Quality control of fat fraction quantification in magnetic resonance imaging: A two-center phantom study

BACKGROUND: Assessment of quantitative parameters using magnetic resonance imaging (MRI) is a relevant trend. Fat fraction (FF) calculation provides new opportunities for accurate diagnosis and will replace invasive methods such as biopsy in the future. Quantification enable reliable dynamic monitoring assessment drug therapy. However, radiologists clinical specialists must be confident accuracy reliability measures.
 AIM: To assess FF measurement phantom simulation range 0% to 60%.
 METHODS: Emulsions oil-in-water type were chosen simulate objects study. Concentrations vegetable oil-based emulsions presented 60%. Tubes containing placed cylindrical phantom. Scans performed on an Optima, MR450w 1.5 Tesla (GE) tomograph Lava Flex mode Ingenia, (Philips) DIXON mode. was determined by formulas images In-phase Out-phase based signal characteristics (FF=[InOut]/2∙In∙100) weighted Water data (FF=Fat/[Fat+water]∙100.
 RESULTS: The fat percentage with technique identical that Flex. measured concentration systematically overestimated relation target values average 57.6% absolute difference 17.2%. In addition, irregular underestimation 20% 40% detected.
 CONCLUSIONS: Phantom direct allowed control performance Dixon sequences quantifying FF. For correct quantification, from image formula FF=Fat/(Fat+water)∙100 preferable. Calculations provide ambiguous results. modes correction factor. use allows proper quality calibration MRI scanner makes widely available.