The utility of renal sonographic measurements in differentiating children with high grade congenital hydronephrosis

Abstract

Current grading systems for hydronephrosis include a subjective determination of parenchymal 'thickness' and suffer from poor reliability. Use of more objective ultrasonographic measurements including medullary pyramidal thickness (PT) may be useful in augmenting current grading systems by decreasing subjectivity and enhancing prognostic ability. To evaluate the utility of PT measurements in patients with SFU grades 3 and 4 hydronephrosis, we assessed the: 1) the inter-rater reliability of PT measurements, 2) the correlation between relative renal function on nuclear renal scan and PT, and 3) the pyeloplasty predictive ability of PT alone and in combination with SFU grade and/or other sonographic measurements in multivariate statistical models. We retrospectively reviewed 110 children with SFU grade III and IV hydronephrosis. Most patients presented with a history of prenatally detected hydronephrosis at a median age (IQR) of 1.7 months (0.6-5.2). Sixty-two kidneys were followed without operative intervention while 63 underwent pyeloplasty. Indications for surgery included an obstructive drainage pattern with a T1/2>20min on diuretic renal scan in addition todecreased relative renal function less than 40%, increasing hydronephrosis on serial ultrasounds, and/or a decline in relative renal function >10% onserial renal scans. The median age at the time of pyeloplasty was 5.7 months (2.8-13.7). The median time from initial presentation to final follow-up for all patients was 28 months (18.3-44.6). The PT, APD, and renal length were measured on sonographic images. The inter-rater reliability for SFU grading was only fair whereas it was excellent for PT measurements. Receiver operating characteristic (ROC) curves were generated for inclusive multivariate models for prediction of pyeloplasty with and without SFU grade. PT is a reliable and useful measurement to characterize the hydronephrotic kidney parenchyma and a PT>3mm occurs significantly more frequently in patients with a DRF≥45% (p=0.0056). PT alone was predictive of subsequent pyeloplasty (AUC=0.781). A novel pyeloplasty predictive score (PPS) using only objective measurements including PT, APD, and renal length was more accurate than a PPS that incorporated SFU grade (AUC of 0.885 and 0.866, respectively). Utilization of PT≤3mm as a criterion for 'thinned parenchyma' in the SFU, UTD, and other hydronephrosis grading systems should be considered if confirmed by additional studies.