Impact of Prematurity on the Tissue Properties of the Neonatal Brain Stem: A Quantitative MR Approach

DISCUSSION

In this study, the impact of prematurity on the tissue properties of the neonatal brain stem, determined at term-equivalent age, was investigated using the MR data postprocessing software SyMRI. Considerable correlations were observed between GA at birth and T1R/T2R, characterized by lower GA, which was associated with longer T1R/T2R and vice versa. In most cases, the physical properties of premature (GA: <37 + 0 weeks) and term infants differed significantly. Moreover, T1R/T2R determined in the midbrain enabled a reliable differentiation between extremely preterm, preterm, and term neonates. Overall, PD values showed the lowest correlations with GA at birth. Based on the data presented here, SyMRI, using quantitative MR parameters, can enable the characterization of cerebral development in different stages of prematurity and in term-born infants.

According to descriptions in the literature, myelination processes primarily lead to changes in T1R, T2R, and PD.^14,15 These alterations affect the appearance of white matter on MR contrasts and allow brain maturation to be assessed qualitatively.^1,5 However, based on the visual evaluation of conventional sequences, myelin-induced MR signal differences in preterm and term infants may be subtle and remain undetected in many cases.⁶ In addition, currently, techniques for the quantification of incomplete myelination are scarce. As described by Yakovlev and Lecours,²⁸ the human brain stem already contains a multitude of myelinated fibers at the time of birth compared with other brain areas. Hence, the physical properties of this region best reflect the current stage of brain maturity.

In this study, the strongest correlations were observed between T1R and GA at birth. It is hypothesized that T1 shortening already occurs at stages of “premyelination” due to interactions between water molecules and isolated myelin components, such as cholesterol and glycolipids.^29-31 This might explain the fact that T1R of the midbrain enabled a reliable discrimination between different stages of prematurity. Nonetheless, T1 shortening of other areas of the brain stem did not reveal similar results, though significant differences between premature and term infants were observed. As described in previous studies, iron deposition also affects the T1R and leads to specific MR signal changes during cerebral development.^32,33 Thus, more pronounced T1 shortening in the mesencephalon, caused by the initial accumulation of iron components in the red nucleus, seems credible and could explain these findings.³² In addition, melanin shows biochemical interactions with ferric iron.^32,33 However, it has been suggested that iron bound to the substantia nigra and other brain stem nuclei that contain melanin do not affect relaxation parameters before the 10th month of life.³³

Furthermore, T2R of the midbrain revealed significant differences between extremely preterm, preterm, and term neonates. T2 shortening is initiated by tightening of fully developed myelin sheaths, which becomes most evident in advanced states of brain maturation.^29-31 In extremely preterm infants, myelination is already detectable in the midbrain.¹ Predominantly, the superior cerebellar peduncles and their decussation show an advanced maturation at this stage.^1,28 Between 30 + 0–36 + 0 weeks GA, the myelin amount increases, which is detectable histologically.¹ However, myelin-related changes are hardly detected during this period using a qualitative MR imaging approach.¹ Our data indicate that quantitative MR metrics allow for the discrimination between different stages of prematurity on the basis of T2 shortening, whereas these subtle changes may not be detected visually.¹ Based on T2R of the pontine tegmentum, differences between preterm and term neonates were observed. In contrast to other regions of the developing brain, myelin deposition is not detectable before 32 weeks GA in this part of the brain stem.^1,28 Myelination progresses rapidly in this region and corroborates that T2 shortening is related to advanced states of brain maturity.^1,29-31 However, no significant differences were found on the basis of T2R of the medulla oblongata. This region shows considerable myelination even before 30 weeks GA.¹ Hence, the effect of T2 shortening on the differences between the groups may be limited. Compared with T2R, T1R indicated maturation differences more clearly. These results are consistent with the literature and may be attributed to the immediate response of T1 parameters to subtle maturation processes.^4,29-31

Compared with T1R and T2R, PD indicated neonatal brain maturity less clearly. PD values refer to the concentration of water protons and may indicate the content of water in the investigated tissue.³⁴ There is evidence that immature or demyelinated brain tissue contains higher proportions of water, which leads to increased PD values.^14,21,35 As demonstrated by Lee et al,¹⁴ PD decreases in most brain areas during regular development. However, as opposed to T1 and T2 shortening, changes in PD occur considerably later and more slowly.¹⁴ This could explain the fact that no correlations between PD and GA at birth were observed in most areas of the neonatal brain stem. Nonetheless, significant results were found in the pons, particularly in the pontine tegmentum. PD determined in the tegmentum pontis revealed significant differences between preterm and term neonates. These data might confirm that maturity-related PD changes appear in a similar manner as T1R/T2R but more slowly.¹⁴ Interestingly, compared with preterm neonates, higher PD values were detected in the midbrains of term infants. These nonsignificant observations may be attributed to the impact of the cerebral aqueduct on PD determined in the mesencephalon. There is evidence that the size of the aquaeductus mesencephali increases in the course of development, which might lead to higher PD values because of higher H₂O components.³⁶ However, descriptions are rare with regard to the proportions of the ventricular system in mature and former premature infants.

As demonstrated in a previous study, delayed myelination is detectable in preterm neonates at the regular, expected due date.⁶ The data presented in this study are in line with descriptions in the literature.⁶ The specific novelty presented in this study is the fact that distinct maturational differences between different stages of prematurity can be identified noninvasively by novel MR acquisition strategies.

Prematurity leads to structural alterations of infratentorial regions.³⁷ Furthermore, extremely preterm delivery is associated with disturbances in brain stem functions.³⁸ As described by Antinmaa et al,³⁹ impairment of the auditory pathway is highly associated with poor future verbal abilities in former preterm infants. Thus, preterm birth interferes with regular brain stem development and plays an important role in the pathogenesis of neurologic and psychiatric disorders.^37,39,40 SyMRI enables brain maturation to be quantified and therefore allows for a more precise assessment of the neonatal brain stem at early stages of development. Quantitative MR metrics represent novel, clinically feasible diagnostic tools, enabling a more differentiated assessment of functionally important maturational stages in neonatal brain development. This opens further perspectives in pediatric brain imaging and may help physicians to predict neurodevelopmental disabilities in later life. However, this was outside the scope of the present study and should be the subject of future investigations.

This study has several limitations. The sample size was small and included pathologic cases that made the data provided inappropriate as reference data. Nonetheless, this technique bears the potential to make quantitative MR baseline data available for the assessment of brain maturity in preterm and term neonates. At the date of the MR imaging examination, term-born infants were relatively older than those born prematurely. Hence, statistical methods were used that allow for correction of the effect of GA at MR imaging on the data presented. The physical properties of supratentorial regions, such as the posterior limb of the internal capsule, which is considered highly myelinated by the regular due date, were not determined.²⁸ This is attributed to the possibility of incorrect determination of T1R, T2R, and PD in participants with telencephalic brain damage. To determine brain maturity–related physical tissue properties unadulterated, participants without abnormal supratentorial findings should be investigated, which was not the case in the presented study cohort. However, the quantifiability of supratentorial brain maturity is an important topic and should be investigated in future studies. Finally, DWI and DTI have been demonstrated to be MR techniques that are sensitive enough to quantify the process of “premyelination.”^41,42 An assessment of the relationship between DTI metrics and the mapping technique presented here was outside the scope of this work and should be further addressed in the future.