Vesicoureteral Reflux and Urinary Tract Infection
| Article DOI | https://doi.org/10.1016/j.cmpb.2021.106369 |
| Objective | To determine VUR grade using images from VCUGs |
| AI Approach | Hybrid CNN (+K-nearest neighbors or + SVM) |
| Data Source(s) | Institutional series (1228 images) |
| Model Input | Raw VCUG images |
| Model Outcome | Normal vs. VUR grade (respective) |
| Model Metrics | AUROC = 0.99, accuracy = 97% |
| Model Usability | NA |
AI = Artificial intelligence, AUROC = Area under the reciever operator characteristic, VUR = Vesicoureteral Reflux, VCUG = Voiding cystourethrogram
AIPEDURO 
Leave a comment