Non-contrast-enhanced magnetic resonance imaging can be used to assess renal cortical and medullary volumes-A validation study

Background: Magnetic resonance imaging (MRI) biomarkers can diagnose and prognosticate kidney disease. Renal volume validation studies are however scarce, and measurements are limited by use of contrast agent or advanced post-processing. Purpose: To validate a widely available non-contrast-enhanced MRI method for quantification of renal cortical and medullary volumes in pigs; investigate observer variability of cortical and medullary volumes in humans; and present reference values for renal cortical and medullary volumes in adolescents. Materials and Methods: Cortical and medullary volumes were quantified from transaxial in-vivo water-excited MR images in six pigs and 15 healthy adolescents (13-16years). Pig kidneys were excised, and renal cortex and medulla were separately quantified by the water displacement method. Both limits of agreement by the Bland-Altman method and reference ranges are presented as 2.5-97.5 percentiles. Results: Agreement between MRI and ex-vivo quantification were -7 mL (-10-0 mL) for total parenchyma, -4 mL (-9-3 mL) for cortex, and -2 mL (-7-2 mL) for medulla. Intraobserver variability for pig and human kidneys were <5% for total parenchyma, cortex, and medulla. Interobserver variability for both pig and human kidneys were <= 4% for total parenchyma and cortex, and 6% and 12% for medulla. Reference ranges indexed for body surface area and sex were 54-103 mL/m(2) (boys) and 56-103 mL/m(2) (girls) for total parenchyma, 39-62 mL/m(2) and 36-68 mL/m(2) for cortex, and 16-45 mL/m(2) and 17-42 mL/m(2) for medulla. Conclusion: The proposed widely available non-contrast-enhanced MRI method can quantify cortical and medullary renal volumes and can be directly implemented clinically.