CT-Derived Hounsfield Units Correlate with Dorr Classification and Objectively Identify Dorr Type C Morphology in MAKO-Assisted Total Hip Arthroplasty

<title>Abstract</title> <p> Background Radiographic Dorr classification used before total hip arthroplasty (THA) is subjective. Preoperative computed tomography (CT) scans for robotic THA allow Hounsfield unit (HU) quantification, providing an objective measure of bone quality. Methods We retrospectively reviewed 578 patients undergoing MAKO-assisted primary THA (2019–2024); 489 met inclusion criteria. Femoral neck HU and an average composite HU (L4, femoral neck, intertrochanteric) were compared across Dorr types. ROC and regression analyses identified thresholds predictive of high-risk Dorr Type C morphology. Results Femoral neck HU decreased stepwise across Dorr A (171.9), B (124.5), and C (77.9) hips ( <italic>P</italic>  &lt; 0.001), as did average HU (186.5, 136.6, and 92.2, respectively; <italic>P</italic>  &lt; 0.001). Femoral neck HU predicted Dorr C morphology with an AUC of 0.81 (cutoff ≤ 87 HU; 62.4% sensitivity, 84.0% specificity). Average HU demonstrated stronger discrimination with an AUC of 0.88 (cutoff ≤ 117.6 HU; 82.2% sensitivity, 80.2% specificity). On multivariable regression adjusting for age, BMI, and sex, lower average composite HU remained independently associated with higher Dorr grade (OR 1.43 per 10 HU decrease, <italic>P</italic>  &lt; 0.001) and maintained strong predictive performance for Dorr C morphology (adjusted AUC 0.879). Conclusion CT-derived HU measurements correlate strongly and stepwise with Dorr classification. Average composite HU outperformed isolated femoral neck HU for identifying Dorr Type C morphology. A threshold of ≤ 117.6 HU serves as a robust, covariate-independent screening tool to objectify preoperative planning and guide stem selection. </p>