FAU/Baptist Health AI Spine Model Could Transform Back Pain Treatment
Nearly 3 in 10 adults in the U.S. have experienced lower back pain in any three-month period, making it the most common musculoskeletal pain.
Nearly 3 in 10 adults in the United States have experienced lower back pain in any three-month period, making it the most common musculoskeletal pain. Back pain remains one of the leading causes of disability worldwide, affecting millions and often leading to chronic discomfort, missed work and invasive procedures.
Researchers and clinicians are increasingly turning to lumbar spine modeling, which bridges engineering and medicine, creating a virtual, patient-specific model of the lower back. This technology simulates how the spine moves, where mechanical stress builds up, and what might be causing pain or dysfunction.
These detailed models are used to plan surgeries, evaluate spinal implants and develop personalized treatment strategies tailored to each patient’s anatomy. Despite its promise, current lumbar spine modeling is slow, manual and demands specialized expertise, limiting scalability and personalization. This hinders clinical application and results in inconsistent outcomes.
Researchers from the College of Engineering and Computer Science at Florida Atlantic University and the Marcus Neuroscience Institute at , part of Baptist Health, have reached a major milestone in lumbar spine modeling by integrating artificial intelligence with biomechanics to transform spine diagnostics and personalized treatment planning.
They are the first to create a fully automated finite element analysis pipeline specifically for lumbar spine modeling. Their breakthrough involves integrating deep learning tools like nnUNet and MONAI with biomechanical simulators such as GIBBON and FEBio.
Results of the study, published in the journal , show that this new approach reduced lumbar spine model preparation time by 97.9% – from more than 24 hours to just 30 minutes and 49 seconds – without compromising biomechanical accuracy. The fully automated pipeline enables rapid, patient-specific simulations that support preoperative planning, spinal implant optimization and early detection of degenerative spine conditions.
Tests showed that the virtual spine reacted just like a real one, with realistic disc movement, ligament tension and pressure in the back of the spine during bending and stretching. Because the system runs with very little manual work, it’s much faster and more consistent than traditional methods, making it a valuable tool for doctors and researchers alike.
“What sets our approach apart is its ability to automatically convert standard medical images like CT or MRI scans into highly accurate, patient-specific spine models,” said Maohua Lin, Ph.D., corresponding author and a research assistant professor, FAU Department of Biomedical Engineering. “Traditional manual methods require complex geometry processing, meshing and finite element simulation setup, making them not only time-intensive but also highly dependent on the operator’s expertise. Our automated pipeline significantly reduces the time required, cutting what once took several hours or even days down to just minutes.”
For the study, researchers used advanced AI to automatically identify important parts of the spine – like bones and discs – from medical scans. These were then turned into smooth 3D models that included bones, cartilage and ligaments. They mapped where the ligaments attach and shaped the cartilage based on common patterns. Lastly, researchers ran computer simulations to see how the spine responds to movements like bending and twisting, helping them understand where stress builds up and how the spine moves in real life.
“Beyond advancing research, automated lumbar spine modeling plays a critical role in preoperative planning,” said , corresponding author and chief of neurosurgery at Marcus Neuroscience Institute. “This technology quickly generates patient-specific models to predict mechanical complications, optimize implant design and reduce surgical risks. By removing manual steps, it also improves speed and consistency, helping clinicians make more informed decisions.”
This research builds upon previous work by the research team published in leading journals including Artificial Intelligence Review and the North American Spine Society Journal, investigating related AI-driven biomechanical modeling techniques.
“This groundbreaking work exemplifies the game-changing power of uniting engineering and medicine to address complex health care challenges,” said Stella Batalama, Ph.D., dean of the FAU College of Engineering and Computer Science. “FAU and Baptist Health researchers are not only pushing the boundaries of innovation, they are also delivering real-world solutions that can improve patient outcomes and redefine spine care.”
Study co-authors are Mohsen Ahmadi, a Ph.D. student in the FAU Department of Electrical Engineering and Computer Science; Xuanzong Zhang, an American Heritage High School student; Yufei Tang, Ph.D., an associate professor, FAU Department of Electrical Engineering and Computer Science and FAU Sensing Institute fellow; Erik Engeberg, Ph.D., a professor, FAU Department of Biomedical Engineering and Department of Ocean and Mechanical Engineering, a member of the FAU within the Charles E. Schmidt College of Science, and a member of the FAU Stiles-Nicholson Brain Institute; and Javad Hashemi, Ph.D., inaugural chair and professor of the Department of Biomedical Engineering and associate dean for research, FAU College of Engineering and Computer Science.
This research was supported by Boca Raton Regional Hospital, part of Baptist Health, the Helene and Stephen Weicholz Foundation, the National Science Foundation, pilot grants from the FAU College of Engineering and Computer Science, the FAU Stiles-Nicholson Brain Institute, the FAU Center for Smart Health, and the FAU Sensing Institute.
-FAU-
Tags: medicine | AI | brain institute | research | ֱ | engineering | faculty and staff | technology