Random Forest Algorithm for Predicting Stability of Dorsal Intra-articular Bone Mass After Simple Palmar Plate Internal Fixation and Model Construction and Validation
Objective To analyse the factors influencing the stability of the dorsal intra-articular bone mass in patients with distal radius fractures who underwent internal fixation with a simple palmar plate based on the random forest algorithm,and to construct a random forest model to predict the risk of stability of the dorsal intra-articular bone mass after surgery.Methods A convenient sampling method was used to select 320 cases of distal radius fracture treated with simple volar plate internal fixation in Henan Coal Industry(Group)Co.,Ltd.Central Hospital from January 2017 to January 2022 as the research objects.The stability of the bone mass in the dorsal joint after surgery was analyzed,and the patients were divided into stable group and unstable group according to whether the bone mass was displaced or not.General data,diagnosis and treatment information related to fracture and operation were collected from the two groups.Univariate and multivariate logistic regression analysis were used to analyze the factors affecting the stability of the bone mass in the dorsal joint after surgery.The stochastic forest model was established based on logistic regression analysis,and the predictive value of the model to the stability of the bone mass in the postoperative dorsal joint was analyzed.Results The incidence of bone mass displacement in dorsal joint was 27.22％within 12 months after operation.C3 fracture,locking compression plate internal fixation,total displacement distance of bone mass before surgery,transverse width of bone mass,transverse and longitudinal ratio of bone mass and bone mass volume were the influencing factors of postoperative stability of bone mass in the dorsal joint(P＜0.05).Significance analysis showed that the top 4 variables affecting the stability of postoperative dorsal joint bone mass from high to low in importance were the total displacement distance of preoperative bone mass,internal fixation with locking compression plate,bone mass volume,and transverse width of bone mass.The sensitivity and specificity of random forest model to predict the postoperative stability risk of bone mass in the dorsal joint were 91.02％and 82.36％.Conclusion The prediction model based on random forest algorithm has a certain value in predicting the stability of internal bone mass of dorsal joint in simple volar plate internal fixation,which is helpful for early identification of high-risk patients,and has a guiding value for the selection of dorsal plate.
万方数据
维普
