Need for surgery in UPJO
| Article DOI | https://doi.org/10.1038/s41598-024-72271-9 |
| Objective | To predict obstructive hydronephrosis requiring surgery from renal US in children with prenatal hydronephrosis |
| AI Approach | CNN |
| Data Source(s) | Multi-institutional series (294 patients, 1645 US images) |
| Model Input | 256 x 256 pixel images of renal US |
| Model Outcome | Requiring surgery |
| Model Metrics | AUC of 0.93, accuracy of 58% |
| Model Usability | No available code or dataset. Model explainability provided with Grad-CAM (overlayed on input images). |
AI = Artificial intelligence, UPJO = Ureteropelvic junction obstruction, CNN = Convolutional neural network, AUC = Area under the receiver operator characteristic
AIPEDURO 
Leave a comment