How well do we predict renal dysplasia? Comparison between prenatal ultrasound and postnatal renal scintigraphy in CAKUT

To assess the accuracy of prenatal sonographic estimation of renal dysplasia and reduced renal function compared to objective postnatal renal scintigraphy. A retrospective cohort study was conducted in a single tertiary referral centre. Cases diagnosed prenatally with CAKUT, maintained follow-up in the pediatric nephrology institute, and underwent renal scintigraphy, comprised the study group. Prenatal ultrasounds were revised and classified for the risk for dysplasia and future renal function according to the renal parenchymal appearance (thickness, echogenicity, cysts). Cases that demonstrated intrauterine renal failure were excluded. The possibility for renal dysplasia was compared to postnatal renal scintigraphy results (mild and severe dysplasia). A total of 116 kidneys were evaluated (58 cases). The final diagnosis was ureteropelvic junction obstruction in 34 cases, vesicoureteral reflux in 9 cases, and other diagnoses such as posterior urethral valve, ureterocele, and ureterovesical junction obstruction in the remaining cases. Overall, 50 kidneys had dysplastic features on postnatal renal scintigraphy. The sensitivity, specificity, PPV, and NPV for prenatal classification of renal dysplasia were 76%, 84%, 70.5%, and 68.3%, respectively. Although postnatal renal scintigraphy is not performed routinely, the presented results shed light on the prenatal ability to predict renal dysplasia in CAKUT. Estimating renal parenchyma is subjective and operator-dependent, and the predictive value is currently suboptimal even in a tertiary referral centre. Creation of objective tools may improve the predictive ability of prenatal ultrasound to predict postnatal renal function.