The effect of reducing spatial resolution by in-plane partial volume averaging on peak velocity measurements in phase contrast magnetic resonance angiography.

BACKGROUND: The aim of this study was to quantify the degree of the effect of in-plane partial volume averaging on recorded peak velocity in phase contrast magnetic resonance angiography (PCMRA).
METHODS: Using cardiac optimized 1.5 Tesla MRI scanners (Siemens Symphony and Avanto), 145 flow measurements (14 anatomical locations; ventricular outlets, aortic valve (AorV), aorta (5 sites), pulmonary arteries (3 sites), pulmonary veins, superior and inferior vena cava)- in 37 subjects (consisting of healthy volunteers, congenital and acquired heart disease patients) were analyzed by Siemens Argus default voxel averaging technique (where peak velocity = mean of highest velocity voxel and four neighbouring voxels) and by single voxel technique (1.3×1.3×5 or 1.7×1.7×5.5 mm3) (where peak velocity = highest velocity voxel only). The effect of scan protocol (breath hold versus free breathing) and scanner type (Siemens Symphony versus Siemens Avanto) were also assessed. Statistical significance was defined as P<0.05.
RESULTS: There was a significant mean increase in peak velocity of 7.1% when single voxel technique was used compared to voxel averaging (P<0.0001). Significant increases in peak velocity were observed by single voxel technique compared to voxel averaging regardless of subject type, anatomical flow location, scanner type and breathing command. Disabling voxel averaging did not affect the volume of flow recorded.
CONCLUSIONS: Reducing spatial resolution by the use of voxel averaging produces a significant underestimation of peak velocity. While this is of itself not surprising this is the first report to quantify the size of the effect. When PCMRA is used to assess peak velocity recording pixel averaging should be disabled.