IBIIS-AIMI Seminar: Beyond the Lowest T-Score: Rethinking Bone Density Assessment Through Multivariate and Opportunistic Imaging Approaches - Ronnie Serbo, MD, PhD

Abstract: Osteoporosis affects millions of individuals worldwide and remains a major contributor to fracture-related morbidity, mortality, loss of independence, and healthcare expenditures. Despite advances in imaging and fracture risk prediction, contemporary osteoporosis diagnosis continues to rely primarily on a single measurement—the lowest T-score obtained from lumbar spine, femoral neck, or total hip dual-energy X-ray absorptiometry (DXA). This approach implicitly discards substantial information contained within the remaining skeletal measurements and may incompletely characterize an individual's overall skeletal fragility.

We will discuss osteoporosis diagnosis from the perspective of a statistician and imaging scientist. Using principles from multivariate statistics, we will discuss the correlation structure among DXA-derived bone mineral density measurements and demonstrate how integrating information across multiple skeletal sites may alter diagnostic classification and improve risk stratification. Findings from our work suggest that lumbar spine, femoral neck, and total hip measurements provide complementary information regarding skeletal health and that a multivariate framework may identify individuals at elevated fracture risk who would not otherwise be recognized using current approaches.

The talk will also explore emerging opportunities for opportunistic osteoporosis screening using routinely acquired computed tomography (CT) examinations. Advances in quantitative imaging now permit extraction of clinically meaningful bone density information from scans originally obtained for unrelated indications, creating new opportunities for large-scale population screening without additional imaging burden.

About: Ronnie A. Sebro, MD, PhD, is a tenured Professor of Radiology, Epidemiology, and Imaging Physics at The University of Texas MD Anderson Cancer Center. A board-certified musculoskeletal radiologist, statistical geneticist, and biostatistician. He specializes in artificial intelligence applications in medical imaging, opportunistic screening (e.g., osteoporosis via CT/radiographs), radiomics, and AI validation in oncology and musculoskeletal disorders. Previously, he held faculty positions at Mayo Clinic and the University of Pennsylvania. Dr. Sebro serves as Deputy Editor for Radiology: Artificial Intelligence and Associate Editor for Journal of Imaging Informatics in Medicine.

Attendance is open to the Stanford and AIMI affiliate community. Please contact aimicenter@stanford.edu for the Zoom link if you would like to attend virtually.