Consistency of Multiple Renal Functional MRI Measurements Over 18 Months

• Published in: Journal of magnetic resonance imaging 2018-08, Vol.48 (2), p.514-521
• Main Authors: Li, Lu‐Ping, Thacker, Jon, Li, Wei, Tan, Huan, Wang, Chi, Kohn, Orly, Sprague, Stuart, Prasad, Pottumarthi
• Format: Article
• Language: English
• Subjects: apparent diffusion coefficient; arterial spin labeling; Bias; Blood flow; blood oxygenation level dependent; chronic kidney disease; Coefficient of variation; Consistency; Correlation analysis; Diffusion coefficient; Field strength; Functional magnetic resonance imaging; Image acquisition; Image processing; kidney; Kidneys; Longitudinal studies; Magnetic resonance imaging; Medical imaging; Neuroimaging; Oxygenation; Patients; Perfusion; Renal cortex; Scanners; Spin labeling; Statistical analysis; Statistical tests; Two dimensional flow
• ISSN: 1053-1807; 1522-2586
• DOI: 10.1002/jmri.26001

Background Identification of patients with progressive chronic kidney disease (CKD) and those likely to respond to candidate therapeutics is urgently needed. Functional MRI measurements have shown promise. However, knowledge about the consistency of the measurements is essential to conduct longitudinal studies. Purpose/Hypothesis To investigate the consistency of repeated functional MRI measurements in healthy subjects. Study Type Prospective, longitudinal study. Subjects Seventeen healthy subjects were examined on two different occasions, 18 months apart. Field Strength/Sequence Multiple gradient‐recalled‐echo, 2D navigator‐gated flow‐sensitive alternating inversion recovery True‐FISP and spin‐echo planar diffusion‐weighted sequences were used on a 3T scanner. Images were acquired on two different scanner configurations. Assessment Blood oxygenation level‐dependent (BOLD) R2*, arterial spin labeling (ASL) perfusion‐derived blood flow (BF) and apparent diffusion coefficient (ADC) maps were analyzed using a custom image processing toolbox. Regions of interest (ROIs) were placed on renal cortex, medulla, and whole kidney. Multiple researchers were involved in defining the ROIs. Statistical Tests Intra‐ and intersubject coefficients of variation (CV) and Bland–Altman plots were used to measure consistency and evaluate bias in the measurements. A nonparametric Wilcoxon test was used to compare differences between two timepoints. Results The intrasubject CV for R2* and ADC were 6.8% and 5.3% with small (−3.8 and 5.3%) bias, respectively, comparing baseline and 18‐month data. Intrasubject CV for renal cortex BF was higher (18.7%) compared to R2* and ADC, but comparable to prior literature values over shorter durations. It also exhibited a larger bias (−15.4%) between two timepoints and significantly lower values (P = 0.022) at 18‐month data. Data Conclusion All three MRI parameters over 18 months, even with a scanner upgrade and involving multiple observers, showed good consistency. These results are useful for the interpretation of longitudinal data and support the use of these methods to monitor progression in patients with CKD. Level of Evidence: 1 Technical Efficacy: Stage 1 J. MAGN. RESON. IMAGING 2018;48:514–521.