The construction of a nomogram based on ultrasound signs combined with clinical parameters can effectively predict testicular torsion
XIAO Kunkun
WU Guozhu
Abstract：Objective To investigate the predictive potential of building a nomogram model for testicular torsion(TT)by combining clinical characteristics and ultrasound signals.Methods The clinical data of 272 patients with scrotal pain in Inner Mongolia People's Hospital from March 2011 to April 2024 were retrospectively analyzed.According to the 7:3 random machine,they were divided into training set(n=190)and validation set(n=82).The training set consisted of 97 non-TT patients and 93 TT patients,based on the findings of surgical exploration and follow-up.The validation set consisted of 41 non-TT patients and 41 TT patients.The clinical information and several ultrasonic characteristics of the patients were compared using the training set.R software was used to create the risk prediction model of TT visual nomogram in scrotal discomfort after univariate analysis and multivariate logistic regression were used to check for independent influencing factors.The model's performance was examined and assessed using the ROC curve,calibration,and decision curves.Results Age,internal testicular parenchymal echogenicity,testicular blood flow signal grade,Spermatic cord whirlpool,and neutrophil/lymphocyte ratio were all independent risk variables for TT prediction,according to the results of multifactorial logistic regression analysis(P<0.05).A nomogram prediction model was established using the aforementioned criteria.Both the validation set and the model training set had corresponding area under the curves of 0.980 and 0.995.The calibration curve demonstrated that the model's TT prediction probability in both the training and validation sets tended to match the likelihood of actual occurrence.In the training and validation sets,the average absolute error between the expected and actual outcomes was 0.018 and 0.040,respectively.The decision curve demonstrates that the net benefit of predicting TT in the training set and the verification set is higher when the model's probability thresholds are 1%-100%and 1%-93%,respectively.Conclusion In addition to providing a visual aid for clinical decision-making,the nomogram model based on ultrasound indicators in conjunction with clinical factors can accurately estimate TT risk.
Keywords：testicular torsionultrasoundnomogrampredictive models
Publication Date：2025-06-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：9( 690-698 )