Associate Professor Walter O’Dell, PhD, and seven UF Radiation Oncology biomedical engineering students will attend the prestigious Biomedical Engineering Society (BMES) annual conference, October 8-12, 2025, at the San Diego Conference Center, San Diego, CA.
The theme of the 2025 event is “Bridging Healthcare Gaps: The Role of Biomedical Engineering”, which reflects the conference’s focus on health and wellness through engineering innovation. Some 5,500 professionals and students across academia, industry, and healthcare are expected to attend.
The UF group will present three posters at the conference:
- Variability in 3D VS 2D Cardiac MRIs (Smith C, Sultz M, Lopez A, Burns M, O’Dell W)
- “We showed several years ago that women with left-sided breast cancer who receive radiation treatment experience a decrease in heart function that is observable within 9 months and that scales as a 5% decrease in left ventricular ejection fraction, or LVEF, for every 1 Gray of mean heart radiation dose exposure, or MH),” explains Dr. O’Dell. “Quizzically, several groups have published that they observed no significant difference in LVEF at these early time points after breast cancer treatment. The difference is that we use non-conventional, high fidelity MR imaging and 3D LV anatomical modeling that I have previously demonstrated achieves far superior accuracy in LVEF measurements than the conventional echocardiography and 2D MRI that is used in the clinic and was used in the other studies. Last year at this conference, we showed that inter-reader variability in estimating LVEF from echocardiography images was around +/- 15%. The changes that we observed with patients was less than this amount, thereby demonstrating that the inherent uncertainty in measuring the LV with echocardiography explains why those other studies failed to see a decrease in LVEF when one actually existed. This year, our first poster looks at whether conventional 2D MRI can be used to detect the changes in LVEF due to RT, by measuring the inter-reader variability with 2D MRI vs 3D MRI. The students show that while 3D is better, 2D MRI is adequate for patients who receive higher MHD.”
- Deformable Image registration using high-order finite element geometry: application to tracking heart contours (O’Dell W, Park M, Tamayo A, Smith C)
- “The problem with our 3D MRI approach is that the large time needed to manually contour a full 3D heart MRI dataset makes it clinically impractical,” says Dr. O’Dell. “Our second poster investigates a novel approach combining finite element modeling (FEM) with deformable image registration (DIR) with the goal to fully-automate the segmentation of the heart from 3D MRI. We showed superiority of FEM-based DIR over conventional B-spline DIR for the task of tracking the LV s the heart contracts, as a first step in developing and test a fully 3D approach.”
- B-TRAC – Breast Tissue Rotation and Compression Apparatus for Calibration (Hernandez-Perez D, Wang A, Holtz D, O’Dell W)
- According to Dr. O’Dell, the third poster “addresses the problem that breast oncologists and surgeons face when trying to mentally match breast anatomy and tumors from a 2D X-ray mammogram to the patient’s 3D MRI. The 3D MRI gives oncologists and surgeons more complete information that is needed for treatment/surgical planning. The mammogram is used to identify suspicious lesions and perform directed biopsies. When there are multiple suspicious lesions it can be challenging to match what is seen on the mammogram where the breast is compressed, to the MRI where the breast is freely suspended. Our team of physics and medical physics majors have designed and created a breast compression device to simulate compression during mammography. While some members are creating and testing the device, others are creating realistic 3D breast phantoms with implanted markers based on the patient’s 3D MRI. And still others are working out the mathematical description of the deformation of the breast using machine learning.”
For more information about BMES, visit the organization’s website.
