Hydronephrosis
| Article DOI | https://doi.org/10.1111/bju.16159 |
| Objective | To predict the risk of renal obstruction on diuretic renography using routine reported ultrasonography findings |
| AI Approach | Random Forest Model (multi-class classification) |
| Data Source(s) | Internal: 304 patients (The Hospital for Sick Children) Validation: 64 patients (Children’s Hospital of Philadelphia) |
| Model Input | Age, Sex, Kidney laterality, Kidney length, Anteroposterior diameter, SFU grade |
| Model Outcome | T1/2 time on diuretic renogram (<20 min, 20-60 min, >60 min) |
| Model Metrics | Multi-class AUROC = 0.75 Binary AUROC (>20 min) = 0.84 |
| Model Usability | https://sickkidsurology.shinyapps.io/AERO/ |
AIPEDURO 
Leave a comment