Tracking Anesthetic Spread with AI-Powered Ultrasound Imaging
In a study conducted at Showa Medical University Northern Yokohama Hospital, the automated segmentation of local anesthetic spread during ultrasound-guided rectus sheath block was successfully tested using a Mask R-CNN-based artificial intelligence model.
In an exploratory observational study conducted on 40 adult patients between April 2025 and February 2026, an artificial intelligence model was developed to automatically segment the distribution of local anesthetic agent during ultrasound-guided rectus sheath block (RSB).
The research conducted at Showa Medical University Northern Yokohama Hospital utilized the Mask R-CNN architecture and the Detectron2 platform. Trained with reference masks manually marked by experienced anesthesiologists, the model was tested on images extracted from RSB videos at a rate of 10 frames per second.
Within the scope of the study, 20 cases were allocated for training and 20 cases for the test data set. According to the test results, the model's case-based mean Dice coefficient was measured at 0.822, sensitivity at 0.867, and precision at 0.874. These data indicate that artificial intelligence can be used to track dynamically changing anesthetic spread.
Sources
- cureus.com · Link