Article of the Month - May 2021


Front Pediatr. 2020 Jan 28;7:527. doi: 10.3389/fped.2019.00527 PMID: 32047727; PMCID: PMC6997479.

Damasio MB, Bodria M, Dolores M, Durand E, Sertorio F, Wong MCY, Dacher JN, Hassani A, Pistorio A, Mattioli G, Magnano G, Vivier PH.

Reviewed by Siddharth Shah


Congenital obstructive uropathy, a part of the CAKUT spectrum, is a common neonatal urinary tract anomaly and may be seen in up to 2% of normal pregnancies. 25% of neonates born with severe hydronephrosis may require surgical correction to prevent renal scarring and dysfunction. Dynamic renal scintigraphy (DRS) is the current standard to evaluate split renal function (SRF) and renal drainage (curve) function. Functional MR urography is recently validated to assess Ureteropelvic junction (UPJ) obstruction. Previous studies have also shown MR urography as a helpful alternative study to evaluate drainage curves and SRF in obstructive uropathy. When compared to DRS, MR urography offers advantages of obtaining the volume of enhanced renal parenchyma and studying renal volume loss in detail, better depicting anatomical information in complex and familiar phenotypes of CAKUT, and an absence of ionizing radiation with this study.

What was the purpose of the study?

The purpose of the study was to compare functional MR urography with DRS in measuring SRF and the classification of renal drainage curves in patients affected by different obstructive CAKUT phenotypes.

What were the study design and methods?

Patient selection was retrospective. Patients who had severe hydronephrosis (anterior-posterior renal pelvis diameter of ≥ 10 mm) and who underwent both DRS and MR urography within six months without surgical intervention or urinary tract infection in between were selected. All patients had GFR>90 ml/min/1.73m2 at recruitment. Patients with vesicoureteral reflux (VUR) and posterior urethral valves on VCUG were excluded. In patients with unilateral hydronephrosis, normal kidney and the urinary tract was used as control. The functional MR urography imaging was blindly evaluated twice by two Pediatric radiologists from different centers for a total of 4 readings to obtain both inter-reader and intra-reader reproducibility. A single radiologist blindly assessed DRS imaging for a total of 2 readings to obtain intra-reader reproducibility.

In DRS, the SRF was determined using the Rutland-Patlak method, or Area under the Curve method. The drainage was classified as ‘normal,’ ‘borderline,’ or ‘accumulation’ using Tmax and renal output efficiency. In MR Urography, a T2-weighted sequence was obtained for renal volume measurement, and a T1-weighted dynamic contrast-enhanced sequence was used to assess SRF (using Area under Curve and Rutland-Patlak method) and drainage curve function (classified as normal, borderline, or accumulation using O’Reilly DRS-drainage curve classification). The MR urography drainage curve patterns for these classifications were introduced and figuratively shown in the paper to compare with DRS curve patterns.

What were the results of the study?

52 children (age range: 1-19 years, median: 1 year) were selected, and 104 kidney-urinary tracts were initially considered. For drainage curves, 88 kidney-urinary tract imaging was found suitable for analysis. 14/52 children had bilateral hydronephrosis. No adverse effects were seen with DRS or MR urography.

  1. SRF Consistency

    For MR urography, the intra-reader repeatability (using correlation coefficient) to assess SRF in normal kidneys for readers 1 and 2 was 0.97 and 0.99, while the exact repeatability to assess pathological kidneys was 0.95 and 0.96, respectively. The inter-reader repeatability was also high at 0.93 for normal kidneys and 0.95 for pathological ones. There was excellent consistency in the results. For DRS, the intra-reader repeatability also had a high correlation of 0.98.

  2. SRF Concordance

    There was good concordance for SRF obtained from DRS and MR urography with Area under the curve (AUC) % for normal kidneys showing the mean bias of 0.50 (agreement limits: -14.35, 15.34)  and AUC% for pathological kidneys had a mean bias of -0.37 (agreement limits: -13.66, 12.91). This was demonstrated in Bland-Altman plots in the paper.

  3. Equivalence in the classification of drainage curves

    Reader 1 and Reader 2 had observed agreement of 72% (Cohen’s Kappa-0.557) in MR urography drainage curve classification reflecting fair agreement. The DRS-drainage curve classification was compared with the MRU-drainage curve classification (obtained from reader 2) using a 3x3 consistency table (normal, borderline, accumulation patterns). There was a moderate level of concordance and fair agreement among drainage curve classification between DRS and MR urography (observed agreement=71.6%, Cohen’s Kappa=0.563). In 63/88 kidney urinary tract images, both DRS and MR urography rendered similar classification (normal-32, borderline-18, accumulation-13). The misclassification was seen in 25/88 kidney urinary tract images, particularly among borderline drainage curves.

    This misclassification between DRS and MR urography curve patterns may be related to the difference in timing of diuretic administration (T0 in MRU vs. T20 in DRS).

What was the conclusion?

Functional MR Urography is a valuable emerging imaging technique to combine detailed anatomical information with SRF and functional evaluation in children with obstructive uropathy. The significant limitations of widespread use of this technique include cost, lack of experienced radiologists, limited access, and time for post-procedure acquisition analysis that requires 45 minutes per patient compared to < 10 min with DRS.

Why does this matter?

MR urography has apparent advantages in providing finer anatomical details and volumetric assessment of kidneys that are important in evaluating complex phenotypes of CAKUT, assessing prognosis, and making surgical decisions. MR urography is a safe technique, does not carry a risk of ionizing radiation, and may also be considered in children with moderate renal dysfunction using low dose macrocyclic, which is safer than linear gadolinium-based contrast agent and which significantly decreases the risk of nephrogenic systemic fibrosis. Functional MR urography has a high level of consistency and concordance for the evaluation of split renal function. Further studies with more extended time follow-up may be needed to assess drainage curve patterns.