COMPARISON STUDY OF DMSA SCINTIGRAPHY AND NONCONTRAST MRI IN DETECTING RENAL SCARRING IN CHILDREN AFTER ACUTE PYELONEPHRITIS
Daniel B. Herz, MD, MS.
Connecticut Children's Hospital, Hartford, CT, USA.
BACKGROUND: DMSA scan is the gold standard for imaging the renal cortex for renal scarring and dysplasia. However, in the “real world” DMSA is not uniformly available and the study often takes several hours to complete. A reliable alternative to DMSA with the same gold standard sensitivity and specificity as well as image resolution that is more readily available and does not involve radiation exposure is needed. We present a prospective comparison of DMSA and MRI in children after clinical pyelonephritis.
Methods: Under IRB approval and as part of a larger group of children prospectively followed after an episode of acute pyelonephritis, we studied a subgroup of children that had both MRI and a DMSA performed at the same time within 6 months after acute infection. Using DMSA as the gold standard, our purpose was to test the reliability of MRI in detecting renal scarring as well as the inter-observer reliability with kappa statistics. DMSA was performed with parallel/pin hole imaging using between 1-4 mCi. MRI used short time inversion recovery or fat-saturated T2-weighted fast spin imaging. Studies were independently evaluated by 2 radiologists per modality. Radiologists were blinded to patient specific clinical information. Each reader identified the presence of scarring as present or absent in each kidney imaged. A Likert scoring system of 1 to 4 (1=bad, 2=fair, 3=good, 4=excellent) was rated for the quality of studies.
Results: Over a 5-year period, 27 children fit inclusion criteria with imaging of the kidneys with both DMSA and MRI. Sedation or anesthesia was required in 18 of 27 children. Renal scarring was identified in 15 (55.6%) and 17(62.9%) of 27 children by DMSA and MRI, respectively. Positive agreement between MRI and DMSA occurred in 15 children, and negative agreement occurred in 10 children. Disagreement was noted in 2 (7.4%) children with MRI stating presence and DMSA stating absence of renal cortical scarring. Using a coefficient of 0.52 the Kappa statistic between DMSA and MRI was 0.85 indicating very good strength of agreement. On average MRI was read with a quality score of 4.2 while DMSA was read with a quality score of 3.8, (p=0.21). Using DMSA as gold standard, MRI demonstrated a sensitivity of 100% and specificity of 86%. Interobserver agreement in the MRI group was 96.4% [95% CI (93.2-99.5%].
Conclusions: MRI demonstrates excellent sensitivity, specificity, and inter-observer agreement in detection of renal scarring when compared to DMSA in children after an episode of acute pyelonephritis. We speculate that rapid imaging non-contrast MRI may serve as a replacement for DMSA in the imaging of pediatric renal scarring and aid clinical management. However, further investigation in a larger population is warranted.
Back to 2023 Abstracts