Region-specific evaluation of Hounsfield units on elbow CT as a predictor of osteoporosis in patients at risk.

Background

Assessment of osteoporosis is crucial for favorable outcomes after trauma surgery, but a formal DXA scan may not be available preoperatively. The aim of the present study was to assess standardized measurements of Hounsfield units (HU) on CT images of the elbow as a predictor for the presence of osteoporosis.

Methods

Patients who underwent both a computed tomography (CT) scan of the elbow and a dual energy X-ray absorptiometry (DXA) measurement within a period of two years were included. Standardized HU measurements with a diameter of 3 mm each were carried out in defined regions of interest (ROIs) on the humerus, ulna and radius. The lowest T-value of the respective DXA measurement was correlated with the mean value of the HU of the corresponding ROI. Subsequently, the discriminatory capacity of the HU for osteoporosis determination was evaluated by calculating the area under the ROC curve (AUC).

Results

40 patients (72.6 ± 12.2 years; 83 % female) met the inclusion criteria with an average of 1.1 ± 3.8 months between DXA measurement and elbow CT scan. Significant correlations were identified between DXA measurements and HU measurements at all individual ROIs: HCoL (r = 0.43, p = 0.015), HCoM (r = 0.54, p = 0.002), HCap (r = 0.64 , p < 0.001), HTrL ( = 0.52, p = 0.003), HTrM (r = 0.66, p < 0.001), UO (r = 0.49, p = 0.004), UN (r = 0.57, p = 0.001), UC (r = 0.49, p = 0.003), RH (r = 0.46, p = 0.006), ), RN (r = 0.38, p = 0.021), RT (r = 0.51, p = 0.001). The highest discriminatory power for the presence of osteoporosis was detected at the ulnar notch (area under the receiver operating characteristic curve (AUC): 0.88) followed by the area of the humeral capitulum (AUC: 0.86).

Conclusions

CT-based assessment of the elbow joint indicates a promising method for estimating bone quality in at-risk patients. Regional HU measurements may serve as a preliminary reference for osteoporosis and could support treatment planning and implant selection in acute trauma care.