Identifying Risk Factors for Altered Foot and Ankle Bone Mineral Density Using Weight Bearing CT-Derived Hounsfield Units

Research Type: Level 3 - Retrospective cohort study, Case-control study, Meta-analysis of Level 3 studies Introduction/Purpose: Reduced bone mineral density (BMD) has been associated with adverse events following foot and ankle procedures. Poor patient-reported outcomes after ankle fracture fixation and increased periprosthetic fracture rates in ankle arthroplasty are reported. However, there remains a lack of understanding of the risk factors that contribute to changes in BMD. Hounsfield units (HU) derived from computed tomography (CT) have emerged as a valid measure of BMD, and the progressive introduction of Weight Bearing CT (WBCT) clinically has enabled regular access to bone density data under physiologic conditions. The present study aimed to identify patient-related risk factors influencing foot and ankle BMD using WBCT-derived HU data. Methods: A retrospective analysis was conducted of patients who had either a unilateral or bilateral foot/ankle WBCT scan with no significant metal artifact. Potential risk factors studied included patient age, gender, body mass index (BMI), and Vitamin D level. Only patients with a laboratory measurement of 25-Hydroxyvitamin D within one day of their WBCT scan were included to reduce confounding due to Vitamin D supplementation. WBCT datasets were processed using 3D slicer (Version 5.9.0). HU threshold values of -500 to +500 and +500 to +2000 were used to isolate a soft tissue (ST) and bone (B) segment for each dataset. Mean HU values for ST and B were calculated, and a normalized ratio of bone to soft tissue HU values (B:ST) was utilized in final analysis. Univariate analysis was performed, with p < 0.05 considered significant. A LASSO regression analysis with cross-validation was performed to identify the key predictors of B:ST. Results: 22 patients (8 male, 14 female) were included, with a mean ± standard deviation age of 32.0 ± 9.2 years, BMI of 32.6 ± 10.0 kg/m2, and 25-Hydroxyvitamin D of 45.0 ± 25.1 ng/mL. On univariate analysis, 25-Hydroxyvitamin D (p = 0.04, R = 0.44) and female gender (p = 0.03, R = 0.48) were found to be significantly correlated with B:ST. On LASSO regression analysis, an optimal penalty parameter (λ = 0.546) was selected, resulting in a model that retained only female gender (β = 1.84) and 25-Hydroxyvitamin D (β = 0.03) as significant predictors, while the coefficients for age and BMI were shrunk to zero, indicating only female gender and Vitamin D were significant predictors of BMD in this sample. Conclusion: In summary, this study identified female gender and Vitamin D level as significant predictors of foot and ankle BMD when measured via HU on WBCT. This study was limited by small sample size and a relatively young patient cohort. However, this initial pilot study advances our understanding of risk factors affecting BMD in the foot and ankle and highlights the potential for targeted preoperative optimization. This study also demonstrates the utility of clinical access to WBCT datasets in evaluating BMD. Future research in larger cohorts is indicated to validate these results and understand clinical implications.