Aim To assess the prevalence and severity of short-term postoperative complications (≤30 days) following anorectal malformation (ARM) reconstruction, and explore potential risk factors for minor and major complications. Methods This retrospective cohort study included all ARM patients born between January 2000 and December 2025 who underwent definitive reconstruction. Short-term postoperative complications were classified according to the Clavien-Madadi classification. Univariable and multivariable logistic regression analyses were performed. Results 278 patients were included, of whom 49.6% were female. Median age (IQR) at reconstruction was 4 months (3–6). Short-term postoperative complications occurred in 78 patients (28.1%), including 53 (19.1%) minor and 25 (9.0%) major complications. Wound dehiscence occurred most frequently (n = 57; 20.5%). In multivariable analysis, surgical techniques other than ASARP or PSARP were associated with lower odds of minor complications (OR 0.14, 95% CI 0.03–0.62). Major complications were associated with VACTERL association (OR 6.27, 95% CI 2.43–16.22) and inversely with pre-reconstructive colostomy creation (OR 0.26, 95% CI 0.08–0.82). Conclusions Of all patients, 28.1% experienced short-term postoperative complications after ARM reconstruction. Minor complications were associated with type of surgery, whereas major complications were associated with VACTERL association and inversely with pre-reconstructive colostomy creation.
Background Low- and middle-income countries (LMICs), particularly Sub-Saharan Africa, experience a high burden of surgical disease and a disproportionately high postoperative morbidity and mortality. Although machine learning based pediatric surgical risk calculators (SRCs) have been developed to support clinical decision making, most are derived from high-income settings and may not generalize well to LMIC populations. We previously developed a pediatric surgical risk calculator using multinational LMIC data; however, its performance in independent African cohorts remains unknown. Methods We conducted an external validation study of a previously developed pediatric surgical risk calculator using prospective data from children undergoing surgery at four tertiary referral hospitals in Ethiopia and Nigeria between July 2024 and February 2025. Model performance was assessed by evaluating discrimination (area under the receiver operating characteristic curve) and calibration (slope and intercept). Predicted risk was stratified into low, intermediate, and high-risk categories to enhance clinical applicability. Model recalibration was performed to optimize performance in the external cohort. Results A total of 1670 pediatric surgical patients were included from the four hospitals. In-hospital mortality was 6.8%. Mortality varied significantly by site, age, weight, sex, prematurity, preoperative sepsis, transfusion, supplemental oxygen use, antibiotic use, ASA class, type of surgery (emergent vs elective), need of ICU postoperatively and reoperation. The risk prediction model had good discrimination overall with an area under the curve (AUC) of 0.84, outperforming ASA classification alone (AUC = 0.77). Discrimination remained high for general pediatric surgery cases (AUC = 0.83) and was highest for emergency cases (AUC = 0.89), but was low in neonates (AUC = 0.64). The model demonstrated systematic miscalibration, which improved following logistic calibration. Patients were stratified into meaningful risk categories, which reliably distinguishes between low-and-hgih risk patients (LR = 0.1 and LR = 5.4 respectively). However, 66.5% fall in the intermediate risk group which is not useful in decision making (LR = 1.0). Conclusion The pediatric surgical risk calculator demonstrated good performance in the African context, with improved clinical utility following recalibration and risk stratification. Reduced performance in neonates suggests the need to incorporate additional risk factors and potentially separate models for this subgroup. Future efforts should focus on supporting the use of the SRC in risk communication, clinical decision making, and resource prioritization.
Background Artificial intelligence (AI) applied to surgical video is often evaluated by model performance alone, leaving clinical usefulness uncertain. We evaluated a U-Net-derived semantic segmentation model as an assistive tool for identifying the tracheoesophageal fistula (TEF) during thoracoscopic esophageal atresia repair. Methods The model was trained on 7287 annotated frames from 26 surgical videos and evaluated by four-fold case-based cross-validation and an independent five-case test set. Forty-eight pediatric surgeons (7 experts and 41 novices) completed paired video-based identification tasks with and without AI assistance and a separate frame-based segmentation task. AI-on/off comparisons used surgeon-level medians. Real-time deployment was assessed in two operations. Results The median cross-validation Dice coefficient was 0.628. In clip-based testing, AI identified the TEF faster than experts and novices (1.05 versus 2.87 and 6.92 s). Among novices, median identification time was 263 (190–453) vs. 390 (272–555) frames (AI-on vs. AI-off; P < 0.001), and confidence was 4 (4–5) vs. 3 (3–4) (P < 0.001). Among experts, median identification time was 186 (174–201) vs. 172 (157–196) frames (P = 0.22), and confidence was 5 (5–5) vs. 5 (4–5) (P = 0.031). In the controlled frame task, AI achieved a higher Dice coefficient (0.881) than experts (0.803) and novices (0.780). Real-time overlay was technically feasible in two cases without observed AI-related adverse events. Conclusions These findings support prospective, multi-institutional evaluation of AI assistance for pediatric thoracoscopic navigation and training.
