Return To Abstract Listing

Scientific Session 15 — Musculoskeletal - CT

Wednesday, May 3, 2017

Abstracts 2331-2956

2331. Predicting Future Hip Fractures at Abdominal CT Using Opportunistic Osteoporosis Screening Measures

Lee S*,  Pickhardt P. University of Wisconsin School of Medicine and Public Health, Madison, WI

Address correspondence to S. Lee (

Objective: Fragility fractures of the hip are one of the most devastating consequences of osteoporosis and osteopenia. Treatment of low bone mineral density reduces fracture risk, but this condition remains underdiagnosed. Measurement of the vertebral trabecular attenuation at CT (e.g., at L1) and asynchronous quantitative CT (QCT) of the hip have been explored as a means of opportunistic CT screening for osteopenia and osteoporosis and can be performed on abdominal CT scans obtained for any indication. The purpose of this case-control study was to perform these two measurements in patients with hip fracture on CT performed before fracture and compare results with age- and sex-matched control subjects to determine if they are associated with future hip fracture risk.

Materials and Methods: Consecutive hip fracture cases over a 10-year period (February 2006 to September 2015) were reviewed to identify patients with abdominopelvic CT performed within 6 years (mean interval to fracture, 24.8 months) before their fracture date that included the lumbar spine, hips, or both. Studies not performed at 120 kV were excluded. The final hip fracture cohort was composed of 204 patients (130 women, 74 men; mean age, 74.3 years). An age- and sex-matched control group without hip fracture was randomly selected from a screening CT colonography population in a 1:1 ratio (n = 204), yielding a total sample size of 408. L1 attenuation was measured using an ROI placed within the trabecular bone and femoral neck. T scores were calculated for both hips using asynchronous QCT software (CTXA, QCTPro, Mindways). A conditional logistic regression model was used to determine the relationship between lowest hip T score, L1 attenuation, prevalent vertebral fractures, and hip fracture outcome.

Results: Mean L1 attenuation values and femoral neck T scores were 98.5 ± 36.8 HU and –1.6 ± 1.0 for the hip fracture cohort and 129.7 ± 44.9 HU and –0.2 ± 1.5 for control subjects, respectively (p < 0.0001). There was moderate or severe vertebral compression fracture at CT in 18.1% of the cases compared with 2.5% of control subjects (OR, 7.4; p < 0.001). The sensitivity and specificity of L1 attenuation in distinguishing patients with hip fracture from control subjects at an optimal threshold of 107 HU were 61.8% and 67.2%, respectively (ROC AUC = 0.70). Sensitivity and specificity values for CTXA at an optimal threshold of –1.5 were 61.1% and 81.7%, respectively (ROC = 0.775

Conclusion: L1 trabecular attenuation, CTXA femoral neck T scores, and prevalent vertebral fractures obtained as opportunistic measures at abdominal CT can provide valuable information regarding patients at increased risk for future hip fragility fractures.