Fluid-attenuated inversion recovery MR imaging: identification of protein concentration thresholds for CSF hyperintensity

Using human volunteers and phantoms emulating CSF, we analyzed the effects of varying protein concentration on the signal intensity of saline solution. Also, for different fluid-attenuated inversion recovery (FLAIR) sequences, we compared protein concentration thresholds above which the signal from these solutions becomes hyperintense to that from brain parenchyma.Nine albumin solutions of varying concentrations (3.9 mg/dl to 2500 mg/dl) were imaged using fast FLAIR MR sequences (TR, 6000 msec: inversion time, 1730 msec: echo train length, 20) at different effective TEs (110, 150, 200, and 250 signal-to-noise ratios from the different albumin solutions versus albumin concentration were generated and correlated with average signal-to-noise ratios from brain parenchyma and CSF.We saw a gradual increase in signal-to-noise ratios from the albumin solutions as a function of albumin concentration. As the effective TE increased, the point of intersection between the plots and the average signal-to-noise ratio from brain parenchyma occurred at lower albumin concentrations.FLAIR MR imaging is potentially useful to evaluate pathologic conditions that increase CSF protein concentration. Using phantoms and healthy volunteers, we defined a protein concentration threshold above which the signal from saline solutions becomes hyperintense to that from brain parenchyma. This threshold depends on the effective TE used in the FLAIR sequence and is 250 mg/dl for an effective TE of 110 msec, 125 mg/dl for 150 msec. 110 mg/dl for 200 msec, and 95 mg/dl for 250 msec.