When surgeons remove an aggressive bone tumor from the pelvis, saving a patient's life can mean sacrificing a large portion of the spine and hip. Putting that fragile structure back together in a way that will let the patient sit, stand and walk safely is one of the most difficult challenges in orthopedic oncology and requires a team of experts with highly specialized skills.
Now, a new collaboration between Rice University engineers and physicians at the University of Texas MD Anderson Cancer Center is giving surgeons a powerful new way to plan those reconstructions before they ever step into the operating room. The research was recently published in the Journal of the Mechanical Behavior of Biomedical Materials .
"These cases are notoriously complex," said Raudel Avila , corresponding author of the study and an assistant professor of mechanical engineering at Rice. "After a hemipelvectomy, you're left with a big gap where the tumor and part of the pelvis used to be. Surgeons have to decide what bone to graft, what implants to use and how to assemble everything, so the reconstruction will withstand everyday loading. Until now, there hasn't been a clear engineering road map for those decisions."
For Justin Bird , orthopedic oncologic surgeon at MD Anderson and co-author on the study, the stakes are personal and immediate.
"When I'm in the operating room, I'm making decisions that will affect a patient's mobility for the rest of their life," Bird said. "Having a patient-specific model that shows how different grafts and implants will behave under real-world loads gives us a tremendous advantage in choosing the safest, most durable reconstruction."
Working from de-identified CT and MRI scans of a young adult patient with a complex spino-pelvic osteosarcoma, the Rice engineering team built highly detailed 3D computer models of the patient's pelvis before and after surgery. Using finite element analysis — a computer modeling approach that breaks complex structures into small pieces to predict stress and deformation — the team simulated the mechanical loads the reconstructed pelvis would experience during early recovery, when the grafts and implants are most vulnerable to failure.
"We essentially created a digital model of the reconstructed pelvis," said first author Ritika Menghani , a mechanical engineering postdoctoral fellow in Avila's lab at Rice. "That let us test different bone grafts and implant materials through a computer model, which can in turn inform surgical decisions and hopefully improve the mechanical outcomes of pelvic reconstructions, potentially leading to better patient recovery and reduced risk of complications."
One of the clearest findings: Bigger really is better when it comes to bone grafts. By comparing geometry, material and placement of femoral, tibial and fibular grafts, the models showed that grafts with larger cross-sectional area, like the femur and tibia, experienced much lower stresses and smaller deformations than thinner bones such as the fibula.
"In our simulations, the femoral graft consistently carried the load with the lowest stress, and the tibia wasn't far behind," Menghani said. "The fibula saw almost three times as much stress as the femur. That kind of information can give surgeons more confidence when they're choosing which bone to harvest for a reconstruction."
The team also explored what happens when different implant materials — titanium, stainless steel, magnesium and polymer-based options such as PEEK (polyether ether ketone) — are used to anchor the graft to the pelvis and spine. Stiffer metals provided the best protection against screw breakage in high-load scenarios, while polymer implants reduced "stress shielding," a phenomenon where very stiff implants take so much load that the surrounding bone stops remodeling and weakens over time. The models make explicit trade-offs surgeons often face implicitly: whether to prioritize immediate mechanical stability or long-term bone remodeling.
"It's really a trade-off," said Rice undergraduate Amy Li, a member of Avila's lab and co-author on the study. "Titanium and stainless steel are great if you're worried about things breaking under heavy loads, but they can shield the bone from stress. Polymers like PEEK are gentler on the bone and may support long-term healing, but they can't tolerate the same extreme forces. Our models help quantify those trade-offs, so surgical teams can pick what makes sense for each patient."
To make the approach more accessible for busy clinicians, the researchers didn't stop at full 3D simulations. They also developed a simplified 2D model that captures the key mechanics of the reconstruction and can be quickly adapted to different geometries and graft choices.
"The 3D models are the gold standard, but they're time-intensive," Avila said. "The reduced 2D model is like a quick sketch you can run in seconds. It lets you compare different graft sizes, orientations and materials without losing the essential physics, and that's something we envision integrating into presurgical planning."
Although the study focused on a single, highly complex hemipelvectomy case, the underlying pipeline — going from clinical images to validated mechanical models — is designed to be flexible. The team says it could be extended to other spino-pelvic tumors, spinal reconstructions, joint replacements and even reconstructive surgeries for other types of cancer.
"What excites us most is the idea of simulation-guided surgery," said Valerae O. Lewis , chair of orthopedic oncology at MD Anderson and co-author of the study. "If patient-specific models can be built from the scans surgeons are already taking, we can stress-test different reconstruction strategies on the computer first, instead of relying only on past experience and trial and error."
Ultimately, the researchers say they hope their work will translate into fewer complications and a better quality of life for patients facing these operations.
"These are people who want to go back to school, to work, to their families," Avila said. "By bridging engineering and medicine, we're giving surgical teams another way to predict what will happen after they close the incision — and that has real potential to improve outcomes for patients."
