Non-contrast renal MR angiography: a within-subject comparison of imaging performance at 1.5 T, 3 T, and 5 T

Abstract

This within-subject study compared non-contrast renal MRA (NC-MRA) at 1.5 T, 3 T and 5 T via comprehensive quantitative and subjective imaging assessments, and further explored imaging disparities between standard and high-resolution 5 T sequences. Thirty-nine healthy volunteers underwent standardized same-day non-contrast MRA scans across the three magnetic field strengths; both standard and high-resolution protocols were implemented on the 5 T scanner. Quantitative parameters including apparent signal-to-noise ratio (aSNR), apparent contrast-to-noise ratio (aCNR) and vessel sharpness were calculated using the erector spinae muscle as the local noise reference, an approach intended to reduce g-factor related noise bias between different scanners. Two senior radiologists performed subjective image scoring independently as blinded readers. All images were fully anonymized by removing scanner and sequence identifiers, and the presentation order of image datasets was randomly rearranged for each reviewer to minimize potential reading bias. Statistically significant inter-group differences were detected for all quantitative indicators (all p  < 0.05). The 5 T standard protocol yielded the highest median aSNR and aCNR values, while the 5 T high-resolution sequence achieved optimal vessel sharpness and clearer visualization of distal subsegmental renal branches. Both 5 T protocols were superior to 1.5 T and 3 T sequences in terms of artifact suppression, vascular contour delineation and subjective diagnostic confidence. The two tested 5 T NC-MRA protocols were associated with improved image-quality metrics and artifact scores in healthy volunteers. These findings provide quantitative baseline data for normal renal vasculature and support further clinical validation of NC-MRA protocols in patients with renal artery disease or contraindications to gadolinium-based contrast agents.