Summary
This study aimed to assess the potential of artificial intelligence (AI) in enhancing the diagnosis of pediatric appendicitis using ultrasonography (US). The AI neural network was trained on 50 abdominal ultrasound videos. In test 1, the AI’s performance was evaluated on 20 additional videos, with some videos demonstrating a significant positive impact on detection and accuracy.
However, in test 2, when the AI failed to detect appendicitis in videos with a shallow scan depth, it harmed examiners. The study suggests that AI can assist in identifying appendicitis in the US, especially when the appendix is shallow or partially detected, but its limitations should be considered.