Table 2.
Underexplored AI applications in pediatric surgery compared to adult surgery
| Area of application | Status in adult surgery | Status in pediatric surgery | Challenge |
|---|---|---|---|
| AI-guided preoperative risk stratification | Widely used (e.g., LOS prediction) | Limited tools; rare data registries and small cohorts | Lack of large pediatric datasets; condition heterogeneity |
| AI-Enhanced Surgical Robotics and Autonomy | Limited clinical use used in urology, colorectal with AI-driven tasks | Minimal use | Size constraints, regulatory barriers, lack of pediatric-specific platforms |
| AI for Intraoperative Decision Support (e.g., computer vision) | Experimental phase (e.g., structure recognition) | Largely unexplored; no pediatric datasets or validated tools | Few annotated surgical videos; case rarity |
| AI in Postoperative Complication Prediction/Long-term outcomes | Widely used to predict infection, readmission, bleeding risks, QoL | Limited tools; mostly in research phase | Lack of integrated perioperative data systems for children |
| NLP for Operative Notes and Clinical Documentation | Used for quality control, adverse event detection, auto-documentation | Rare use in pediatrics; models not adapted to pediatric language | Need for pediatric-specific ontologies |
| AI in Surgical Education and Simulation | AI-enhanced simulators, skill tracking, rare case training available | Very limited; few pediatric-specific simulators with AI | Case complexity, limited training datasets |