Preliminary Establishment Of A Clinical Efficacy Prediction Model For Ultrasound-guided High-intensity Focused Ultrasound For Uterine Fibroids-single-center Summary

ObjectiveIn order to help patients with uterine fibroids for treatment options,save time and cost,reduce unnecessary waste of resources,through screening of intensive ultrasound focused ultrasound treatment factors,and these influencing factors into the analysis,preliminary establishment of uterine fibroids patients accept hifu treatment before predicting postoperative treatment effect model,provide the basis for clinical decision,is conducive to better Promotion of the technology.MethodsAll the number of cases of high-intensity focused ultrasound in the electronic medical record system of our center were summarized as the modeling population.SPSS 21.0 was employed to conduct univariate and multivariate Logistic regression analysis,and the prediction model of high-intensity focused ultrasound in uterine fibroids was then examined.The model was established by R language.The model was visualized in the form of nomogram and evaluated from differentiation and calibration.The net benefit of the model was assessed by drawing decision curve analysis and the model was internally validated by a ten-fold cross-validation method.ResultsAtotal of 218 Patients were included in this study,and the postoperative response rate of ultrasound-guided intensive focal ultrasound treatment for uterine fibroids was 69.27%.After univariate and multivariate Logistic regression analysis,the type of uterine fibroids,site of uterine fibroids,flow grade of uterine fibroids under ultrasound,and uniformity of contrast distribution at enhancement were selected as the final Predictor,and the nomogram was drawn through R language.The AUC value of the discrimination evaluation index of the model was0.898(AUC>0.5)and a cut-off value of 0.680.The R~2in the regression equation constructed by R language is 0.614(R~2>0.1).The calibration curve highly coincides with the predicted curve in the calibration map of the model.Combining the cut-off values in the ROC curve yields the overall net benefit of the model of 60%–65%.The average AUC obtained by the ten-fold cross-validation method was 0.881(95%CI:0.868,0.894),and the calculated mean Cindex:0.852(95%CI:0.836,0.868)indicates that our model is stable and has good discrimination.ConclusionThis study established a predictive model to predict the clinical efficacy of high-intensity focused ultrasound in the treatment of uterine fibroids,which showed excellent performance in differentiation,calibration and internal verification.This model can provide a reference for Chinese patients when choosing the treatment mode of uterine fibroids.