Construction And Validation Of Nomogram Model For Estimating The Risks Of Surgical Intervention In The Neonates With Necrotizing Enterocolitis

Purposes We aimed to use LASSO to screen the risk factors for surgical intervention of neonates with necrotizing enterocolitis(NEC),and develop a nomogram model that early assesses the risks of surgical intervention in the neonate with Necrotizing enterocolitis(NEC).Methods Retrospective analysis was conducted on neonates diagnosed as NEC(Bell stage II and above)by Guangzhou Women and Children’s Medical Center from January 2016 to July 2020.They were divided into surgical group and conservative group according to whether to undergo surgical treatment.Collect the clinical data,radiographic Results and laboratory results of the two groups of children at the time of onset,use LASSO regression to screen out the risk factors for NEC children requiring surgical treatment,and further incorporate multivariate logistic regression to establish a prediction model using R language to construct a nomogram.Then analyze the predictive effectiveness of the model through the ROC,correction curve,and decision curve.Select the NEC children admitted to our hospital from June 2021 to December2022 for internal verification.Results A total of 194 neonates with NEC were included,including 73 cases(37.6%)in the surgical group and 121 cases(62.4%)in the conservative group.(1)Using LASSO regression,seven risk factors were selected,including male infants,birthweight,sepsis before onset,shock before onset,presence of portal vein gas at onset,absolute value of monocytes at onset,and the ratio of neutrophils to lymphocytes at onset.(2)These factors were further incorporated in Logistic regression was and constructed a nomogram base on these factors by using R language.Using the modeling set data to draw the ROC curve,the area under the curve(AUC)was 0.890(95% CI: 0.844～0.936),the optimal cutoff value was 0.334,the sensitivity was 86.3%(95% CI: 75.8%～92.9%),and the specificity was 81.8%(95% CI: 73.5%～88.0%).The model was internally validated using the Bootstrap method(sampling times: 1000),indicating that the correction curve and the ideal curve fitted well,this indicates that the nomogram model has a high prediction ability.(3)The validation set included 16 patients in the surgical group and 23 patients in the conservative group.Base on the nomogram model to estimating the risks of surgical intervention of neonates in validation set,the sensitivity was 81.3%(95% CI: 53.7%to 91.7%),the specificity was 73.9%(95% CI: 51.3% to 88.9%),the positive predictive value was 68.4%(95% CI: 43.5% to 86.4%),and the negative predictive value was 85.0%(95% CI: 61.1% to 96.0%).Based on the model established by the modeling set,draw the ROC curve of the validation set,with an area under the curve of 0.861.Conclusions This study has constructed a nomogram model that includes seven predictive indicators,including gender,birth weight,sepsis,shock,portal vein pneumatosis,absolute monocyte value at onset,and NLR value,to predict the risk assessment of surgical treatment in children with NEC.The ROC curve,calibration curve,and DCA curve suggest that this nomogram model has good predictive ability and clinical decision-making benefits,it can effectively assist clinicians in early identification and early diagnosis of NEC children requiring surgical treatment,thereby improving the prognosis of such children.