Can Hounsfield Unit Evaluation of Preoperative CT Scans Predict Surgeon's Preference of Cementless Fixation in Total Knee Arthroplasty?

Abstract

Adequate bone quality is necessary for reliable cementless fixation during total knee arthroplasty (TKA). Hounsfield Units (HUs) can be a useful surrogate of bone density, and therefore a proxy of bone quality. We sought to determine whether HU would correlate with subjective intraoperative assessment of bone quality and hence use of cementless fixation. A retrospective review of patients who underwent a robotic-assisted TKA between June 2022 and June 2024 by a single surgeon was conducted. HU measurements were obtained at three regions of interest (ROI) within the tibial plateau on preoperative axial-CT scans. Patients were divided into cementless or cemented tibial implant fixation groups based upon subjective intraoperative bone quality assessment. Demographic data such as age, sex, and BMI were collected. The final cohort included 630 patients. Cementless fixation was used in 58.3% of cases. Cementless patients had a higher combined overall average HU measurement (152.7 vs. 105.6, p < 0.001). Point-biserial correlation showed moderate inverse correlations between implant fixation type and HU measurements (all p < 0.001). ROC analysis and predictive modeling found that the best performance was derived from the combined model followed by combined demographics alone, and then HU measurements alone (AUC: 0.853 vs. 0.808 vs. 0.757). Standalone demographic resulted in the poorer predictive modeling. Ultimately, cementless fixation was associated with higher HU measurements compared to cemented fixation in all 3 ROI. HU in conjunction with combined demographic data offered the greatest confidence in predicting fixation type and therefore may be utilized to aid in decision-making and surgical planning for TKA.