Preoperative prediction of bleeding complications in percutaneous nephrolithotomy using a machine learning model based on CT radiomics and clinical variables

Abstract

Background and objective Radiomics and machine learning play a significant role in clinical medical research, particularly in the development of prediction models.This study aims to utilize radiomic features and clinical variables in combination with machine learning to predict the risk of postoperative bleeding after percutaneous nephrolithotomy (PCNL). Materials and Methods A retrospective study analyzed 151 patients who had PCNL at the Second Affiliated Hospital of Nanchang University.Clinical variables linked to postoperative bleeding were identified using univariate analysis,and radiomic features were screened using the least absolute shrinkage and selection operator algorithm(lasso regression).Logistic regression,Random Forest(RF),and Support Vector Machine(SVM) were then used to develop prediction models based on the correlated clinical variables and radiomic features.The predictive accuracy of these models was assessed through identification and calibration. Results The postoperative statistics revealed that the postoperative bleeding rate was 31.1%(n=47),the blood transfusion rate was 1.42%(n=3),and the final probability of requiring vascular embolization was 0.94%(n=2).The accuracy rates for predicting postoperative bleeding in patients with PCNL using logistic regression,RF and SVM algorithms were 75.6%,75.6%,and 71.1% respectively.The corresponding area under the curve AUC(95% CI) were 0.76(0.72-0.81),0.74(0.69-0.79) and 0.63(0.54-0.68).The top four prediction importance scores in logistic regression and RF algorithms were wavelet-HLH_glrlm_ShortRunLowGrayLevelEmphasis,wavelet-HLH_glrlm_LowGrayLevelRunEmphasis,stone shape,operation time and stone shape,stone diameter,operation time,Wavelet-HLH_glrlm_ShortRunLowGrayLevelEmphasis. Conclusion The logistic regression model demonstrated the highest efficiency in predicting postoperative bleeding in PCNL.Our study successfully developed an effective machine learning model that can assist urological surgeons in making appropriate treatment decisions for predicting postoperative bleeding in PCNL.