BACKGROUND:Proximal humeral fracture in older adults is one of the three major osteoporotic fractures.Anatomic locking plate fixation is the first choice for most scholars to treat difficult-to-reduce and complex fracture types.However,the probability of reduction failure after the operation is high,which seriously affects patients'quality of life. OBJECTIVE:To investigate the correlation between deltoid tuberosity index and postoperative reduction failure of proximal humeral fractures in the elderly,analyze and filter preoperative independent risk factors for reduction failure of proximal humeral fractures in the elderly,and construct and verify the effectiveness of a clinical prediction model. METHODS:The clinical data of 153 elderly patients with proximal humeral fractures who met the diagnosis and inclusion criteria and received open reduction and locking plate surgery in Foshan Hospital of TCM from June 2012 to June 2021 were collected.The patients were divided into the reduction failure subgroup and the reduction maintenance subgroup.The independent risk factors were selected by multivariate Logistic regression analysis,and the nomogram was constructed by R language.After 1000 times of resampling by Bootstrap method,the Hosmer-Lemeshow goodness of fit correlation test,receiver operating characteristic curve,calibration curve,clinical decision,and influence curve were plotted to evaluate its goodness of fit,discrimination,calibration ability,and clinical application value.Fifty-five elderly patients with proximal humeral fractures from June 2013 to August 2021 were selected as the model's external validation group to evaluate the prediction model's stability and accuracy. RESULTS AND CONCLUSION:(1)Of the 153 patients in the training group,44 patients met reduction failure after internal plate fixation.The prevalence of postoperative reduction failure was 28.8%.Multivariate Logistic regression analysis identified that deltoid tuberosity index[OR=9.782,95%CI(3.798,25.194)],varus displacement[OR=4.209,95%CI(1.472,12.031)],and medial metaphyseal comminution[OR=4.278,95%CI(1.670,10.959)]were independent risk factors for postoperative reduction failure of proximal humeral fractures in older adults(P<0.05).(2)A nomogram based on independent risk factors was then constructed.The Hosmer-Lemeshow test results for the model of the training group showed that χ2=0.812(P=0.976)and area under curve=0.830[95%CI(0.762,0.898)].The calibration plot results showed that the model's predicted risk was in good agreement with the actual risk.The decision and clinical influence curves showed good clinical applicability.(3)In the validation group,the accuracy rate in practical applications was 86%,area under curve=0.902[95%CI(0.819,0.985)].(4)It is concluded that deltoid tuberosity index<1.44,medial metaphyseal comminution,and varus displacement were independent risk factors for reduction failure.(5)The internal and external validation of the risk prediction model demonstrated high discrimination,accuracy,and clinical applicability could be used to individually predict and screen the high-risk population of postoperative reduction failure of proximal humeral fractures in the elderly.The predicted number of patients at high risk is highly matched to the actual number of patients who occur when the model's threshold risk probability is above 65%,and clinicians should use targeted treatment.

BACKGROUND:Internal fixation and open reduction with locking plate is the main treatment for proximal humeral fractures with medial column instability.However,reduction failure is one of the main postoperative complications,and accurate risk factor assessment is beneficial for screening high-risk patients and clinical decision selection. OBJECTIVE:To construct four types of prediction models by different machine learning algorithms,compare the optimal model to analyze and sort the risk variables according to their weight scores on the impact of outcome,and explore their significance in guiding clinical diagnosis and treatment. METHODS:262 patients with proximal humeral fractures with medial column instability,aged(60.6±10.2)years,admitted to Foshan Hospital of Traditional Chinese Medicine between June 2012 and June 2022 were included.All patients underwent open reduction with locking plate surgery.According to the occurrence of reduction failure at 5-month follow-up,the patients were divided into a reduction failure group(n=64)and a reduction maintenance group(n=198).Clinical data of patients were collected,and model variables and their classification were determined.The data set was randomly divided into a training set and a test set according to a 7:3 ratio,and the optimal hyperparameters were obtained in the training set according to a 5-fold cross-over test.Four machine learning prediction models of logistic regression,random forest,support vector machine,and XGBoost were constructed,and the performance of different algorithms was observed in the test set using AUC,correctness,sensitivity,specificity,and F1 scores,so as to comprehensively evaluate the prediction performance of the models.The best-performing model was evaluated using SHAP to assess important risk variables and to evaluate its clinical guidance implications. RESULTS AND CONCLUSION:(1)There were significant differences between the two groups in deltoid tuberosity index,fracture type,fracture end with varus deformity before operation,fragment length of inferior metaphyseal of humerus,postoperative reduction,cortical support of medial column of proximal humerus,and insertion of calcar screw(P<0.05).(2)The best-combined performance of the four machine models was XGBoost.The AUC,accuracy,and F1 scores were 0.885,0.885,and 0.743,respectively;followed by random forest and support vector machine,with both models performing at approximately equal levels.Logistic regression had the worst combined performance.The SHAP interpretation tool was used in the optimal model and results showed that deltoid tuberosity index,medial humeral column cortical support,fracture type,fracture reduction quality,and the status of the calcar screw were important influencing fators for postoperative fracture reduction failure.(3)The accuracy of using machine learning to analyze clinical problems is superior to that of traditional logistic regression analysis methods.When dealing with high-dimensional data,the machine learning approach can solve multivariate interaction and covariance problems well.The SHAP interpretation tool can not only clarify the importance of individual variables but also obtain detailed information on the impact of dummy variables in each variable on the outcome.