Delivering a MarginDx prototype in one year: A look inside with Stephen Boppart

10/6/2025

Written by

Last August, the Interdisciplinary Health Sciences Institute (IHSI) at the University of Illinois Urbana-Champaign announced an up to $33 million award from the Advanced Research Funding Agency for Health (ARPA-H) for an end-to-end solution that will drastically improve surgical outcomes for cancer tumor removal. The project, Margin Diagnostics (MarginDx), combines optical imaging technologies with AI screening tools to ensure, in real time, that tumor tissue and cells are completely removed during surgery.

IHSI Director Stephen BoppartDr. Stephen Boppart, a professor of electrical and computer engineering and bioengineering at Illinois, and director of IHSI, is the principal investigator for the multi-team project, which includes a multi-institutional and multidisciplinary team of researchers and clinicians at Illinois and Mayo Clinic, along with industry and clinical collaborators.

One year into the 5-year project, we asked Boppart to share with us the progress and challenges of the ambitious MarginDx development process. Here are a few things to know about the project:

  • MarginDx will allow for immediate microscopic assessment of the tissue that the surgeon is removing or leaving behind, essentially giving the surgeon real-time microscopic vision. This is different from the current standard-of-care, which relies on review from a pathologist outside of the operating room, often while the patient and surgical team are waiting in the OR.
  • MarginDx will integrate two types of optical imaging. The first is optical coherence tomography (OCT), which can generate high-speed images that allow visualization and assessment of a large tumor cavity in minutes. The second is non-linear optical (NLO) imaging, which can achieve high sub-cellular resolution at the same scale as standard histology for precision imaging and diagnostics.
  • Both OCT and NLO images are collected label-free, which means it does not require injecting any contrast agents or applying dyes or stains that are sometimes used to show contrast between tumor and normal tissues.
  • The imaging system will integrate directly with an AI-driven digital tissue analysis system to classify regions as either cancer or normal, and provide real-time feedback to the surgeon for immediate surgical decision-making.
  • To ensure immediate use in the clinic, MarginDx must be delivered affordably and fit within existing clinical workflows. The team will engineer a freehand scanning surgical imaging OCT probe accompanied by a robotic arm for high-precision NLO imaging.
  • The goal at the end of the five-year project period is for the MarginDx platform solution to be clinically proven and commercially viable for rapid dissemination and use across the nation and globe.

Q: First, congratulations on a successful first year of the MarginDx project! How would you describe the work to this point? How is it different from other types of projects you’ve been involved in?

A: Thank you for the opportunity to share this update on our MarginDx project. It has been a whirlwind year for all of us, but one that is just as exciting and stimulating as it has been challenging and a lot of work! We have an incredibly talented team of over 75 researchers and staff across multiple institutions and multiple areas of expertise, including The Grainger College of Engineering, IHSI, and National Center for Supercomputing Applications (NCSA) at Illinois, clinical experts from Mayo Clinic and Carle Health, and industry partner Eleuthra Photonics. The scope of the project makes it very different from the federal research grants from NIH or NSF that many of us have been a part of to investigate fundamental questions or problems in biomedical science and technology. In fact, this project is a contract from ARPA-H to explicitly build and demonstrate a new platform to solve a specific taxing problem in surgery, that of high positive margin rates following surgical resection of tumors. Our project and team are operating in a mode that is more like a start-up company than an academic research laboratory, which requires a very different mindset and milestone focus in order for us to be successful. I would say that this has not been a surprise, however, knowing that ARPA-H was created to go after big high-risk high-reward solutions to address a critical need in a relatively short period of time, and rapidly commercialize and disseminate this technology so it can begin having immediate impact on the outcomes of so many cancer patients.

Q: Representatives from ARPA-H visited U. of I. in early September. What did the team need to have ready for this first year site visit? How did it go?

operating room with surgical robot arm, simulated patient, and three people in lab coatsA: We were thrilled to host the ARPA-H program leadership during their visit with us at Illinois. It was our opportunity to demonstrate in-person just how far we had come in one year, and to have them actually engaged with and use our bench-top system to do some imaging themselves. Our Year-One Prototype was a fully functional bench-top system that had all the various imaging, software, and computational pieces built and functional, just not optimized or integrated into a single cart platform. Getting to this point on such a fast timeline really was a heroic effort, particularly coordinating all the progress between our various optics, robotics, AI, data, human factors, clinical, health economics, and translation teams.  I think the visit went well. All the teams showcased their hard work, and our integrators and administrative staff seamlessly coordinated the day. Everyone’s pride in what was accomplished was evident. The ARPA-H visitors recognized this too and were excited that they could interact and use the system at this stage of development. While we still have an enormous amount of work ahead of us in the next year, we were able to identify what was working well and where we needed to focus.

Q: What challenges has the team encountered and what types of adjustments have you had to make during your first year?

A; Many of our challenges have stemmed from the incredibly fast timelines of our milestones and deliverables. These were all known, but to actually keep that pace was challenging. It is always the unexpected or unanticipated details that keep us busy after hours and on weekends. Because our project team is so large, with such a diverse set of expertise, we realized early how important it was for cross-team communication, collaboration, and coordination, and this is something that we continue to work on and make adjustments to every week. Very few on our team have worked on a single project team this large, and so we each have had to spend more time sharing our expertise with others and weaving all of our knowledge and skills together toward a common focus and goal.

Q: What is most exciting to you about what the team has been able to accomplish and what is ahead?

A: In addition to bringing together all the pieces for our Year-One Prototype, and to demonstrate that it works, it has been most exciting to see everyone adopt this project and goals as their own. I believe everyone knows that if we are successful, we will really change the standard of care for cancer patients undergoing surgery. I believe our team members know that opportunities like this are rare, and everyone is dedicated and committed to seeing this technology become realized and used. I’m most excited about what’s ahead…being able to see all the parts and pieces come together into a single cart and instrument. That will make the platform technology even more real when we can see it in a form that will be ready to be used clinically in the operating room.

Q: What are the next steps for MarginDx?

MarginDx platformA: The next steps for the MarginDx team include physically combining all the components into a single cart, and integrating the computer control software that will coordinate the different operations of the system such as hand-held OCT imaging, AI image analysis, robotic arm positioning, nonlinear imaging, and final AI classification to identify any residual cancer cells and inform the surgeon intraoperatively in real-time.  We have designs to make the imaging probes smaller so they can be used in smaller tumor cavities, and we will have more evaluations by our Human Factors and Clinical Teams to ensure that this system will seamlessly integrate into the procedures and workflow in the operating room. All of this will culminate in our next landmark achievement when our completed system will undergo external review and performance evaluation at the end of the second year.

Q: How will the Mayo Clinic & Illinois Alliance build on this experience in the future?

A: The idea for this project evolved from a long history of steadily building research investigations and feasibility studies over many years, but launched and took off from there because of our conversations, connections, and collaborations through our long-standing Mayo Clinic & Illinois Alliance for Technology-Based Healthcare. Our MarginDx project represents just one example of what is possible for our partnering institutions, and our ability to go after the grander challenges in healthcare that cannot be addressed by single institutions, single groups, or smaller teams alone. I hope more investigators from Illinois and Mayo Clinic begin to conceptualize ideas and envision innovations, solutions, and teams at this scale, knowing that our staff and support teams at IHSI, our Alliance, and Mayo Clinic are ready to help make these impactful solutions possible.