Phase difference enhanced imaging (PADRE) is a post-processing imaging method using magnitude and phase data. PADRE has shown the detailed structure of the white matter fiber bundle and nucleus.^1-3 Recent reports suggest that phase shift in the white matter may be related to myelin content. We, therefore, hypothesized that PADRE might reflect the age-related change in the white matter in children. The purpose of this study was to compare signal changes in the white matter structures on PADRE with age in children.

Twenty-seven small children (age, 0‒73 months; mean 18.7), who had screened at a 3T-MRI, resulting in no abnormal findings, were enrolled. PADRE images were obtained by three-dimensional principles of echo shifting using a train of observations (PRESTO) sequence (repetition time/echo time = 18.5/22.7‒26.7/38.2 ms; flip angle = 10 degrees; matrix = 212‒294; thickness = 2 mm; field of view = 160‒250 mm). Magnitude and phase images were transferred to a workstation, where phase wraps on the phase images were removed using the homodyne filer with a kernel size of 40. Then a mask was generated by selecting relative small positive phase shift in the right handed system; the final PADRE image was created by a combination of this mask and magnitude image.

PADRE images were coronally reformatted. Signal changes indicating phase shift in the white matter were assessed by two experienced neuroradiologists using four-point scale: 0 = no signal changes, 1 = slight signal change, 2 = moderate signal change, and 3 = prominent signal change. The following three slices were used for this assessment: slice 1, at the level of the anterior horn of the lateral ventricle; slice 2, at the level of the pyramidal tract; and slice 3, at the level of the posterior hone of the lateral ventricle. In turn, the following anatomical locations were evaluated at each slice: subcortical white matter in the frontal lobe at slice 1, subcortical white matter in the Rolandic area and the temporal lobe, and the pyramidal tract at slice 2, and the subcortical white matter in the parietal lobe. Differences in the score between the two readers were solved by a consensus.

These scores were correlated to the age of the children according to each anatomical location by Spearman correlation coefficients.

The relationship between the assessed scores and age of children according to each anatomical location is shown in Figure 1, The assessed scores were 0‒2, 1‒3, 1‒3, 0‒2, 0‒2, and 1‒3, for subcortical white matter in the frontal lobe, Rolandic area, temporal lobe, parietal lobe, and the pyramidal tract, respectively. The scores were generally increased with the increment of age; Spearman ρ values (P value) were 0.75 (P < 0.001), 0.49 (P = 0.009), 0.66 (P < 0.001), 0.37 (P = 0.057), and 0.96 (P < 0.001) for the subcortical white matter of the frontal lobe, Rolandic area, temporal lobe, parietal lobe, and the pyramidal tract, respectively (Figure 2). The scores for the pyramidal tract were highest among the assessed locations, and showed stronger correlation to the age of children, whereas the scores for the temporal lobe were lower, compared to the other locations, and showed less correlation to the age of children.

We found age-related phase changed in relatively small children on PADRE. The signal changes on PADRE increased with age in the subcortical white matter as well as the pyramidal tract, which indicated more phase shift in these anatomical locations. According to the previous studies,^1,2 myelin content was suggested as a dominant source of phase changes in the cerebral white matter. Therefore, we suggest signal changes in the white matter in this study reflect the development of the myelin related structures in the pediatric brain. However, progression of signal changes on PADRE seen in this study differs from the pattern of myelination generally known on T1-and T2-weighted images. This suggests that the progression of phase shift in the white matter reflect not only myelination but denser myelin content along with age. Less signal changes were noted in the temoral lobe, which suggests slower development of myelin related structure, but an effect of angulation of myelin bundle to the static magnetic field are also considered.¹

Age-related progression of phase sift was found at the subcortical white matter as well as the pyramidal tract on PADRE images. These changes may reflect the development of the myelin related content.