. The initial process took an entire summer.

That made the technology nearly impossible to use in real world clinical care. But in late 2024, researchers achieved a major breakthrough. With the help of engineers at the University of Illinois Urbana-Champaign, AIM lab engineers trained neural networks on hundreds of high-quality CT scans and automated the entire process. Now, what once took months can be done with a single click.

“Have you ever seen a flip book where you see a drawing move? That’s what we’re doing with these 3D models in a digital, virtual space.” Dr. Bramlet explains, “We’re just turning them off and on in rapid succession until we have that beating heart.”

For surgeons, the benefits are game-changing. They can see not only the anatomy but also how it changes with every beat. This is especially critical for conditions such as hypertrophic obstructive cardiomyopathy, where the problem lies in how the heart contracts. Even more surprising, the AI began rendering valves — structures notoriously difficult to see on CT scans.

“Now we can see that valve which we really couldn’t see before. So, repeat valve procedures are very valuable for this. TAVR or any valve in the cath lab, we’re finding good use with these as well as the hypertrophic cardiomyopathy cases.”

Hospitals across the country are taking notice, sending scans to OSF and receiving beating heart files in return. Some don’t even have the technology to view them yet. Dr. Bramlet says OSF Innovation is pushing the boundaries so far forward, some other health systems will need to get on board with the technology to be able to visualize the data being sent to them.

For now, the ability to view a 4D Heart is already changing how surgeons think about the body’s most vital organ. In some cases, the new view is confirming a surgery plan. In other cases, including a most recent pediatric heart case from another hospital, it changed the approach. Dr. Bramlet says as surgeons interact with the models, they are amazed.

He shares, “I see those lightbulbs go on and there’s nothing like being able to observe and interact with these 3D models in a format, and especially a 4D model, in a way they’ve never been able to see them before, and it will become the standard of care, I have no doubt.”

The ultimate vision is bold: a patient enters the CT or MRI scanner, and their digital twin heart comes out the other side – ready for analysis, surgical planning and even analysis of AI-driven measurements like ejection fraction or blood flow strain to help with diagnosis and a treatment plan.

How long until that happens? Hard to say because as Dr. Bramlet points out, “seeing is believing,” and it might take time for health systems to invest in the technology to make the vision become reality. Government reimbursement for the care will also drive adoption. For now, the work is being covered through research grants, most notably; the Jump ARCHES endowment through the Health Care Engineering Systems Center at the University of Illinois Urbana Champaign and OSF HealthCare.

Dr. Mathew Bramlet

B-roll-4D Beating Heart

Sepsis is a critical health issue worldwide, but the advancements in artificial intelligence are starting to offer a better path forward for early diagnosis and treatment.

OSF HealthCare Saint Francis Medical Center in Peoria, Illinois was one of 10 hospitals taking part in clinical trials that led to the first-ever Federal Food and Drug Administration authorization to market an AI diagnostic tool for sepsis. Chicago-based company Prenosis, which spun out of the University of Illinois in Urbana Campaign (UIUC), created The Sepsis ImmunoScore^TM.

OSF Saint Francis hospitalist Anwaruddin Syed, MD, began working with researchers at UIUC while in residency at a hospital in Champaign. He took that experience to OSF Saint Francis where he was the lead investigator, continuing research in the Prenosis clinical trial. As a hospitalist - a doctor who provides care for patients at a hospital - he wanted to help solve the challenge of a serious and often deadly complication.

“In the US alone, there’s an estimate of 1.7 million American adults who develop sepsis every year out of which 350,000 people die or go into hospice. The numbers are eye opening.” (:19)

Dr. Syed explains sepsis happens when the body tries to fight an infection.

“When you have a really bad infection and your immune system is compromised by that infection, then the infection takes over. That’s why it’s important to identify that infection early on, before that infection can take you over.” (:19)

OSF contributed to the pool of 100,000 blood samples and de-identified patient data to help train algorithms to identify patient vital statistics, such as heart rate and oxygen levels, most associated with developing sepsis. It looks at biomarkers – biological molecules found in blood, tissues and other bodily fluids that lead to disease. The ImmunoScore is based on a total of 22 factors to identify risk, which it then put into one of four categories.

The risk score can be integrated with a patient’s electronic health record. Care team members can see what factors impacted the patient’s risk score. Dr. Syed says it can be an add-on to the protocols and information hospitals already use to help better identify patients at risk and for managing those with a confirmed sepsis diagnosis. In other words, it’s one more tool to help physicians.

According to a 2022 study published in Nature Medicine, in severe cases, the then-UIUC AI model detected sepsis an average of six hours earlier than traditional methods. Dr. Syed says Prenosis will conduct additional studies to demonstrate the tool’s accuracy and impact on clinical decision-making.

In the meantime, Dr. Syed is excited OSF has participated in such important research.

“It honestly feels amazing, knowing that our work and the trust and contributions from our patients here at OSF played a role in developing this first-ever FDA sepsis AI tool. It’s really rewarding. It’s a game-changer in patient care and I’m proud to have been a part of it.”

Other companies have developed sepsis diagnostic tools, but Prenosis wanted the FDA approval before marketing its software. The company worked for 18 months to receive the go-ahead to market and sell The ImmunoScore under a process in which the FDA reviews and regulates artificial intelligence/machine learning software as if it is a medical device.

Dr.  Anwaruddin Syed

B-roll of sepsis patients, treatment