Association of Hypoxic-Ischemic Injury of the Brain With MRI-Derived Glymphatic Function Parameters in Neonates

Abstract

Objective

To evaluate the association between hypoxic-ischemic injury (HII) of the brain and glymphatic function using MRI-derived parameters in neonates.

Materials and Methods

This retrospective, single-institution study collected brain MRI scans of 127 neonates between July 2020 and July 2022. The volume and fraction of the basal ganglia perivascular space (BG-PVS) were automatically extracted using three-dimensional T2-weighted image processing. Diffusion-tensor imaging (DTI) along the PVS (DTI-ALPS) index values were derived from the DTI maps. BG-PVS and DTI-ALPS parameters were compared between neonates with and without HII. The correlations between MRI-derived glymphatic parameters and corrected gestational age (CGA), as well as between BG-PVS measurements and the DTI-ALPS index, were analyzed using Spearman coefficients. Multivariable logistic regression adjusted for age, sex, birth weight, and mode of delivery was performed to examine the associations between each glymphatic parameter and HII.

Results

This study included 97 neonates without HII (median gestational age [GA]: 252 days) and 30 with HII (median GA: 252 days). Neonates with HII had smaller BG-PVS volumes (19 mm³ vs. 33 mm³, P = 0.001) and fractions (0.29% vs. 0.54%, P = 0.003) compared to neonates without HII. The DTI-ALPS index values did not differ significantly between neonates with and without HII (P = 0.54). CGA correlated negatively with BG-PVS measurements (ρ = -0.21 to -0.26, all P < 0.05) and positively with DTI-ALPS index values (ρ = 0.22, P = 0.014). BG-PVS measurements and DTI-ALPS index values were not significantly correlated (ρ = -0.28 to -0.08, all P > 0.05). Multivariable logistic regression revealed a negative association between BG-PVS volume (odds ratio [OR]: 0.96 per mm³ increase, 95% confidence interval [CI]: 0.93–0.99) and fraction (OR: 0.15 per % increase, 95% CI: 0.03–0.79) with HII, while DTI-ALPS index values were not significantly associated with HII (OR: 0.10, 95% CI: 0.00–25.41).

Conclusion

Neonates with HII demonstrated smaller BG-PVS volume and fraction compared with those without HII, indicating potential alterations in glymphatic function among affected newborns.

INTRODUCTION

Neonatal brain injury can arise from various processes, among which hypoxia-ischemia is a significant cause. Neonatal encephalopathy, often resulting from perinatal hypoxic-ischemic injury (HII), affects approximately 3 per 1000 live births in high-income countries [1]. Approximately 25% of neonates with hypoxia-ischemia encephalopathy exhibit subsequent neurological disabilities [2,3], ranging from cognitive deficits to behavioral abnormalities. Additionally, children with a history of neonatal HII are more likely to develop attention-deficit/hyperactivity disorder (ADHD) and autism spectrum disorders [4]. Although long-term neurodevelopmental assessment remains the gold standard for identifying impairments and disabilities, early detection and intervention may increase the likelihood of improved outcomes [5]. Thus, a key area of research has focused on identifying potential imaging biomarkers using brain MRI.

The glymphatic system, a glial cell-dependent perivascular network responsible for removing brain waste products, is critical for neuronal health and has been linked to various pathologies in children [6]. Its function relies heavily on the structural integrity of perivascular spaces (PVSs), which are pathways for fluid circulation and metabolic waste clearance [7]. Brain MRI has been used to study the glymphatic system and its components, including MRI-visible PVS volume [8] and diffusion-tensor imaging (DTI) analysis along the PVS (DTI-ALPS) [9]. The glymphatic system may play an important role in the neonatal brain because of the high metabolic demands associated with rapid growth and development [10]. Studies assessing PVS visibility and volume [11], along with DTI-ALPS index values [12], have provided insights into how glymphatic function changes with age and in association with preterm birth. A recent study also reported lower DTI-ALPS index values in neonates with birth asphyxia than in those without [13]. However, since birth asphyxia does not always lead to HII evident on brain MRI, and the previous study included only preterm neonates, evidence regarding the association of HII with the glymphatic system in both preterm and term neonates is lacking. In addition, previous studies evaluated either PVS volume [8,11] or DTI-ALPS index values [13,14], making it difficult to comprehensively assess the glymphatic system and evaluate the relationship between the two imaging biomarkers.

Therefore, this study aimed to investigate the association between HII and glymphatic function using MRI-derived parameters in neonates.

MATERIALS AND METHODS

Study Sample

This retrospective study was approved by the Institutional Review Board of Eunpyeong St. Mary's Hospital (IRB No. PC24RISI0104), which waived the requirement for informed consent. The study was conducted in accordance with the principles of the Declaration of Helsinki. The study sample comprised neonates admitted to the neonatal intensive care unit who underwent brain MRI between July 2020 and July 2022. All scans were performed for clinical purposes at the discretion of the neonatologist. The clinical indications included suspicion of brain injury, neurological symptoms, or abnormal findings on brain ultrasonography. Among 254 neonates initially identified who underwent brain MRI during the study period, we excluded 40 neonates who underwent brain MRI due to congenital abnormalities, follow-up studies, meningitis, seizures, metabolic disorders, or trauma. An additional 33 neonates were excluded because their brain MRI scans were not obtained at term-equivalent age, and 14 more were excluded because of MRI motion artifacts or the absence of DTI data. The remaining neonates were then categorized as those with or without MRI-visible HII. A flowchart of the patient selection process is shown in Figure 1.

Fig. 1. Flow chart of study group selection. NICU = neonatal intensive care unit.

In both preterm and term neonates, brain MRI findings of HII were classified by severity into profound and mild-to-moderate. In preterm neonates, profound HII typically involves the thalamus, basal ganglia (BG), and brainstem. In term neonates, profound HII affects the dorsal brainstem, anterior cerebellar vermis, thalamus, BG, corticospinal tract, and perirolandic cortex. Mild-to-moderate HII in preterm neonates is characterized by injury to the periventricular or deep white matter. In term neonates, mild-to-moderate HII involves cortical and white matter injury, which is limited to the white matter in mild cases and extends to involve both the parasagittal watershed cortex and white matter in moderate cases. Neonates with either profound or mild to moderate HII were identified and included in the study [15,16]. Neonates with minor brain injuries and hemorrhages, including small subdural hemorrhage or choroid plexus hemorrhage, were excluded. The final study sample included 127 neonates: 97 without HII and 30 with HII.

Clinical information of the neonates, including sex, gestational age (GA) at birth, corrected GA (CGA) at scan, birth weight, mode of delivery, and Apgar scores (1 and 5 minutes), was collected.

MRI Acquisition

All MRI data were obtained using a 3T MRI scanner (Magnetom Vida, Siemens Healthineers, Erlangen, Germany). The neonates were scanned without sedation using a customized 64-channel receiver coil. Three-dimensional (3D) T1-weighted images (Magnetization-Prepared Rapid Acquisition with Gradient Echo, or MPRAGE), and 3D T2-weighted images (Sampling Perfection with Application-Optimized Contrasts using Different Flip Angle Evolution, or SPACE), DTI, and susceptibility-weighted imaging (SWI) were obtained. Detailed imaging protocols are described in the Supplementary Material.

Automated Volumetric Measurements of PVSs

We first automatically segmented the brain regions using Infant FreeSurfer [17]. Frangi filtering was then performed to highlight enlarged PVS voxels [11,18]. Next, we performed multiple thresholding by adding multiples of the standard deviation to the mean in the BG region (including the caudate, putamen, and pallidum) to extract relatively high-intensity voxels. Subsequently, a researcher (D.B.) performed manually refined the segmentation under the guidance of a radiologist with 15 years of experience in pediatric neuroradiology (H.G.K.) to improve quantification accuracy. This process included removing false-positive PVS voxels, correcting false-negative PVS voxels, and determining the optimal threshold values to enhance the label consistency of the BG and PVS boundaries. The BG-PVS volume was calculated by counting the number of extracted voxels per unit voxel volume. The PVS volumes were collected from both hemispheres and their sum. The BG-PVS fraction was calculated by dividing the PVS volume by the segmented BG volume. Manual refinement was performed using ITK-SNAP version 4.2.0 [19], whereas other processes were performed using Python version 3.8 (Python Software Foundation; https://www.python.org).

DTI-ALPS Index

The DTI-ALPS index was used to assess the glymphatic system activity within the PVS by analyzing multidirectional diffusivity profiles generated from DTI data, as described previously [9,14,20]. DTI-ALPS was measured exclusively in white matter regions because this technique requires the presence of a directional fiber architecture for accurate diffusion measurements along the PVS. With SWI serving as a reference for identifying the direction of the medullary veins, regions of interest (ROIs) were placed on the projection and association fibers. Although previous studies performed DTI analysis of the BG [21], ROI placement and measurement of DTI-ALPS metrics in this region were not possible in the present study owing to the lack of a reliable reference for identifying the fiber orientation necessary for these measurements. An axial section at this level was selected on a color-coded DTI-derived fractional anisotropy (FA) map and aligned with the SWI section. On this axial image, two readers (A.C. and H.G.K.) independently placed three spherical ROIs (3–4 mm diameter) in the projection, association, and subcortical neural fiber regions of the left hemisphere. Custom software (DTI-ALPS Analyzer) developed in MATLAB 9.6 (version 2019a, The Mathworks, Inc., Natick, MA, USA) was used to automatically calculate nine distinct DTI-ALPS index values from the ROIs: 1) Dx in the projection (Dxproj), association (Dxassoc), and subcortical (Dxsubc) neural fibers, 2) Dy in the same regions (Dyproj, Dyassoc, Dysubc), and 3) Dz in the same regions (Dzproj, Dzassoc, Dzsubc). The ALPS index was derived using the following formula: ALPS index = mean (Dxproj, Dxassoc)/mean (Dyproj, Dzassoc). Reader 1 performed a second set of DTI-ALPS measurements one month after the initial assessment. The DTI-ALPS index values obtained by reader 1 were used to evaluate the association with HII and correlation with PVS parameters. The DTI-ALPS index values from Reader 1's first and second assessments were used to evaluate the intra-observer agreement, whereas indices from Readers 1 and 2 were analyzed to assess inter-observer agreement.

Statistical Analysis

Continuous variables were expressed as means ± standard deviation (SD) or medians with interquartile ranges (IQRs), while categorical variables were expressed as frequencies and percentages. Normality was assessed using the Shapiro–Wilk test. For group comparisons, we used chi-square tests for categorical variables, t-tests for normally distributed continuous variables, and Mann–Whitney U tests for non-normal or ordinal variables. Spearman's correlation analyses were performed to assess the relationships between glymphatic parameters and CGA, as well as between BG-PVS measurements and DTI-ALPS index values. We performed multivariable logistic regression analyses to evaluate the associations between HII in neonates and each of the BG-PVS measurements and the DTI-ALPS index. Three separate multivariable models were constructed using HII as the dependent variable. In Models 1, 2, and 3, the main independent variables were BG-PVS volume, BG-PVS fraction, and DTI-ALPS index, respectively. All models were adjusted for potential confounding factors including sex, GA, CGA, birth weight, and mode of delivery. For categorical variables, we specified reference categories (male vs. female for sex and cesarean vs. vaginal delivery for mode of delivery). For continuous variables, we standardized units (per mm³ for BG-PVS volume, percentage for BG-PVS fraction, and per kg for birth weight). To quantify the relationships between each variable and the outcome, adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were calculated, and statistical significance was assessed using P-values. Intra- and inter-observer agreements of the DTI-ALPS index were assessed using the intraclass correlation coefficient (ICC) with a two-way random effects model for absolute agreement. ICC values were interpreted as excellent (≥0.75), good (0.60–0.74), fair (0.40–0.59), and poor (<0.40), with 95% CIs calculated for all measurements [22]. All tests were two-tailed, with significance set at P < 0.05. Analyses were performed using R version 4.4.0 (2024, R Core Team, Vienna, Austria).

RESULTS

Characteristics of the Study Sample

The clinical and demographic characteristics are summarized in Table 1. The total study sample of 127 neonates (median GA [IQR], 252 [240–266] days, 62 males) included 97 neonates without HII (median GA [IQR], 252 [241–264] days, 46 males), and 30 neonates with HII (median GA [IQR], 252 [239–267] days, 16 males). All 30 neonates with HII had mild to moderate HII findings on brain MRI scans and included 23 neonates with white matter injury; 5 neonates with both white matter injury and subdural hemorrhage; and 2 neonates with white matter injury along with subdural hemorrhage, subarachnoid hemorrhage, or intraventricular hemorrhage. GA, CGA, birth weight, and 1-minute and 5-minute Apgar scores did not differ significantly between neonates with and without HII (all P > 0.05) (Table 1). Birth type differed between the two groups, with a higher percentage of neonates without HII born cesarean section compared with those with HII (17 [57%] vs. 85 [88%], P = 0.001).

Table 1. Characteristics of study sample.

Characteristic		Total (n = 127)	Neonates with hypoxic-ischemic injury (n = 30)	Neonates without hypoxic-ischemic injury (n = 97)	P
Sex					0.76*
	Male	62 (49)	16 (53)	46 (47)
	Female	65 (51)	14 (47)	50 (53)
Age, days
	Gestational age at birth	252 [239–266]	252 [239–267]	252 [241–264]	0.87
	Corrected gestational age at scan	267 [262–276]	266 [261–277]	267 [262–276]	0.53
Birth weight, kg		3 [2–3]	3 [2–3]	3 [2–3]	0.91
Mode of delivery					0.001*
	NSVD	25 (20)	13 (43)	12 (12)
	C-section	102 (80)	17 (57)	85 (88)
Apgar score
	1 minute	6 [5–8]	6 [4–8]	6 [5–8]	0.2
	5 minutes	8 [7–9]	8 [7–9]	8 [7–9]	0.24
BG-PVS volume, mm3		31 [18–47]	19 [13–33]	33 [20–51]	0.001
BG-PVS fraction, %		0.48 [0.28–0.73]	0.29 [0.22–0.54]	0.54 [0.32–0.78]	0.003
DTI-ALPS index		1.18 ± 0.09	1.18 ± 0.10	1.19 ± 0.09	0.54†

Data are presented as mean ± standard deviation, median [interquartile range], or number (%) unless otherwise indicated.

^*P-value is from the Chi-square test, ^†P-value is from the t-test; Others are from the Mann-Whitney U test.

NSVD = normal spontaneous vaginal deliveries, C-section = cesarean section, BG-PVS = basal ganglia perivascular space, DTI-ALPS = diffusion-tensor imaging analysis along the perivascular space

Difference in Glymphatic Parameters Between Neonates With and Without HII

BG-PVS volume and fraction differed significantly between neonates with and without HII, while the DTI-ALPS index values did not (Table 1, Fig. 2). Neonates with HII had smaller BG-PVS volumes (neonates with HII vs. neonates without HII: median [IQR], 19 [13–33] mm³ vs. 33 [20–51] mm³, P = 0.001) and BG-PVS fractions (median [IQR], 0.29 [0.22–0.54]% vs. 0.54 [0.32–0.78]%, P = 0.003) compared with neonates without HII. DTI-ALPS index values did not differ significantly between the two groups (mean ± SD, 1.18 ± 0.10 vs. 1.19 ± 0.09, P = 0.54). Representative images of neonates with and without HII are shown in Figures 3 and 4, respectively.

Fig. 2. Box plots comparing glymphatic parameters between neonates with and without HII. A-C: While BG-PVS volume (A) and BG-PVS fraction (B) differed between neonates with and without HII, the DTI-ALPS index (C) did not. HII = hypoxic-ischemic injury, BG-PVS = basal ganglia perivascular space, DTI-ALPS = diffusion-tensor imaging analysis along the perivascular space.

Fig. 3. Representative brain MRI of a female neonate with hypoxic ischemic injury (born at 38 weeks GA, MRI performed at 41 weeks CGA). A-E: Axial T2-weighted images show BG-PVS (A), with segmented areas marked in red (B). An axial T1-weighted image (C) shows hyperintense lesions in periventricular white matter (arrows), suggesting hypoxic-ischemic injury. Color-coded, DTI-derived fractional anisotropy map with ROIs (yellow circles) for DTI-ALPS (D), accompanied by a reference SWI illustrating medullary vein direction (E) are shown. The three ROIs represent projection, association, and subcortical fibers, respectively, from left to right. The diffusivity of the projection and association fibers was used to calculate the DTI-ALPS index. The segmented BG-PVS was 25 mm³ with a fraction of 0.3%. DTI-ALPS index was 1.14. GA = gestational age, CGA = corrected gestational age, BG-PVS = basal ganglia perivascular space, DTI = diffusion-tensor imaging, ROI = region of interest, DTI-ALPS = diffusion-tensor imaging analysis along the perivascular space, SWI = susceptibility-weighted imaging.

Fig. 4. Representative brain MRI of a female neonate without hypoxic ischemic injury (born at 39 weeks GA, MRI performed at 41 weeks CGA). A-D: Axial T2-weighted images show BG-PVS (A) with segmented areas marked in red (B). Color-coded, DTI-derived fractional anisotropy map with ROIs (yellow circles) for DTI-ALPS index (C), with a reference SWI showing medullar vein direction (D) are shown. The segmented BG-PVS was 39 mm³ with a fraction of 0.7%. The DTI-ALPS index was 1.17. The three ROIs represent projection, association, and subcortical fibers, respectively, from left to right. Diffusivity of the projection and association fibers was used to calculate the DTI-ALPS index. GA = gestational age, CGA = corrected gestational age, BG-PVS = perivascular space in basal ganglia, DTI = diffusion-tensor imaging, ROI = region of interest, DTI-ALPS = diffusion-tensor imaging analysis along the perivascular space, SWI = susceptibility-weighted imaging.

Correlations Between Glymphatic Parameters and CGA

To evaluate the potential maturation-related effects on glymphatic parameters, we examined the correlation between CGA and each glymphatic parameter (Fig. 5). Across all neonates, CGA showed significant negative correlations with BG-PVS volume (rho [ρ] = -0.21, P = 0.020) and BG-PVS fraction (ρ = -0.26, P = 0.004). Group analysis revealed that these correlations were significant and stronger in neonates without HII (BG-PVS volume: ρ = -0.33, P < 0.001; BG-PVS fraction: ρ = -0.35, P < 0.001), whereas we observed no significant associations in neonates with HII (BG-PVS volume: ρ = 0.07, P = 0.709; BG-PVS fraction: ρ = -0.001, P = 0.997). Additionally, CGA was positively correlated with DTI-ALPS index values across all neonates (ρ = 0.22, P = 0.014), although subgroup analyses of neonates without HII (ρ = 0.17, P = 0.099) and with HII (ρ = 0.36, P = 0.055) did not reach statistical significance.

Fig. 5. Correlations between CGA and glymphatic parameters. A: CGA vs. BG-PVS volume. CGA is negatively correlated with BG-PVS volume in all neonates and in neonates without HII but not in neonates with HII. B: CGA vs. BG-PVS fraction. CGA is negatively correlated with BG-PVS fraction in all neonates and in neonates without HII, but not in neonates with HII. C: CGA vs. DTI-ALPS index. CGA is positively correlated with DTI-ALPS index in neonates overall but not in neonates with or without HII. Blue dots and lines: neonates without HII; pink dots and lines: neonates with HII; black dashed lines: trend for all neonates combined. Linear regression lines are shown. CGA = corrected gestational age, BG-PVS = basal ganglia perivascular space, HII = hypoxic-ischemic injury, DTI-ALPS = diffusion tensor image-analysis along the perivascular space.

Correlations Between BG-PVS Measurements and DTI-ALPS Index

The relationships between BG-PVS volume and DTI-ALPS index values and between BG-PVS fraction and DTI-ALPS index values are shown in Figure 6. BG-PVS volume and DTI-ALPS index values were significantly correlated in neonates overall, neonates without HII, or neonates with HII (range of ρ = -0.28 to -0.08, all P > 0.05). Similarly, the BG-PVS fraction was not significantly correlated with DTI-ALPS index values in any of the three neonate groups (range of ρ = -0.18 to -0.15, all P > 0.05).

Fig. 6. Scatter plots illustrating the relationship between BG-PVS parameters and the DTI-ALPS index. A: BG-PVS volume vs. DTI-ALPS index. B: BG-PVS fraction vs. DTI-ALPS index. No significant correlations were observed in all neonates or in subgroups with or without HII. Blue dots and lines: neonates without HII; pink dots and lines: neonates with HII; black dashed lines: trend for all neonates combined. Linear regression lines are shown. BG-PVS = basal ganglia perivascular space, DTI-ALPS = diffusion-tensor imaging analysis along the perivascular space, HII = hypoxic-ischemic injury.

Associations Between Glymphatic Parameters and HII

The results of the multivariable logistic regression analyses of the associations between each glymphatic parameter and HII, adjusted for potential confounding factors, are presented in Table 2. In Model 1, BG-PVS volume was significantly negatively associated with HII (adjusted OR: 0.96 per mm³ increase, 95% CI: 0.93–0.99, P = 0.007). Similarly, in Model 2, BG-PVS fraction and HII showed a negative association (adjusted OR: 0.15 per % increase, 95% CI: 0.03–0.79, P = 0.026). However, in Model 3, the DTI-ALPS index values were not significantly associated with HII (adjusted OR: 0.10, 95% CI: 0.00–25.41, P = 0.411).

Table 2. Multivariable logistic regression to analyze the association between glymphatic parameters and hypoxic-ischemic injury in neonates.

Variable		OR (95% CI)	P
Model 1: BG-PVS volume
	BG-PVS Volume, mm3	0.96 (0.93−0.99)	0.007
	Sex, male	1.19 (0.46–3.08)	0.712
	Gestational age, day	0.96 (0.92–1.00)	0.055
	Corrected gestational age, day	0.97 (0.92–1.02)	0.251
	Birth weight, kg	1.64 (0.59–4.53)	0.341
	Mode of delivery, C-section	0.11 (0.03–0.37)	<0.001
Model 2: BG-PVS fraction
	BG-PVS Fraction, %	0.15 (0.03−0.79)	0.026
	Sex, male	1.21 (0.48–3.08)	0.685
	Gestational age, day	0.97 (0.93–1.01)	0.110
	Corrected gestational age, day	0.97 (0.92–1.02)	0.236
	Birth weight, kg	1.29 (0.47–3.54)	0.618
	Mode of delivery, C-section	0.11 (0.03–0.37)	<0.001
Model 3: DTI-ALPS index
	DTI-ALPS index	0.10 (0.00−25.41)	0.411
	Sex, male	1.17 (0.48–2.89)	0.729
	Gestational age, day	0.96 (0.92–1.01)	0.086
	Corrected gestational age, day	0.98 (0.93–1.03)	0.493
	Birth weight, kg	1.41 (0.53–3.81)	0.492
	Mode of delivery, C-section	0.09 (0.03–0.29)	<0.001

All models are adjusted for sex, gestational age, corrected gestational age, birth weight, and mode of delivery. BG-PVS volume is measured in mm³ and BG-PVS fraction is expressed as a percentage. For Birth weight, odds ratio represents per 1 kg increase; for mode of delivery, vaginal delivery is the reference category.

OR = odds ratio, CI = confidence interval, BG-PVS = perivascular space in the basal ganglia, C-section = cesarean section, DTI-ALPS = diffusion-tensor imaging analysis along the perivascular space

Intra- and Interobserver Agreement of DTI-ALPS Index

Intra- and inter-observer agreement analyses demonstrated the excellent reliability of the DTI-ALPS index values. Both agreements showed excellent reliability, with ICC of 0.825 (95% CI: 0.76–0.87) and 0.868 (95% CI: 0.82–0.91), respectively.

DISCUSSION

Brain glymphatic function may be altered following HII in neonates. However, the associations of imaging biomarkers such as BG-PVS measurements and DTI-ALPS index with HII have not been studied. Moreover, the relationship between these two imaging markers has also not yet been investigated. Neonates with HII had lower BG-PVS volume (P = 0.001) and fraction (P = 0.003) than those without HII. In all neonates, CGA showed significant negative correlations with BG-PVS volume (P = 0.020) and BG-PVS fraction (P = 0.004) and a significant positive correlation with the DTI-ALPS index (P = 0.014). We observed no significant correlations between BG-PVS measurements and DTI-ALPS index values (ρ = -0.28 to -0.08, all P > 0.05) across all neonates. The results of the multivariable logistic regression revealed a negative association between BG-PVS measurements and HII.

Our study findings revealed significant differences in glymphatic system parameters of BG-PVS volume and fraction between neonates with and without HII, indicating that HII may alter the glymphatic system. This result aligns with a previous study showing smaller BG-PVS volumes among preterm neonates compared with controls [11]. The finding of a smaller BG-PVS volume in diseased neonates contrasts with findings in older children, who exhibit larger PVS volumes in disorders such as ADHD [23] and autism [24]. Therefore, altered glymphatic function may affect PVS volume differently depending on patient age. In the neonatal brain, a smaller BG-PVS volume can be attributed to abnormally high expression of aquaporin-4, which mediates the influx of cerebrospinal fluid into the PVS. In addition, ischemic injury can lead to astrocyte swelling [25] which may result in the dysfunction of the immature glymphatic system.

The median PVS volume of control neonates in our study (median 33 mm³, IQR, 20–51 mm³) differs from that reported in a previous study [11] using the same method of automatic quantification (median 48 mm³, IQR, 21.2 mm³). The most likely explanation for this discrepancy is a difference in imaging protocols. Although the spatial resolution setting of the 3D T2-weighted scan was identical to that of the Developing Human Connectome Project (dHCP) dataset (0.8 mm³), the actual resolution, considering the voxel spread function and image contrast, differed. Even when the nominal resolution is the same, variations in repetition time, echo time, and acceleration factors can influence the effective image resolution. Additionally, the signal distribution may vary depending on the RF coil and image reconstruction method, which can affect the segmentation outcome, particularly in threshold-based approaches. Given these factors, using relative differences and within-dataset analyses may be more appropriate than directly comparing PVS volumes across datasets unless segmentation methods are standardized and generalized to ensure higher reproducibility. In future multicenter or longitudinal studies, harmonization of these imaging parameters is essential.

BG-PVS parameters but not DTI-ALPS index values differed between neonates with and without HII. This finding contrasts with a previous study that reported lower DTI-ALPS index values in neonates with birth asphyxia compared with those without asphyxia (0.98 ± 0.08 vs. 1.08 ± 0.07, P < 0.05) [13]. This difference can be attributed to several factors, including variations in the study samples and MRI timing. First, our neonates had a median GA of 36 weeks, whereas the previous study included more preterm neonates, with a median GA of 29–30 weeks. Second, the timing of the MRI acquisition differed. Our neonates were scanned at a term-equivalent age, whereas the previous study conducted MRIs at a median age of 31 weeks. At term-equivalent ages, the BG and white matter are at different maturation stages [26]. Regional differences in vulnerability to HII have also been documented [16]. These developmental and anatomical differences may explain why the DTI-ALPS index, measured in the white matter, showed lower sensitivity than the BG-PVS measurements from the BG. Our study included both preterm and term neonates within each group, covering a range of brain maturation and HII susceptibilities. Because the sensitivity of glymphatic imaging markers may vary with developmental stage, the results may differ depending on the characteristics of the neonatal population. Our results suggested that BG-PVS volume or fraction values may be more sensitive markers for detecting glymphatic alterations in neonates who undergo brain MRI at term-equivalent age, even after adjusting for GA and CGA.

The results of the correlation analyses in the present study revealed significant associations between CGA and glymphatic parameters, particularly in neonates without HII. These findings are consistent with previous studies suggesting that glymphatic system development parallels brain maturation during the perinatal period [11,27]. In our study, BG-PVS volume and fraction were negatively correlated with CGA, whereas the DTI-ALPS index showed a positive correlation. Since both BG-PVS volume and DTI-ALPS index were evaluated in the same individuals, our results provide further evidence of an inverse relationship between these two MRI-based markers of glymphatic function. These two parameters also show a distinct age-related pattern compared with older populations, in which BG-PVS volume tends to increase with age [11], while the DTI-ALPS index shows a negative association [27].

The results of the multivariable logistic regression analyses revealed a significant association between BG-PVS parameters and HII in neonates after adjusting for sex, GA, CGA, birth weight, and mode of delivery. The initial group comparison showed no significant demographic differences between neonates with and without HII, except for mode of delivery. After adjusting for these potential confounding variables, the association between BG-PVS measurements and HII remained significant, suggesting that BG-PVS parameters may serve as independent imaging markers associated with HII. In contrast, the DTI-ALPS index values showed no significant association with HII after adjusting for confounding variables. Longitudinal studies with early measurements of glymphatic function parameters and subsequent outcome assessments are required to establish the predictive value of neonatal glymphatic function for HII. Additionally, as HII may result in neurodevelopmental disorders [2,3], it is important to investigate whether BG-PVS measurements influence the likelihood of developing such disorders following HII.

Based on prior research [11,14] showing promising associations between BG-PVS measurements and DTI-ALPS index values and age in neonates, we hypothesized that these measures are related. However, we observed no significant correlations between them, contrary to findings in adults with Parkinson's disease, where decreased DTI-ALPS index values were correlated with increased PVS burden [28]. As discussed above, regional differences in the developmental stages of the glymphatic system may influence the relationship between these two parameters, which differ from those observed in mature brains [16]. Additionally, while DTI-ALPS measurements in adult studies used ROIs 5 mm in diameter, the smaller fiber tracts in the neonatal brain limited the ROI sizes to 3–4 mm. Although our inter-observer agreement was excellent, the smaller ROI size may have introduced methodological instability in the DTI-ALPS metrics. As our results showed inverse relationships between these two parameters according to CGA, future studies with larger sample sizes, refined grouping, and better control of covariates may yield different outcomes regarding the relationship between BG-PVS and DTI-ALPS parameters.

Our study has certain limitations. First, as this was a cross-sectional retrospective study, establishing a causal relationship between MRI-derived glymphatic markers and HII in neonates was challenging. Although HII may result in smaller BG-PVS measurements, this relationship should be further explored in studies assessing glymphatic parameters earlier in the perinatal period. Second, the PVS was measured only in the BG and not in the white matter, as discussed above. This restricted regional assessment limited our ability to comprehensively characterize the glymphatic system throughout the neonatal brain. Consequently, whether white matter PVS plays a distinct role, undergoes different alterations in glymphatic function in neonates, or has a relationship with DTI-ALPS index values remains unclear. Future studies should aim to develop and validate techniques for evaluating white matter PVS and whole brain glymphatic metrics in neonates.

In conclusion, HII in neonates is associated with smaller BG-PVS volumes and fractions. The DTI-ALPS index values were not significantly associated with HII, and no correlation was observed between BG-PVS measurements and DTI-ALPS index values. These findings suggest that the glymphatic function may be altered in association with HII in neonates. Our results also indicate distinct roles for BG-PVS and DTI-ALPS parameters in neonatal brain pathology, implying that they may reflect different aspects of glymphatic function. Future longitudinal studies are required to explore how these markers evolve with maturation, and their potential value for predicting neonatal neurodevelopmental outcomes.

Availability of Data and Material

The datasets generated or analyzed during the study are not publicly available due to Institutional Review Board (IRB) restrictions to protect patient privacy but are available from the corresponding author on reasonable request.

Supplement

The Supplement is available with this article at https://doi.org/10.3348/kjr.2025.0300.