Thalamic volume loss is greater in children than in adults following MCA territory arterial ischemic stroke

Abstract

Background:

Younger stroke patients may suffer worse outcomes than older patients, however the extent to which age at stroke impacts remote areas of the brain remains unclear. The objective of this study was to determine thalamic volume changes ipsilateral to middle cerebral artery territory strokes based on age at acute ischemic stroke (AIS) onset.

Methods:

AIS patients <9 years, 9–18 years, and >18 years old were retrospectively recruited from a large quaternary care system. Each subject underwent an acute (<72 hours from AIS) and chronic (> 90 days) MRI scan. Manual thalamic segmentation was performed.

Results:

Younger and older children had significantly greater stroke-side thalamic volume loss compared to adults (48.2%, p = 0.022; 40.7%, p = 0.044, respectively).

Conclusions:

Stroke-side thalamic volumes decreased across the age spectrum but to a greater degree in pediatric patients. This observation can affect functional and cognitive outcomes post-stroke and warrants further research.

Introduction

Stroke is the fifth cause of death and a leading cause of disability in the United States, with about 795,000 new acute ischemic events and 137,000 deaths every year¹. Even though the greatest stroke burden affects the older adult population, acute ischemic stroke (AIS) can occur at any stage of life. Approximately 10 to 15% of all AIS occurs in patients younger than 50 years old and about 1% occurs in the pediatric population^{2, 3}. AIS patients suffer from an increase in mortality and morbidity, including long-term neurological injuries such as cognitive and motor disability^{4, 5}. Stroke risk factors, symptoms, and recovery present differently in children and adults⁴. While stroke recovery in adults is largely influenced by stroke size and location, recovery in children is more complex⁶. Some controversy has existed over whether stroke recovery in children is influenced more by enhanced plasticity (i.e., improved recovery due to the pediatric brain’s ability to rewire or regenerate following injury) or selective vulnerability (i.e., worsened recovery due to impaired maturation of distant brain structures requiring the damaged areas for normal neurodevelopment). More recent studies have shown that younger children often experience relatively worse cognitive outcomes after AIS, with additional influence based on location of stroke^{7, 8}. Conversely, hippocampal volume on the stroke-side in children under nine years of age is relatively preserved compared to older children, implying a mechanism of enhanced plasticity⁹.

The thalamus has been observed to be affected by strokes remote from it in both adults and children^10–12. However, the extent to which age impacts post-stroke degeneration of brain structures remote from the injury remains unclear.

The purpose of this study was to determine the effect of remote ischemia on thalamic volumes over time in pediatric and adult patients with unilateral middle cerebral artery (MCA) territory AIS as a surrogate indictor of differential brain plasticity and selective vulnerability. Specifically, we aimed to determine differences in thalamic volume changes in both the stroke and contralateral cerebral hemispheres by age (ie., children younger than nine year of age, older children, and adults).

Material and Methods

Study Population

Patients were retrospectively recruited from University of Colorado Hospital and Children’s Hospital Colorado, a large quaternary health system. Subjects were eligible if they were diagnosed with an AIS involving a unilateral MCA cortical distribution between 2011 and 2017 and had undergone a magnetic resonance imaging (MRI) of the brain within 72 hours of the AIS admission and at > 90 days of initial AIS admission. To ensure thalamic volume was not influenced by other pathology, exclusion criteria included any history of brain tumors, intracranial hemorrhage with secondary AIS, history of epilepsy, new strokes (ischemic or hemorrhagic) prior to the repeat MRI scan at >90 days, AIS involving bilateral middle cerebral arteries, poor image quality, or any evidence of chronic hippocampal sclerosis on the initial AIS MRI. Additionally, neonates were not included in this study. Demographic and clinical information, including age at stroke, sex, race, ethnicity, presenting signs and symptoms, and treatment, were recorded via chart review. This study protocol was approved by the institutional review board at the University of Colorado Anschutz Medical Center (Aurora, CO).

Imaging and Segmentation Protocol

Each subject underwent an acute (<72 hours from AIS admission) and chronic (> 90 days from AIS admission) MRI scan. Scans were performed on multiple different MRI platforms with different protocols, but all included diffusion weighted imaging (DWI) and T1-weighted sequences. All acute images were reviewed by a neurologist and neuroradiologist to confirm isolated MCA AIS sparing subcortical structures. Commercially available segmentation software (Aquarius iNtuition, TeraRecon, Forest City, CA) was used to manually segment thalamic volumes¹³ on T1-weighted imaging according to published protocols, as well as volumes of the individual cerebral hemispheres and acute stroke volumes on DWI. An example of segmentation can be found in Supplemental Material Figure I. Stroke-side and contralateral-side thalamic and cerebral hemisphere volumes were measured on both the acute and chronic scans.

Volumetric Variables and Definitions

Change in volume over time was calculated for stroke-side and contralateral-side and is defined as:

$$ChangeinVolume=\frac{VolumeatChronicScan-VolumeatAcuteScan}{VolumeatAcuteScan}*100\%$$

In addition, since many of the subjects were children, we used the contralateral side, which was unaffected by AIS, to account for differences in brain size and growth over time. Stroke-side volumes were normalized to stroke/contralateral (S/C) ratios and are defined as:

$$\frac{S}{C}Ratio=\frac{StrokeSideVolume}{ContralateralSideVolume}$$

Controlled volume change (CVC) represents the change in the stroke-side structures from acute to chronic while using the contralateral structures as a control. This variable encompasses the acute and chronic measurements for the stroke-side and contralateral structures. It allows for the comparison of volume changes between age groups. CVC is defined as:

$$CVC=Chronic\frac{S}{C}Ratio-Acute\frac{S}{C}Ratio$$

Statistical Analysis

Patients were categorized into three groups based on age at the time of AIS: young children (< 9 years), older children (9 – 18 years), and adults (> 18 years). Supported by a previous study, age 9 was used as a cutoff to evenly divide the pediatric age spectrum⁹. Volumetric measurements were compared using three-way or pairwise comparisons. The acute and chronic measurements for each structure were compared using paired t-test. Variables were checked for the distributional assumption of normality using Shapiro Wilks test. Radiomic features were compared using one-way ANOVA when data was normally distributed data, otherwise Kruskal Wallis was used. Lastly, pairwise comparisons within the groups were calculated using the Tukey Honestly Significant Differences test for normally distributed data and the Dunn test otherwise. Pairwise comparison results were adjusted for false discovery rate using Benjamini Hochberg procedure. Effect size, d, was defined as difference in population means divided by pooled population standard deviation. Analyses were performed with the statistical software package R v4.1.0. P-values, p, < 0.05 were deemed as statistically significant.

Results

Study Population

Eight younger children (<9 years old), nine older children (9–18 years old), and 13 adults (> 18 years old) were included in the study (Table 1). One subject under 9 years of age and one subject over 9 years of age were excluded due to poor imaging quality. Stroke etiology in pediatric patients per the CASCADE classification¹⁴ included cardioembolic (4), other (5), bilateral cerebral arteriopathy (1), focal cerebral arteriopathy (5), small vessel arteriopathy (1), and aortic/cervical arteriopathy (1). Stroke etiology in adult patients per the TOAST criteria¹⁵ included cardioembolic (3), small-vessel occlusion (3), large-artery atherosclerosis (3), other (1), and undetermined etiology (3).

Table 1.

Cohort Demographics

	All	Younger Children	Older Children	Adults	P value
Total, n	30	8	9	13
Age at AIS, years	14.93 (2.56 −87.79)	6.77 (2.56 −8.51)	12.08 (9.14 −16.71)	70.59 (50.75 −87.79)	<0.001
Gender, n					0.073
Female	16	7	2	7
Male	14	1	7	6
Stroke side, n					0.96
Right	9	2	3	4
Left	21	6	6	9
Race/ethnicity, n					0.45
Hispanic white	5	1	3	1
Non-Hispanic black	4	1	0	3
Non-Hispanic white	20	6	6	8
Non-Hispanic Asian	1	0	0	1
More than one race	0	0	0	0
NIHSS	5.5 (0 – 21)	12 (2 – 17)	9 (0 – 21)	3(0 – 19)
Presenting signs and symptoms, n					0.36
Headache	6	1	4	1
Hemiparesis	26	8	7	11
Reduced	consciousness	7	3	3	1
Seizure	5	2	2	1
Speech deficit	21	5	5	11
Vomiting	5	2	3	0
Seizure activity > 1-month post-AIS, n	5	2	0	3	0.26
Time between scans, years	1.30 (0.26 −6.46)	2.02 (1.52 −2.97)	0.44 (0.27 −2.04)	1.22 (0.26 −6.46)	0.11
Treatment at diagnosis, n					0.15
ASA	19	5	6	8
UFH	4	3	1	0
TPA	5	0	1	4
Thrombectomy	1	0	0	1
None	1	0	1	0
Uncertain	0	0	0	0
Treatment at discharge, n					0.16
ASA	24	6	6	12
LMWH	4	2	2	0
Clopidogrel	2	0	0	2
Coumadin	3	0	0	3
None	1	0	1	0

Age, NIHSS, and time between scans measurements are reports as median and range.

NIHSS – The National Institutes of Health Stroke Scale

AIS – acute ischemic stroke

ASA- aspirin

UFH – unfractionated heparin

TPA – tissue plasminogen activator

LMWH- Low molecular weight heparin

Thalamic Volume

There was no difference in acute stroke volumes proportional to total brain volume between the three groups (Table 2). All groups experienced significant volume loss in the stroke-side thalamus between the acute and chronic MRI scans ranging from 9% to 48% volume decrease (young children p = 0.002; older children p = 0.002; adults p = 0.03, Figure 1).

Table 2.

Comparison of Volumetric Variables

	Young Children	Older Children	Adults	P value
Acute Stroke Volume (cm3)	68.64 (11.7 – 125)	119.9 (2.27 – 239)	39.29 (1.27 − 23.95)	0.26
Acute Stroke Volume (% of Total Brain Volume)	1.95 (1.56 – 9.42)	3.38 (0.23 – 10.66)	0.85 (0.12 – 2.23)	0.28
CVC	−0.48 (−0.58 – −0.27)	−0.42 (−0.59 – −0.18)	−0.05 (−0.08 – −0.008)	0.031
Change stroke-side (%)	−48.2 (−57.8 – −28.7)	−40.7 (−61.1 – −30.3)	−8.9 (−14.65 – −4.4)	0.014
Change contralateral side (%)	−0.078 (−3.1 – −4.1)	0.27 (−9.12 – 2.9)	−6.1 (−9.23 – 2.7)	0.59

Measurements reported as median and interquartile range. Pairwise p values are in Supplement.

CVC – Controlled volume change

Figure 1. Acute and chronic thalamus volumes (cm³) measured from manual segmentations.

Left boxplots display the summary statistics for the acute measurements while the right boxplots describe the summary statistics for chronic measurements. Each line connects a subject’s acute and chronic thalamus volumes. P-values are located in the bottom left corner of each plot. (A) Stroke side thalamus volumes for children under nine years old (p = 0.002). (B) Stroke side thalamus volumes for children between nine and eighteen years old (p = 0.002). (C) Stroke side thalamus volumes for adults over eighteen years old (p = 0.028). (D) Contralateral side thalamus volumes for children under nine years old (p = 0.81). (E) Contralateral side thalamus volumes for children between nine and eighteen years old (p = 0.39). (F) Contralateral side thalamus volumes for adults over eighteen years old (p = 0.53).

Young and older children experienced significant declines in S/C ratios, indicating volume loss on stroke-side when controlled for contralateral volume (young children: d = 1.56, p = 0.002; older children: d = 1.43, p = 0.003, Figure 2). Adults, however, did not experience a significant decline in S/C ratio, indicating no significant stroke-side thalamic volume loss when controlled for contralateral thalamic volume (d = 0.71, p = 0.067).

Figure 2. Acute and chronic stroke/contralateral thalamic ratios.

To account for differences in brain size and growth, stroke side thalamic measures were normalized to a S/C ratio by dividing the stroke side thalamic volume by the contralateral side thalamic volume. Left boxplots display the summary statistics for the acute S/C ratio while the right boxplots describe the summary statistics for chronic S/C ratio. Each line connects a subject’s acute and chronic ratios. P-values are located in the bottom left corner of each plot (A) S/C thalamic ratios for children under nine years old (d = 1.56, p = 0.002). (B) S/C ratios children between nine and eighteen years old (d = 1.43, p = 0.003). (C) S/C thalamic ratios for adults (d = 0.71, p = 0.067).

There was a significant difference in the CVC between younger children and adults (−0.43, p = 0.023) and older children and adult (−0.37, p = 0.044). There was no significant difference in thalamus CVC between the younger and older children. In other words, while all three age groups showed significant volume loss within the stroke-side thalamus over time, older and younger children experienced significantly greater thalamic loss compared to adults when controlling for contralateral thalamic volume. Figure 3 displays a visual comparison of acute infraction and chronic thalamic volume loss between pediatric and adult patients. Volumetric measures obtained from manual segmentation can be found in supplemental material (Table I). Pairwise comparisons between the three groups can be found in supplemental material (Table II). Supplemental figures II and III display change in thalamic volume and change in S/C ratios, respectively.

Figure 3. Comparison of thalamic volume loss in pediatric and adult patients.

Comparison of acute infarction and chronic thalamic volume loss between pediatric (A – F) and adult patients (G – L). For each pair, the left image is a DWI from the acute stroke presentation, and the right image is a FLAIR image performed at least 3 months post infarction. Black arrows indicate the acute infarction on DWI, and white arrows indicate the ipsilateral thalamus on follow-up (f/u) FLAIR.

Discussion

We found stroke-side thalamic volumes significantly decreased after 90 days of stroke onset, and the decline was more pronounced in children than in adults. Previous studies have found thalamic atrophy in adult patients in the contralateral and ipsilateral side three months after stroke¹⁶.

For decades, scientists believed that children’s brains were highly adaptable and could recover from injury through plasticity^17–19 due to the younger brain’s ability to quickly respond to injury and restore physical phenotypes²⁰. In this theory, the earlier a brain injury occurs, the better the brain recovers. However, many competing factors contribute to the recovery process^{17, 21, 22}. More recent studies have suggested that pediatric stroke recovery may be affected more by the concept of selective vulnerability, where brain development is adversely affected long-term by early injuries^{17, 18, 23}. Westmacott reported that lesion location modulated the effect of age at stroke on cognitive performance, and perinatal stroke patients performed worse on cognitive outcome measures⁷. More recently, younger age at stroke has been associated with poorer outcomes^{8, 24}. Our results indicate subcortical structures remote from MCA territory stroke are injured to a greater degree in younger patients, supporting these recent studies that report differential developmental outcomes.

The thalamus was chosen for this study because it is supplied by the posterior cerebral circulation and should not suffer direct ischemia in MCA territory AIS²⁵. However, the thalamus may be susceptible to secondary injury following an infarction in remote areas of the brain to which it is anatomically connected. Prior work in neonatal AIS patients has described acute diffusion restriction in the ipsilateral thalamus after MCA territory infarction, which the authors hypothesized represented acute secondary network injury due to excitotoxicity or deafferentation¹⁰. Therefore, the relatively greater ipsilateral thalamic volume loss experienced by the younger population of our MCA infarction patients may indicate a greater vulnerability to a similar type of network injury. Another study in neonatal AIS patients also demonstrated decreases in neonatal stroke patients’ thalamic volumes ipsilateral to the infarction. The same study also demonstrated that these patients’ motor outcomes correlated inversely with contralateral thalamic volume, indicating a role for developmental plasticity in the contralesional hemisphere¹¹.

Our findings highlight the need for further exploration into the effect of age at time of injury on long term recovery and outcomes. A better understanding of the differences in response of various subcortical gray matter nuclei after stroke and the related effect on long-term neurodevelopmental outcomes could lead to improved rehabilitative strategies.

Study limitations include small sample size, a typical problem in the study of a rare pediatric process. While there were statistically significant results, the study population size is not large enough to formally declare differences within these subcortical gray matter volumes. While all chronic scans were obtained by our definition more than 90 days after the acute stroke, differences in timing of the chronic scans could result in occasional underestimation of chronic volume loss. Additionally, we assumed that the contralateral side was unaffected. Further work with more subject numbers and longer follow-up is needed to investigate post-stroke thalamic volume differences across age groups and their impacts on functional and cognitive outcomes.

Conclusion

In patients with MCA territory AIS, stroke-side thalamic volumes decreased across the age spectrum, but the volume loss was greater in children compared to adults. This finding supports the notion that younger patients are at risk of worse neurodevelopmental outcomes following AIS.