Journal of pediatric surgery最新文献

Purpose: This study evaluates the effectiveness of machine learning (ML) algorithms for improving the preoperative diagnosis of acute appendicitis in children, focusing on the accurate prediction of the severity of disease.

Methods: An anonymized clinical and operative dataset was retrieved from the medical records of children undergoing emergency appendectomy between 2014 and 2021. We developed an ML pipeline that pre-processed the dataset and developed algorithms to predict 5 appendicitis grades (1 - non-perforated, 2 - localized perforation, 3 - abscess, 4 - generalized peritonitis, and 5 - generalized peritonitis with abscess). Imputation strategies were used for missing values and upsampling techniques for infrequent classes. Standard classifier models were tested. The best combination of imputation strategy, class balancing technique and classification model was chosen based on validation performance. Model explainability was verified by a pediatric surgeon. Our model's performance was compared to another pediatric appendicitis severity prediction tool.

Results: The study used a retrospective cohort including 1980 patients (60.6 % males, average age 10.7 years). Grade of appendicitis in the cohort was as follows: grade 1-70 %; grade 2-8 %; grade 3-7 %; grade 4-7 %; grade 5-8 %. Every combination of 6 imputation strategies, 7 class-balancing techniques, and 5 classification models was tested. The best-performing combined ML pipeline distinguished non-perforated from perforated appendicitis with 82.8 ± 0.2 % NPV and 56.4 ± 0.4 % PPV, and differentiated between severity grades with 70.1 ± 0.2 % accuracy and 0.77 ± 0.00 AUROC. The other pediatric appendicitis severity prediction tool gave an accuracy of 71.4 %, AUROC of 0.54 and NPV/PPV of 71.8/64.7.

Conclusion: Prediction of appendiceal perforation outperforms prediction of the continuum of appendicitis grades. The variables our models primarily rely on to make predictions are consistent with clinical experience and the literature, suggesting that the ML models uncovered useful patterns in the dataset. Our model outperforms the other pediatric appendicitis prediction tools. The ML model developed for grade prediction is the first of this type, offering a novel approach for assessing appendicitis severity in children preoperatively. Following external validation and silent clinical testing, this ML model has the potential to enable personalized severity-based treatment of pediatric appendicitis and optimize resource allocation for its management.

Level of evidence: 3:

Background: Rural facilities that provide pediatric surgical services are a critical resource to local communities. Our aim was to characterize differences in outpatient pediatric cholecystectomy outcomes performed at rural and urban hospitals with the hypothesis that rural hospitals would have similar outcomes.

Methods: The Nationwide Ambulatory Surgery Sample (NASS), which contains ambulatory surgery encounters at hospital-owned facilities, was used to perform a retrospective cohort analysis of pediatric patients age 18-years and younger who had a cholecystectomy (n = 15,449) between 2016 and 2018. Survey-weighted multivariate regression analyses were used to evaluate the association of rural and urban hospital locations with the primary outcome of routine discharge. Models were adjusted for sex, age, comorbidities, insurance, income quartile, indication for operation, and hospital characteristics.

Results: Patients receiving care at rural hospitals were more commonly in the lowest two median income quartiles compared to urban hospitals (91.1 % versus 58.2 %, p < 0.01) and had more complex chronic comorbidities (63.1 % versus 57.2 %, p < 0.01). In both settings, the most common surgical indication was cholecystitis followed by biliary dyskinesia; nearly all procedures were performed laparoscopically (>99 %), and complications were rare with no deaths. Survey weight multivariate regression analysis showed patients cared for at rural hospitals had a higher odds ratio of routine discharge compared to those cared for at urban hospitals (adjusted OR 2.27, 95 %CI 1.04-4.99, p = 0.04).

Conclusions: Pediatric patients at rural hospitals have lower median income and a higher mean morbidity index. Despite these challenges, outpatient cholecystectomy was safe at rural hospitals and associated with increased odds of routine discharge when compared with urban hospitals.

Background: We aimed to analyze the effect of age and weight on 30-day outcomes of single-stage endorectal pull tthrough (ERPT) procedures for Hirschsprung's Disease (HD) using the National Surgical Quality Improvement Program-Pediatric (NSQIP-P) database to identify an optimal time for surgery.

Methods: We queried NSQIP-P for children <2 years with HD who underwent ERPT between 2016-2021 and did not have a preoperative stoma. Patients were stratified by age (<3 months, 3-6 months and >6 months) and weight (<4 kg, 4-8 kg, and >8 kg) at the time of surgery. The primary outcome measure was major adverse outcomes, and the secondary outcomes were 30-day unplanned readmission and reoperation. Multivariable logistic regression was conducted to analyze the association between age and weight and each outcome.

Results: A total of 1002 patients were included with a median age of 31 days (IQR: 10-106 days). Patients were stratified by age (n = 702, 70.1 %; n = 158, 15.8 %; n = 142, 14.2 % respectively from youngest to oldest) and weight (n = 472, 47.1 %; n = 421, 42.0 %; n = 109, 10.9 % respectively from lowest to highest weight) groups. On multivariable analysis, age>6 months and weight>8 kg were independently associated with a higher rate of major complications (aOR = 2.741, 95 % C.I. = 1.234-5.880 and aOR = 4.627, 95 % C.I. = 1.761-11.775 respectively). Conversely, being in the highest age (aOR = 0.359, 95 % C.I. = 0.176-0.699) and weight brackets (aOR = 0.396, 95 % C.I. = 0.171-0.801) independently predicted lower 30-day readmission rates.

Conclusion: Earlier single-stage ERPT for HD was found to be associated with lower complication rates in a select group of patients but may result in increased rates of readmissions and hospital resource utilization.

Level of evidence: Level III evidence.

Type of study: Retrospective cohort study.

Background: The Heineke-Mikulicz anoplasty (HMA) is a technique for addressing skin-level postoperative strictures following posterior sagittal anorectoplasty (PSARP) and posterior sagittal anorectovaginourethroplasty (PSARVUP). We aimed to evaluate the long-term outcomes with HMA as a treatment for postoperative anal strictures.

Methods: A retrospective review was conducted for patients with a history of ARM who underwent HMA for skin-level anal strictures. Data were collected regarding indication for HMA, most recent anorectoplasty preceding HMA (primary or redo), timing of HMA, extent of stricturoplasty, change in caliber of anus, and postoperative outcomes. Descriptive statistical analysis was performed.

Results: From 2015 to 2024, 138 patients with history of ARM underwent HMA. Median age at time of HMA was 4 years [IQR 2-6] with median follow up time of 3.0 years [IQR 1.0-4.0]. HMA was performed after primary PSARP or PSARVUP in 29 % of patients and after redo surgery in 61.6 %. The minimum time between index operation and initial HMA was 2 months and the median was 5 months [IQR 3-24]. 14.5 % of patients required repeat HMA. The median time between index operation and repeat HMA was 24 months (IQR 12-39). There was a significant trend between number of quadrants where HMA was performed and need for repeat HMA.

Conclusion: HMA is a safe and durable procedure that can treat skin-level postoperative anal strictures in most patients following PSARP and PSARVUP. Patients can frequently be managed with a one-time HMA. HMA should be considered as an alternative to dilation for skin-level anal strictures, especially in an older patient after redo surgery.