Development and Validation of a Nomogram Predicting the Probability of in Hirschsprung Children Following Surgery

This retrospective study was conducted on patients diagnosed with HSCR who underwent surgical treatment at our hospital department from January 2000 to December 2019. This research study was reviewed and approved by the ethics committee at Hospital, waiving the requirement for informed consent in alignment with the Declaration of Helsinki. The number of the institutional reviewboard approval is S108. The diagnosis of HSCR was established through preoperative examinations, which included barium enema, anorectal manometry, and suction rectal biopsy. Additionally, postoperative pathological examination was conducted to further confirm the diagnosis. All operations were performed in accordance with the standard procedures by experienced surgical groups in our department.

2.2. Definition of soiling This study refers to the diagnostic criteria for soiling as described in the published literature . Based on the previously established criteria and clinical experience, patients who experienced soiling persisting for more than six months following surgery were ultimately identified for inclusion in this study. In this study, patients with Down's syndrome, and other established syndromes potentially contributed to maldevelopment were excluded as previous described . In addition, patients who had their initial surgery at another institution were excluded from the study, as well as any participants received staged treatment or had missing data related to soiling.

In this study, The investigators collected the clinical information about patients from a thorough review of relevant literatures and clinical judgement. These data included the age for definitive surgery, sex, weight at operation, premature delivery history, family history, meconium, duration of constipation, conservative treatment, surgical history, surgical techniques, length of removed bowel, surgical procedure, and soiling status. The surgical techniques included laparoscopic-assisted pull-through [LAPT], laparotomy-assisted endorectal pull-through [LEPT], and transanal endorectal pull-through [TEPT]. Length of removed bowel included left colectomy, subtotal ectomy, and total colectomy. Surgical procedures were categorized as Soave, Heart-Shaped, Rebbein, Duhamel, and others. Data with a deletion rate >= 20% were excluded from the analysis. The trimming method was introduced to address with abnormal values and a multiple imputation method was employed to interpolate missing data All patients were randomly assigned into training and validation set in a 7:3 ratio as previous reported. No significant differences were observed in demographic and clinical characteristics between the training and validation sets. To determine potential predictive factors, the training set was subjected to LASSO regression analysis, which effectively eliminated several irrelevant or multicollinearity independent variables to reduce high-dimensional data . Then, the multivariable logistic regression analysis was applied to determine the variables of soiling in patients with HSCR following surgery over one year of age. Finally, a nomogram was created based on the training set and validated in the validation set.

Discrimination and calibration were employed to assess and test the predictive accuracy of the established model. Harrell's concordance index [C-index] and the receiver operating characteristic [ROC] curve [AUC] were generated to estimate the discrimination of the nomogram. A C-index close to 1 indicates a high level of predictive ability of the model. Additionally, an AUC > 0.80 indicates relatively good discrimination. The calibration of the nomogram was accompanied with the Hosmer-Lemeshow test were developed to determine the consistency between the predicted and observed occurrence of soiling. Lastly, DCA was performed to evaluate clinical practicability of the nomograms. Both discrimination and calibration were assessed by bootstrapping with 1000 resamples.

For parameters with continuous variables, the normal distribution was expressed as mean ± standard deviation [SD], and the skewed distribution was presented as median [M] and quartile range [Q25-Q75]. Continuous data between two groups were compared using the independent t-test. The unequal variance t-test was used to compare data between groups with unequal variances. In cases of skewed distribution, the Mann-Whitney U test was employed. Furthermore, categorical data were compared using 2 or Fisher's exact test. Statistical analysis was performed using R Software v.4.0.2 [The R Project for Statistical Computing, www.r-project.org] with the "rms" package utilized for logistic regression analysis and nomogram construction. A two-sided P-value less than 0.05 was considered statistically significant.