中医正骨2026，Vol.38Issue(1)：23-28,34,7.
基于铜死亡标志物铁氧化还原蛋白1和脂酰转移酶1的绝经后骨质疏松症风险预测模型构建
Construction of a risk prediction model for postmenopausal osteoporosis based on cuproptosis markers ferre-doxin 1 and lipoyltransferase 1
摘要
Abstract
Objective:To construct a risk prediction model for postmenopausal osteoporosis(PMOP)based on the cuproptosis markers ferredoxin 1(FDX1)and lipoyltransferase 1(LIPT1).Methods:Postmenopausal women who visited our hospital between October 2023 and April 2024 were selected as the study subjects.Bone mineral density(BMD)of L,-L4 was measured using dual-energy X-ray absorptiome-try;peripheral venous blood was collected,and serum levels of β-isomerized carboxy-terminal telopeptide collagen type Ⅰ(β-CTX),type Ⅰprocollagen amino-terminal peptide(P ⅠNP),and osteocalcin(OC)were measured using electrochemiluminescence immunoassay,while ser-um levels of FDX1 and LIPT1 were detected using ELISA.Based on lumbar spine BMD T-scores,the subjects were divided into an osteopo-rosis group(T-scores≤-2.5)and a non-osteoporosis group(T-scores＞-2.5).The differences in serum FDX1,LIPT1,β-CTX,P ⅠNP,OC levels,and BMD-the six variables related to the PMOP risk prediction model-were compared between the two groups.Spearman cor-relation analysis was used to analyze the correlation between serum FDX1 and LIPT1 levels and BMD as well as serum β-CTX,P Ⅰ NP,and OC levels.Random forest algorithm and receiver operating characteristic(ROC)curves were used to evaluate the importance and capability of the above variables in predicting PMOP risk,and a PMOP risk prediction model was constructed based on these variables.Results:①Comparison results of variables related to the PMOP risk prediction model.A total of 90 subjects were included,including 61 cases in the osteoporosis group and 29 cases in the non-osteoporosis group.The differences in serum FDX1,LIPT1,β-CTX,P Ⅰ NP,OC levels and BMD between the two groups were statistically significant(Z=-6.741,P=0.000;Z=-7.420,P=0.000;Z=-3.333,P=0.001;Z=-2.456,P=0.014;Z=-2.931,P=0.003;t=12.355,P=0.000).②Results of correlation analysis.Serum LIPT1 level was posi-tively correlated with BMD(rs=0.579,P=0.000),negatively correlated with serum β-CTX level(rs=-0.257,P=0.015),and not cor-related with serum P Ⅰ NP and OC levels(rs=-0.140,P=0.188;rs=-0.192,P=0.070).Serum FDX1 level was negatively correlated with BMD(rs=-0.466,P=0.000),positively correlated with serum β-CTX level(rs=0.239,P=0.024),and not correlated with serum P Ⅰ NP and OC levels(rs=0.155,P=0.144;rs=0.198,P=0.062).Serum LIPT1 level was negatively correlated with serum FDX1 level(r,=-0.798,P=0.000).③Results of importance and capability evaluation of relevant variables in predicting PMOP risk.Random forest analysis results showed that serum LIPT1 level,BMD,and serum FDX1 level were the top 3 variables in terms of importance for predicting PMOP risk.ROC curve analysis results showed that BMD and serum FDX1,LIPT1,β-CTX,P Ⅰ NP,and OC levels all had certain diagnostic value for PMOP(AUC=0.996,AUC=0.941,AUC=0.985,AUC=0.718,AUC=0.661,AUC=0.692),among which the diagnostic val-ue of BMD and serum FDX1 and LIPT1 levels was higher.④Results of PMOP risk prediction model construction.A PMOP risk nomogram prediction model was constructed based on BMD and serum FDX1,LIPT1,β-CTX,P Ⅰ NP,and OC levels.The calibration curve showed that the calibration curve highly overlapped with the ideal diagonal.The decision curve showed that within the threshold probability range,the clinical net benefit obtained by decision-making using this model was higher than the two extreme strategies of intervening on all patients or not intervening on any patient.In the clinical impact curve,the actual number of events curve closely followed the high-risk number curve.Conclusion:Cuproptosis markers FDX1 and LIPT1 are closely related to the pathogenesis of PMOP.The PMOP risk prediction model con-structed by combining cuproptosis markers FDX1,LIPT1 with BMD and bone turnover markers(β-CTX,P Ⅰ NP,OC)demonstrates good di-agnostic efficacy and accuracy.关键词
骨质疏松,绝经后/细胞死亡/铁氧化还原蛋白1/脂酰转移酶1/列线图表/风险/预测/随机森林/ROC曲线Key words
osteoporosis,postmenopausal/cell death/ferredoxin 1/lipoyltransferase 1/nomograms/risk/forecasting/random forest/ROC curve引用本文复制引用
曾韵杰,刘树华,李小韵,陈桐莹,黄宏兴,杨贻富,杨东升,付赛,万雷..基于铜死亡标志物铁氧化还原蛋白1和脂酰转移酶1的绝经后骨质疏松症风险预测模型构建[J].中医正骨,2026,38(1):23-28,34,7.基金项目
国家自然科学基金项目(82174395) （82174395）
广东省自然科学基金项目(2022A1515012067) （2022A1515012067）
广东省普通高校特色创新项目(自然科学)(2022KTSCX026) （自然科学）
广东省科技计划项目(2024A1111120018) （2024A1111120018）
