Prediction Model Based on Individualized Mechanical Alignment Targets for Outcomes After High Tibial Osteotomy in Knee Osteoarthritis

Ma Xiyao; Dong Kefang; Liu Xiaolan

doi:10.3969/j.issn.1008-5572.2025.10.009

Abstract：Objective To develop a nomogram model based on individualized mechanical alignment targets for predicting the risk of poor outcomes after high tibial osteotomy(HTO)in patients with knee osteoarthritis(KOA).Methods A prospective cohort study enrolled 331 KOA patients scheduled for high tibial osteotomy at the Second Affiliated Hospital of Hunan University of Chinese Medicine between May 2022 and May 2024;328 patients were finally analyzed.One-year postoperative efficacy was evaluated using the hospital for special surgery knee score(HSS≥70=good outcome;HSS＜70=poor outcome).Patients were divided into an effective group(n=266)and a poor-outcome group(n=62).The effective group comprised 136 male and 130 female patients,aged 36-65 years(mean:52.33±8.02).The suboptimal outcome group included 38 male and 24 female patients,aged 33-71 years(mean:53.16±8.41).Baseline characteristics and alignment-correction deviations—mechanical femorotibial angle(mFTA),lateral distal femoral angle(LDFA)and medial proximal tibial angle(MPTA)-were compared.Independent risk factors for poor outcomes were identified by binary logistic regression,and a nomogram was constructed and internally validated by bootstrapping.Model discrimination was quantified with the concordance index(C-index)and receiver operating characteristic(ROC)curve.Results The poor-outcome group showed significantly higher body-mass index(BMI),Western Ontario and McMaster Universities(WOMAC)Osteoarthritis Index scores,proportion of Kellgren-Lawrence(K-L)grade Ⅲ,and rates of alignment deviation(＞3 ° or＜-3 °)in ΔmFTA,ΔLDFA and ΔMPTA(P＜0.05).Logistic regression indicated that high BMI,high WOMAC score,K-L grade Ⅲ,ΔmFTA＞3 ° or＜-3 °,and ΔLDFA＞3 ° or＜-3 ° were independent risk factors for poor outcomes(OR＞1,P＜0.05).The nomogram achieved a C-index of 0.857.ROC analysis yielded an area under the curve(AUC)of 0.857[95％CI(0.810,0.905),P＜0.001]with cut-off 50 20 points,specificity 81.60％,sensitivity 75.80％and Youden index 0.574.A decision-tree model highlighted five key variables-K-L grade,ΔmFTA,ΔLDFA,WOMAC score and BMI-with ΔmFTA as the most important root node(30.10％poor-outcome rate when ΔmFTA was abnormal).Conclusion Individualized alignment deviations(ΔmFTA,ΔLDFA),BMI,K-L grade,and WOMAC score are independent predictors of poor outcomes after HTO.The nomogram model exhibits high clinical predictive value.