Conversion and reliability of two urological grading systems in infants: the Society for Fetal Urology and the urinary tract dilatation classifications system

Background: The urinary tract dilation (UTD) classification system was proposed in 2014. Objective: To evaluate the correspondence and reliability of two US grading systems for postnatal urinary tract dilatation in infants: the Society for Fetal Urology (SFU) and the UTD systems. Materials and methods: We assessed 180 kidneys in infants younger than 1 year. Four radiologists assessed the kidneys twice using both the SFU system (grades 0 to 4) and the UTD system (grades normal, P1, P2, P3). The SFU system was re-categorized into SFU-A (grades 0, 1–2, 3, 4) and into SFU-B (grades 0–1, 2, 3, 4). The Cohen kappa statistic was used for estimating agreement of both UTD–SFU-A and UTD–SFU-B. Results: The Cohen kappa was significantly higher between UTD and SFU-B as compared to the UTD and SFU-A (0.75 vs. 0.50, P < 0.001). Intra-observer agreement was similar for the two grading systems (SFU 0.64–0.88 vs. UTD 0.48–0.92, P = 0.050–0.885). SFU grades 2 and 3 showed fair to moderate inter-observer agreement and corresponding UTD grades P1 and P2 showed moderate to substantial agreement. The overall inter-observer agreement was significantly higher for the UTD system than for the SFU system during the first assessment (95% confidence interval [CI]: right kidney, −0.069 to −0.062; left kidney, −0.048 to −0.043). Conclusion: Correspondence between the systems was poor using a recommended re-categorization (SFU-A). An alternative re-categorization (SFU-B) was found to be more appropriate for establishing correspondence between the systems. Both systems were reliable, with good intra- and inter-observer agreement for the assessment of infant kidneys, but the UTD system had better inter-observer agreement.