Non-contrast MRI as an alternative to Tc-99m DMSA scintigraphy in the detection of pediatric renal scarring

Objectives Tc-99m DMSA scintigraphy has long been the gold standard for detecting pediatric renal cortical scars, but Tc-99m DMSA is currently not available commercially. Historically, its disadvantages included radiation exposure, a possible requirement for restraints to limit patient motion, and limited spatial resolution. In this context, MRI is a promising alternative able to provide high temporal and spatial resolution imaging of the kidneys without radiation exposure and the need of intravenous contrast. We prospectively investigated a non-contrast MRI protocol to determine if MRI is a feasible substitute for DMSA in the detection of renal scars.

Methods A group of 13 patients (9 female, age 11.1± 4.8 years) scheduled for planar pinhole DMSA scans were recruited under IRB-approved protocol to undergo additional non-contrast MRI on the same day (free-breathing axial and coronal T2 HASTE fat-sat imaging, 7-10 minutes total scanning time). The DMSA and MRI studies were independently evaluated by two radiologists per modality. All radiologists were blinded to the patients’ clinical information. Each reader identified the presence of scars (absent or present) in upper, middle, or lower kidney zones by thirds as well as their level of diagnostic confidence (1-lesion(s) certainly absent to 5-lesion(s) certainly present). The quality of the studies (1-bad, no diagnosis possible to 5-excellent) was also rated. A scoring consensus by the DMSA group for the presence of scars was considered the gold standard.

Results DMSA demonstrated scarring in 16 of the 78 (20.5%) evaluated segments. The two MRI readers demonstrated mean sensitivities of 94.7% (95% CI [74.0-99.9%]) and 89.5% (95% CI [66.9-98.7%]), identical specificities of 100% (95% CI [93.9-100.0%]), and diagnostic accuracies of 98.7% (95% CI [93.1-100%]) and 97.4% (95% CI [91.0-99.7%]). Inter-observer agreement in the MRI group was 98.7% (95% CI [93.1-100%]) and 92.3% (95% CI [84.0-97.1%]) in the DMSA group. The MRI readers were marginally more confident than the DMSA readers when the DMSA consensus diagnosis was positive (p=0.054) and significantly more confident when the consensus diagnosis was negative (p=0.02). On average, the MRI readers were more likely to rate the quality of a study as a 5 (84.6%), the highest possible quality rating, than the DMSA readers (57.7%) (p=0.024).

Conclusions MRI demonstrates excellent sensitivity, specificity, diagnostic accuracy, and inter-observer agreement in detection of renal scarring when compared to DMSA. Given the performance of a rapid non-contrast MRI, which eliminated the need for anesthesia and intravenous contrast, MRI may serve as a substitute for DMSA in the imaging of pediatric renal scarring and aid clinical management, especially in times of DMSA shortage. Further investigation in a larger population is warranted.