The utility of renal sonographic measurements in differentiating children with grades 2, 3, and 4 hydronephrosis

Introduction: Prior analysis of children with grade 3 and 4 congenital hydronephrosis demonstrated that renal medullary pyramidal thickness (PT) is predictive of subsequent pyeloplasty (area under the curve [AUC] = 0.78). The objective of this study was to further analyze the utility of sonographic measurements including PT, anteroposterior pelvic diameter (APD), and renal length with an expansion of the number of infants with hydronephrotic kidneys including grades 2, 3, and 4 hydronephrosis.

Methods: A retrospective review was performed of 166 consecutive infants with hydronephrosis (188 hydronephrotic kidneys), out of which 144 had unilateral hydronephrosis and 22 had bilateral hydronephrosis. 109 kidneys were observed and 79 underwent pyeloplasty. Indications for surgery included an obstructive drainage pattern with a T1/2 > 20 min on diuretic renal scan in addition to decreased relative renal function less than 40 %, or increasing hydronephrosis on serial ultrasounds, and/or a decline in relative renal function >10 % on serial renal scans. The PT, APD, and renal length were measured on sonographic images. Receiver operating characteristic (ROC) curves were generated for inclusive multivariate models for prediction of pyeloplasty with and without Society for Fetal Urology (SFU) grade.

Results: The AUC of the ROC constructed using only PT to predict pyeloplasty was 0.82, which indicates strong predictive ability. A PT cutoff of 3 mm would identify 72 % of kidneys requiring pyeloplasty and accurately exclude 81 % of those not needing the procedure. Logistic regression modeling identified PT ≤ 3 mm, APD, SFU grade, and female gender as significant factors associated with increased risk of pyeloplasty. The ROC curves of the predictive performance of multivariate predictive models with and without SFU grade demonstrated an AUC of 0.901 and 0.868, respectively.

Conclusions: These results further support PT as a reliable and useful measurement to characterize the hydronephrotic kidney parenchyma and that a PT ≤ 3 mm is a useful criterion for "thinned parenchyma," which could be used to differentiate kidneys in the SFU, UTD, and other hydronephrosis grading systems. Predictive models using only sonographic measurements, including PT, APD, and renal length, were almost as accurate as models incorporating SFU grade with the former offering the advantage of being more objective and reliable.