Opportunistic screening for osteoporosis: validation study for L1 bone density measurements using contrast-enhanced...

To retrospectively validate the diagnostic power of attenuation values on chest and abdomen/pelvis CECTs, together and separately, compared with dual-energy X-ray absorptiometry (DEXA)-determined osteoporosis diagnoses, and to determine thresholds for accurate osteoporosis diagnosis.

Subjects were identified using the electronic health record. Included patients received DEXA and CECT scans within 60 days of each other. Patients were excluded if taking osteoporosis medication, undergoing dialysis, receiving hormone or cancer therapy, had a history of cancer, osseous metastases, fractures, or compressions. Minimum, mean, and maximum CECT attenuation values of L1 trabecular bone axial cross-sections were measured by a non-physician in Hounsfield units (HUs) using an elliptical region of interest (ROI) tool. DEXA diagnoses were dichotomized as positive (osteoporosis) or negative (osteopenia/normal). The area under the receiver-operator characteristic curves (AUCs) were compared to identify ideal CECT attenuation thresholds.

Two hundred nineteen subjects (mean age 66 ± 0.6 [range 35-92]; 196 (89%) females and 23 (11%) males) were included for analysis. Thirty-one (14%) subjects were positive and 188 (86%) were negative for osteoporosis. Minimum, mean, and maximum combined chest and abdomen/pelvis attenuation values demonstrated AUCs of 0.75 (95% CI 0.67-0.84), 0.931 (95% CI 0.88-0.99), and 0.82 (95% CI 0.73-0.90). The optimal mean attenuation threshold for osteoporosis diagnosis was 120 HU (84% sensitive, 90% specific). There was no statistical difference in diagnostic power between mean attenuation values of chest and abdomen/pelvis CECTs.

CECT mean attenuation values of either chest or abdomen/pelvis CECTs could be used as appropriate thresholds in screening for osteoporosis.