Opportunistic CT screening predicts individuals at risk of major osteoporotic fracture

• Published in: Osteoporosis international 2021-08, Vol.32 (8), p.1639-1649
• Main Authors: Michalski, A.S., Besler, B.A., Burt, L.A., Boyd, S.K.
• Format: Article
• Language: English
• Subjects: Abdomen; Absorptiometry; Bone mineral density; Bone strength; Computed tomography; Dual energy X-ray absorptiometry; Endocrinology; Energy; Femur; Fractures; Medical imaging; Medicine; Medicine & Public Health; Original Article; Orthopedics; Osteoporosis; Predictions; Rheumatology; Spine; Vertebrae; Women
• ISSN: 0937-941X; 1433-2965
• DOI: 10.1007/s00198-021-05863-0

Summary Millions of CT scans are performed annually and could be also used to opportunistically assess musculoskeletal health; however, it is unknown how well this secondary assessment relates to osteoporotic fracture. This study demonstrates that opportunistic CT screening is a promising tool to predict individuals with previous osteoporotic fracture. Introduction Opportunistic computed tomography (oCT) screening for osteoporosis and fracture risk determination complements current dual X-ray absorptiometry (DXA) diagnosis. This study determined major osteoporotic fracture prediction by oCT at the spine and hip from abdominal CT scans. Methods Initial 1158 clinical abdominal CT scans were identified from administrative databases and were the basis to generate a cohort of 490 men and women with suitable abdominal CT scans. Participant CT scans met the following criteria: over 50 years of age, the scan had no image artifacts, and the field-of-view included the L4 vertebra and proximal femur. A total of 123 participants were identified as having previously suffered a fracture within 5 years of CT scan date. Fracture cause was identified from clinical data and used to create a low-energy fracture sub-cohort. At each skeletal site, bone mineral density (BMD) and finite element (FE)-estimated bone strength were determined. Logistic regression predicted fracture and receiver-operator characteristic curves analyzed prediction capabilities. Results In participants with a fracture, low-energy fractures occurred in 88% of women and 79% of men. Fracture prediction by combining both BMD and FE-estimated bone strength was not statistically different than using either BMD or FE-estimated bone strength alone. Predicting low-energy fractures in women determined the greatest AUC of 0.710 by using both BMD and FE-estimated bone strength. Conclusions oCT screening using abdominal CT scans is effective at predicting individuals with previous fracture at major osteoporotic sites and offers a promising screening tool for skeletal health assessment.