Purpose: The purpose of this work was to explore the reproducibility of fat-fraction measurements using Dixon quantitative chemical shift imaging (QCSI) in the lumbar spine (L3, L4, and L5) of healthy volunteers.
Method: Sixteen healthy volunteers were examined at 1.5 T two times to obtain a repeated measurement in the same slice and a third time in three parallel slices. Single slice, two point Dixon SE (TR/TE 2,500/22.3) sequences were used, from which fat-fraction images were calculated. The fat-fraction results are presented as averages over regions of interest, which were derived from the contours of the vertebrae. Reproducibility measures related to repeated measurements on different days, slice position, and contour drawing were calculated.
Results: The mean fat fraction was 0.37 (SD 0.08). The SD due to repeated measurement was small (sigmaR = 0.013-0.032), almost all of which can be explained by slice-(re)-positioning errors.
Conclusion: When used to evaluate the same person longitudinally in time, Dixon QCSI fat-fraction measurement has an excellent reproducibility. It is a powerful noninvasive tool in the evaluation of bone marrow composition.