Magnetic resonance imaging assessment of brain maturation in preterm neonates with punctate white matter lesions

Transcription

1 Neuroradiology (2007) 49: DOI /s y PAEDIATRIC NEURORADIOLOGY Magnetic resonance imaging assessment of brain maturation in preterm neonates with punctate white matter lesions Luca A. Ramenghi & Monica Fumagalli & Andrea Righini & Laura Bassi & Michela Groppo & Cecilia Parazzini & Elena Bianchini & Fabio Triulzi & Fabio Mosca Received: 27 March 2006 / Accepted: 16 October 2006 / Published online: 22 November 2006 # Springer-Verlag 2006 Abstract Introduction Early white matter (WM) injury affects brain maturation in preterm infants as revealed by diffusion tensor imaging and volumetric magnetic resonance (MR) imaging at term postmenstrual age (PMA). The aim of the study was to assess quantitatively brain maturation in preterm infants with and without milder forms of WM damage (punctate WM lesions, PWML) using conventional MRI. Methods Brain development was quantitatively assessed using a previously validated scoring system (total maturation score, TMS) which utilizes four parameters (progressive myelination and cortical infolding, progressive involution of glial cell migration bands and germinal matrix tissue). PWML were defined as foci of increased signal on T1-weighted images and decreased signal on T2-weighted images with no evidence of cystic degeneration. A group of 22 preterm infants with PWML at term PMA (PWML group) were compared with 22 matched controls with a normal MR appearance. Results The two groups were comparable concerning gestational age, birth weight and PMA. TMS was significantly lower in the PWML group than in the control group (mean TMS 12.44±2.31 vs 14.00±1.44; P=0.011). Myelination (mean 2.76±0.42 PWML group vs 3.32±0.55 control L. A. Ramenghi : M. Fumagalli (*) : L. Bassi : M. Groppo : F. Mosca Neonatal Intensive Care Unit Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena, IRCCS, University of Milan, Milan, Italy monifumagalli@hotmail.com A. Righini : C. Parazzini : E. Bianchini : F. Triulzi Department of Radiology and Neuroradiology, Ospedale Pediatrico Buzzi -ICP, Milan, Italy group, P=0.003) and cortical folding (3.64±0.79 vs 4.09± 0.43, P=0.027) appeared to be significantly delayed in babies with PWML. Conclusion Conventional MRI appears able to quantify morphological changes in brain maturation of preterm babies with PWML; delayed myelination and reduced cortical infolding seem to be the most significant aspects. Keywords Preterm. White matter disease. Brain maturation. Magnetic resonance imaging Introduction A wide spectrum of lesions affect the brain of preterm infants, and more recent applications of magnetic resonance imaging (MRI) have been relevant in detecting morphological and volumetric differences of the brain of premature newborns at term postmenstrual age (PMA) compared to term babies. MRI optimization allows the detection of more discrete abnormalities of brain development resulting from impairment of complex maturation processes due to uncommon lesions of prematurity. In addition this technique has proved particularly effective in detecting the milder spectrum of white matter (WM) injury of prematurity compared to ultrasonography [1 3]. Punctate WM lesions (PWML) are frequently identified on brain MRI of preterm infants as small areas of increased signal intensity on T1-weighted images and decreased on T2-weighted images. These WM lesions, often with a punctate appearance, have recently received increasing attention with regard to the neurodevelopmental outcome. Caution is needed when interpreting isolated punctate lesions seen on conventional MRI as their clinical significance seems to be benign for motor problems but more

2 162 Neuroradiology (2007) 49: uncertain, and not yet well established, for longer cognitive follow-up [1, 4, 5]. Diffusion tensor imaging (DTI) and volume MR imaging studies in term PMA infants have shown that early WM injury affects WM maturation in preterm infants [6, 7]. Several scoring systems have been proposed to quantify cerebral maturation of the preterm brain on conventional MRI [8]. Childs et al. [9] validated a method based on the observation of standard MR images which was demonstrated to be reproducible and easy to calculate. The aim of this study was to assess quantitatively brain maturation in preterm infants at term PMA, with and without PWML using a previously standardized and validated scoring system which relies on conventional MRI. The hypothesis to be tested was that PWML, detected by conventional MRI, may impair brain morphological maturation in preterm babies. Methods According to the clinical protocol of the department, the brains of premature infants (with gestational age, GA, 33 weeks) admitted to the Neonatal Intensive Care Unit were imaged by conventional MRI at term PMA. Informed parental consent was obtained to enroll patients into the study. The babies were sedated for imaging with oral chloral hydrate (40 60 mg/kg) and oxygen saturation was monitored by pulse oximetry throughout the procedure. MRI was performed at 1.5 T according to the internal standard protocol for imaging newborn babies, which included axial 3 mm thick T2-weighted fast spin-echo (TR/TE 6000/200 ms, FOV 180 mm, matrix ) and T1-weighted spin-echo (TR/TE 600/15 ms, FOV 180 mm, matrix ) sections. PWML were defined as increased T1 signal abnormalities (Fig. 1a) and decreased T2 signal abnormalities (Fig. 1b) in the cerebral WM with no evidence of cystic degeneration. Babies with cystic periventricular leukomalacia on ultrasound scan or MRI were excluded. The PWML group comprised 22 subjects with PWML identified from a population of 104 preterm babies scanned between October 2003 and December The control group comprised 22 matched normal infants obtained by enrolling the first preterm patient with a normal MR appearance scanned soon after a patient showing PWML. GA and birth weight (BW) were calculated for each patient enrolled in the PWML and control groups. GA and BW were compared between the two groups. The linear regression coefficient was calculated for each group to show the distribution of total maturation score (TMS) according to PMA. The presence of diffuse excessive high signal intensity (DEHSI) on T2-weighted images in the two populations of babies was also evaluated. Scoring system Brain development was quantitatively assessed on MR images using a previously validated scoring system [9] which utilizes four parameters of cerebral maturation giving an increasing score in preterm infants: two phenomena undergoing progressive maturation (myelination and corti- Fig. 1 PWML as detected by MRI: foci of increased T1 signal (a) and decreased T2 signal (b) in the cerebral WM with no evidence of cystic degeneration

3 Neuroradiology (2007) 49: Table 1 Scoring system to assess four parameters of cerebral maturation in preterm newborn infants [9] Parameter Score Definition Myelination (M) M1 Myelination evident in brainstem, cerebellar peduncle, inferior colliculus, cerebellar vermis M2 + Subthalamic nuclei, globus pallidus, ventrolateral thalamus M3 + Caudal portion of the posterior limb of the internal capsule (PLIC) M4 + Complete PLIC M5 + Optic radiation M6 + Corona radiata M7 + Anterior limb of internal capsule Cortical infolding (C) C1 Frontal and occipital cortex completely smooth, insula wide open; thin bright cortical rim on T1-weighted images, generally low-intensity white matter (WM) on T1-weighted images C2 Frontal cortex still very smooth, some sulci evident in occipital cortex; insula still wide with almost smooth internal surface; WM low intensity on T1-weighted images C3 Frontal and occipital cortex similar number of convolutions; frontal sulci still quite shallow; internal surface of insula more convoluted; WM still somewhat low intensity on T1-weighted images C4 Frontal and occipital cortex folded and rich in sulci; frontal sulci obvious along interhemispheric fissure; occipital WM separated into strands by deeper sulci; Insula more convoluted and infolded; WM still slightly low intensity on T1-weighted images C5 Frontal and occipital WM separated into strands by deeper sulci; insula completely infolded; WM still distinguishable from gray matter on T1-weighted images C6 As C5 but WM now isointense with gray matter on T1-weighted images Germinal matrix (G) G1 Matrix seen in posterior horn, at caudothalamic notch (CTN) and anterior horns of lateral ventricles G2 Matrix evident at CTN and anterior horns only G3 Matrix at anterior horns alone G4 No matrix evident Bands of migrating glial B1 Broad band with additional narrower bands cells (B) B2 Broad band alone B3 Narrow band alone B4 No bands seen cal infolding), and two structures undergoing progressive involution (glial cell migration bands and germinal matrix tissue) (Table 1). A single score was determined for each of the four parameters (value ranges: 1 to 7 for myelination, 1 to 6 for cortical infolding, 1 to 4 for germinal matrix, and 1 to 4 for bands of migrating cells) and the TMS was then calculated as the sum of the four different scores (Fig. 2). As previously demonstrated, the scores derived using this scoring system for assessing brain development are Fig. 2 a Axial T2-weighted image from an infant of the control group with a TMS of 14. The caudal portion of the posterior limb of the internal capsule, score 3; cortical infolding, score 4; matrix at anterior horns alone, score 3; no bands of migrating cells, score 4. b Axial T2-weighted image from an infant of the PWML group with a TMS of 12. Myelination, score 2.5 (averaged with T1); frontal sulci still quite shallow, score 3; germinal matrix, score 3; narrow band alone, score 3

4 164 Neuroradiology (2007) 49: significantly related to PMA the score increases with the progressive maturation of the brain [9]. Myelination and germinal matrix distribution were derived from assessment of the entire T1- and T2-weighted axial images, while cortical infolding and glial cell migration scores were obtained from evaluation of the T1- and T2-weighted images closest to the foramen of Monro. Myelination was scored separately on T1- and T2-weighted images and an average value used. The calculation of the TMS was performed by L.A.R. and M.F. in each single patient following routine practice used for each very low BW baby scanned with MRI. To establish and confirm the reproducibility of the TMS method as performed by the operators of our unit, MR images were reviewed independently by two experienced investigators (L.A.R. and M.F.), who performed all the observations, and the level of agreement between the two observers was calculated by intraclass correlation (kappa statistic). The mean values of the assessments by L.A.R. and M.F. were calculated and used in the statistical analysis. Statistical analysis The unpaired t-test was used to determine the significance of differences between the two groups. The interobserver variation in the TMS determined by L.A.R. and M.F. was estimated using the kappa statistic. Linear regression was used to show the distribution of TMS according to PMA. The chi-squared test was used to analyze the differences in the number of DEHSI in the two groups. All statistical analyses were performed using SigmaStat Statistical Software version 2.0. Results The PWML group comprised 22 preterm infants and the control group 22 infants with normal MRI. The two groups were homogeneous in terms of GA at birth, BW and PMA at MR scan. The clinical data of the two groups are shown in Table 2. TMS was significantly lower in the PWML group than in control group (mean±sd TMS 12.44±2.31 vs 14.00±1.44, P=0.011; Fig. 3). When considering the single parameters of Table 2 Clinical data of the two groups. Values are medians (range) PWML group Control group P value Gestational age (weeks) 30 (26 33) 31 (26 33) 0.6 Birth weight (g) 1390 ( ) 1225 ( ) 0.4 PMA at MR scan (weeks) 39 (35 45) 40.5 (38 43) 0.6 Fig. 3 TMS (presented as the average values of the scores of L.A.R. and M.F.) in babies with PWML (PWML group, white circles) and matched normal controls (Control group, gray circles). Each symbol is an individual value. In two babies with PWML the germinal matrix distribution and the bands of migrating cells could not be evaluated because the MR images were of low quality; TMS was therefore calculated in 20/22 babies. The solid lines show the mean values. The difference between the two groups was statistically significant (P=0.011) TMS separately, myelination appeared to be significantly delayed in babies with PWML (mean±sd 2.76±0.42 PWML group vs 3.32±0.55 control group, P=0.003) as well as cortical infolding (3.64±0.79 PWML group vs 4.09± 0.43 control group, P=0.027). No statistically significant differences were observed in germinal matrix distribution or in appearance of bands of migrating cells between the two groups (Fig. 4). The interobserver coefficient of variation was A scatter graph showing the distribution of TMS according to PMA is given for each group of babies (Fig. 5) and the linear regression coefficients (r) were in the PWML group and in the control group. The difference in the number of DEHSI between the two groups was not significant (12/22 in the PWML group vs 5/22 in the control group; P=0.063). Discussion Our study shows how conventional MRI based on T1- and T2-weighted images is able to assess the delay in the cerebral morphological maturation of preterm babies with PWML compared to normal preterm controls at term PMA.

5 Neuroradiology (2007) 49: Fig. 4 Mean values of myelination (circles), cortical folding (squares), germinal matrix distribution (triangles) and bands of migrating glial cells (diamonds) in the PWML group (open symbols) and control group (filled symbols) ( n=22 vs n=22, P=0.003; n=22 vs n=22, P=0.027; n=20 vs n=22, P=0.203; n=20 vs n=22, P=0.622). Values are means±sd In addition, by considering each single variable (germinal matrix, glial cell migration bands, myelination, cortical infolding) of the TMS used to grade the different levels of brain development, it appears that PWML mainly affect myelination and cortical infolding maturation. The capacity of MRI to follow the level of myelination during prenatal development has been well documented previously [10], but the delay in myelination observed in our babies with PWML suggests some observations. From weeks GA up to weeks postnatal age there is a natural slowing of the myelination process as the major milestones achieved over such a long time are essentially the posterior limbs of the internal capsula and the corona radiata. This period of time is nearly equivalent to the median postnatal age of the two groups of babies studied (8 10 weeks) and the differences observed in the myelination score between the two groups may therefore have an even more significant implication. The mean value of myelination Fig. 5 Scatter graphs showing the distribution of TMS according to PMA in the PWML group (a) and the control group (b). The linear regression coefficients are also shown

6 166 Neuroradiology (2007) 49: in the PWML group indicates that in these babies the posterior limbs of the internal capsule are not myelinated, and myelination is a cornerstone of maturity used as a positive prognostic factor in many other conditions [11 13]. The progressive involution of the germinal matrix and of bands of migrating glial cells is unlikely to be impaired by PWML. We remain unclear as to the reasons of this observation, although we postulate that it could be related to the preferential location of WM damage in the posterior periventricular areas and above the corona radiata. On the other hand, around 30 weeks of GA, germinal matrix and bands of migrating glial cells are located mainly in the frontal-anterior part of the brain and thereby their progressive disappearance should not be affected by PWML. The investigators who calculated the TMS following routine practice in very low BW infants were unaware whether the baby did or did not have PWML although more obvious WM abnormalities were detectable in five patients. In these babies the process of scoring might be considered to have been unintentionally influenced; however, bias due to the appearance of WM abnormalities should have affected all the parameters included in the TMS while our results show that two of them (disappearance of germinal matrix and bands of migrating glia cells) were not significantly different between the two groups of patients, with and without PWML. Interestingly, the results obtained in this study are consistent with previous reports suggesting that early and more severe forms of WM injury in preterm babies may affect WM myelination and cortical gray matter development at term as detected by DTI and three-dimensional MRI [6, 7]. Reduction in the volume of brain myelin [7] and abnormalities of water motion, detected by DTI [6], probably reflecting delay in myelination or axonal damage, have been described in these studies. Inder et al. [7] also reported a reduction in cortical gray matter volume in babies with periventricular WM injury, suggesting that WM damage may cause a disturbance in cortical neuronal development. Using specific histopathological methods, Marin-Padilla [14] observed alterations in the morphology and organization of neurons in the cerebral cortex overlying areas of periventricular leukomalacia, probably resulting from destruction of corticopetal, corticofugal and association fibers leading to input deprivation and output isolation of the overlying gray matter and, as a consequence, impaired neuronal differentiation. According to another hypothesis, the tension-based theory of morphogenesis and compact wiring in the central nervous system of Van Essen, tension along axons in the WM can explain how and why the cortex folds in a characteristic species-specific pattern [15]. WM injury might therefore impair this tension-based morphogenesis and alter cerebral gyral development. This explanation appears reasonable for severe periventricular leukomalacia leading to cystic lesions of the WM, but is less convincing for those milder abnormalities of the WM such as those we studied. We excluded from the present study babies with cystic periventricular leukomalacia on sonography and/or MRI as the aim of the study was to compare quantitative brain maturation in preterm infants, at term PMA, with and without milder forms of WM damage such as PWML, as defined by increased T1 signal and decreased T2 signal abnormalities in cerebral WM with no evidence of cystic degeneration. According to recent observations, by imaging babies at term PMA and therefore with a postnatal age of about 9 10 weeks, we may have underestimated the problem of the mildest forms of PWML, as it has been noticed that very minor WM signal abnormalities may disappear in a relatively short time [4]. In other words, the possible transient nature of these lesions may suggest that the babies showing PWML at term equivalent were those with the more severe form of the disease. The distribution of TMS and PMA may have been different in the two groups of babies (Fig. 5). PWML seem to result in a TMS that is more dependent on PMA than in normal control babies whose higher TMS achieved at term equivalent does not appear to be so related to PMA. This finding may reflect a more homogeneous and rapid ex-utero brain maturation in the normal compared to PWML babies. The presence of DEHSI, a supposed more diffuse white matter abnormality, was not significantly different between the two groups of babies, although a trend towards a higher number in the PWML group was observed. We remain uncertain as to the meaning of this finding but the potential role of DEHSI in impairing brain maturation needs to be addressed by specific studies [16]. In conclusion, even with conventional and routine MRI sequences we were able to quantify a delay in maturation of brain structures in those premature babies suffering minor degrees of WM damage as observed with more sophisticated MR techniques. We think this observation represents an important issue as a high percentage of premature babies show minor cognitive problems later in life. The vulnerable and potentially reduced WM of premature babies can be the key point to understanding these later problems. The impairment in WM maturation is likely to produce an alteration in many developing processes of the premature brain, including synaptogenesis and the phenomenon of connectivity which allows many different parts of the cortex to be connected with other cortical areas. This may explain why even mild abnormalities of WM, not severe enough to cause motor problems, may be able to produce minor cognitive problems frequently seen in children who were premature babies when they reach school age. Longer follow-up studies with a complex assessment of cognitive function are needed in these premature babies

7 Neuroradiology (2007) 49: with PWML in order to better investigate the correlation between delay in myelination and cortical maturation and the frequent neurological problems seen in children who were preterm babies. Conflict of interest statement of interest. References We declare that we have no conflict 1. Childs AM, Cornette L, Ramenghi LA, et al (2001) Magnetic resonance and cranial ultrasound characteristics of periventricular white matter abnormalities in newborn infants. Clin Radiol 56: Miller SP, Cozzio CC, Goldstein RB, et al (2003) Comparing the diagnosis of white matter injury in premature newborns with serial MR imaging and transfontanel ultrasonography findings. AJNR Am J Neuroradiol 24: Inder TE, Anderson NJ, Spencer C, Wells S, Volpe JJ (2003) White matter injury in the premature infant: a comparison between serial cranial sonography and MR findings at term. AJNR Am J Neuroradiol 24: Cornette L, Tanner SF, Ramenghi LA, et al (2002) Magnetic resonance imaging of the infant brain: anatomical characteristics and clinical significance of punctate lesions. Arch Dis Child Fetal Neonatal Ed 86:F171 F Dyet LA, Kennea NL, Counsell P, et al (2004) Punctate white matter abnormalities on magnetic resonance imaging of the brain in preterm infants and neurodevelopmental outcome. Pediatr Res Abstract Miller SP, Vigneron DB, Henry RG, et al (2002) Serial quantitative diffusion tensor MRI of the premature brain development in newborns with and without injury. J Magn Reson Imaging 16: Inder TE, Huppi PS, Warfield S, et al (1999) Periventricular white matter injury in the premature infant is followed by reduced cerebral cortical gray matter volume at term. Ann Neurol 46: Hüppi PS, Schunknecht B, Boesch C, et al (1996) Structural and neurobehavioural delay in postnatal brain development of preterm infants. Pediatr Res 39: Childs AM, Ramenghi LA, Cornette L, et al (2001) Cerebral maturation in premature infants: quantitative assessment using MR imaging. AJNR Am J Neuroradiol 22: Counsell SJ, Maalouf EF, Fletcher AM, et al (2002) MR imaging assessment of myelination in the very preterm brain. AJNR Am J Neuroradiol 23: De Vries LS, Groenendaal F, van Haastert IC, Eken P, Rademaker KJ, Meiners LC (1999) Asymmetrical myelination of the posterior limb of the internal capsule in infants with periventricular haemorrhagic infarction: an early predictor of hemiplegia. Neuropediatrics 30: Rutherford MA, Pennock JM, Counsell SJ, et al (1998) Abnormal magnetic resonance signal in the internal capsule predicts poor neurodevelopmental outcome in infants with hypoxic-ischemic encephalopathy. Pediatrics 102(2 Pt 1): Roelants-van Rijn AM, Groenendaal F, Beek FJ, Eken P, van Haastert IC, de Vries LS (2001) Parenchymal brain injury in the preterm infant: comparison of cranial ultrasound, MRI and neurodevelopmental outcome. Neuropediatrics 32: Marin-Padilla M (1997) Developmental neuropathology and impact of perinatal brain damage. 2. White matter lesions of the neocortex. J Neuropathol Exp Neurol 56: Van Essen DC (1997) A tension-based theory of morphogenesis and compact wiring in the central nervous system. Nature 385: Counsell SJ, Allsop JM, Harrison MC, et al (2003) Diffusionweighted imaging of the brain in preterm infants with focal and diffuse white matter abnormality. Pediatrics 112(1 Pt 1):1 7