Intracranial germ cell tumors (GCT), while rare in the West, are more common in Asia accounting for up to 15% of primary intracranial tumors in Asian children. Germinomas are the most common subtype of GCT. Despite of the excellent survival rate for intracranial GCT, the relatively large volumes and high doses of radiation required to treat intracranial GCT can impact neurocognitive and functional outcomes. Treatment-related neurotoxicity was shown to be particularly harmful to specific cognitive processes, such as executive functions (1, 2). Furthermore, a recent study reported that the intelligence quotient (IQ) of nearly 20% of pediatric intracranial GCT survivors was borderline or worse (3).

Because of the high survival rates of childhood intracranial GCT, there is growing concern over the long-term neurocognitive outcomes. Among the different types of intracranial germinomas, basal ganglia germinomas tended to result in the worst neurocognitive outcomes (4). It is believed that the deep grey matter of children is more susceptible to oxidative damage. Damage to the basal ganglia and deep brain nuclei can lead to cognitive impairment as well as loss of motor control (5). Majority of basal ganglia germinomas with poor neurocognitive outcomes were also associated with early cerebral atrophy (6). It was proposed that atrophy occurs as a result of tumor involvement of the internal capsule fibers or thalamic ganglion cells, with Wallerian degeneration and subsequent interruption of thalamocortical connections leading to detrimental effects on motor function and functional outcomes (7). It is thus conceivable that the cognitive and functional impairments in intracranial GCT survivors are likely due to both grey and white matter damage.

Numerous neuroimaging studies have investigated the association between brain changes and neurocognitive and functional outcomes in brain tumor survivors (8, 9). However, previous studies suggested that brain volume or volumetric white matter changes might be a poor predictor of neurocognitive deficits in the later life of cancer survivors (9–11). Conventional MRI methods might not be sensitive enough to identify changes that are predictive of neurocognitive dysfunction.

We have previously demonstrated that fractional anisotropy (FA), a diffusion metric commonly obtained from diffusion tensor imaging (DTI), correlated with the cognitive outcomes of medulloblastoma survivors (12, 13). Diffusion tensor imaging is a diffusion MRI technique that measures the diffusion of water molecules, a process that is very sensitive to tissue microstructural changes. Fractional anisotropy and mean diffusivity (MD) are the two most common diffusion metrics obtained from DTI. Notably, neurological/functional impairments were found to be more related to white matter injury, whereas cognitive impairments were associated with both white and grey matter injuries (14). So far, there have been no investigations on the relation between grey matter changes and cognitive outcomes in brain tumor survivors.

In the past few years, diffusional kurtosis imaging (DKI), an extension to DTI, has been shown to be more sensitive to microstructural changes in white and grey matter in adult neurological diseases such as stroke (15, 16), traumatic brain injury (17), and Alzheimer’s disease (18). We therefore hypothesized that diffusion metrics obtained from DKI would be more sensitive to the white and grey matter changes of intracranial GCT survivors, and these metrics might be associated with functional and neurocognitive outcomes. The aim of this study was to investigate whether white and grey matter microstructural changes as measured by DKI were associated with functional and neurocognitive outcomes in childhood intracranial GCT survivors.

We conducted a cross-sectional study to determine if changes in brain microstructure would correlate with the neurocognitive and functional outcomes of childhood intracranial GCT survivors with a history of cranial radiotherapy. This study was approved by the ethical committees of the Institutional Review Board of the University of Hong Kong and Hospital Authority Hong Kong West Cluster, Kowloon West and Kowloon Central Clusters, and New Territories East and New Territories West Clusters. Informed written consent was obtained from the parents of participants under the age of 18 years or from participants aged ≥ 18 years. Neurocognitive outcomes were assessed using the Full Scale Intelligence Quotient (FSIQ) and functional outcomes were assessed using Lansky/Karnofsky performance scales.

We have successfully demonstrated using diffusional kurtosis imaging that there were extensive microstructural changes in the grey and white matter of intracranial GCT survivors with a history of cranial irradiation when compared to healthy controls. The majority of these changes occurred in brain regions related to cognition and motor functions. Our study provides evidence that these extensive microstructural changes might underlie deficits in cognitive and motor functions, as well as problems with sensory processing. Our study also demonstrated that it is important to look at the microstructural changes of individual brain regions rather than just looking at the whole brain. Brain regions affected by the tumor or focal radiotherapy will be prone to more severe microstructural damages. Therefore, despite of the lack of association between IQ or KPS scores and microstructural changes in the whole brain grey matter, we found significant associations between the microstructural changes of specific grey matter regions and the IQ and KPS scores, which reflect cognitive and functional outcomes respectively.

A recent study showed that childhood leukemia survivors with history of cranial radiotherapy had altered diffusion metrics in the fornix, uncinate fasciculus, and cingulum, which are important structures that subserve episodic memory, learning, and attention (22). Leung et al. found significant change in the FA of the posterior thalamic radiation of medulloblastoma survivors after cranial irradiation and chemotherapy (23). We found GCT survivors had significant microstructural changes in these brain regions, as well as in projection fibers such as the corona radiata, posterior thalamic radiation, and superior longitudinal fasciculus. Furthermore, significant changes were detected in the anterior limb of the internal capsule containing projection fibers between the lentiform nucleus and caudate nucleus, which are responsible for the control of motor and sensory pathways.

It is well known that executive function, attention and memory are regulated by deep brain nuclei in the subcortical regions of the brain. We found significant microstructural changes in the grey matter of intracranial GCT survivors. Significant differences were found in the MD of grey matter regions responsible for language, cognition, and executive functions, including the cingulum, amygdala, temporal gyrus, fusiform, and rolandic operculum. Significant differences were also found in brain regions responsible for visual processing, including the calcarine, lingual gyrus, and cuneus. Other affected brain regions included Heschl’s gyrus, which is the primary auditory cortex, the precentral gyrus containing the primary motor cortex, and the postcentral gyrus containing the somatosensory cortex.

The estimated FSIQ is typically derived from subtests for performance IQ as well as verbal IQ. Our results showed that the IQ scores were negatively correlated with the MD of brain regions responsible for language processing, including those related to Wernicke’s area in the temporal lobe. Heschl’s gyrus is associated with the processing of speech-related cues, which facilitates learning and perception of new speech sounds (24). In addition, lower IQ scores were found to be associated with the MD of brain regions responsible for cognitive function such as the hippocampus, parahippocampal gyrus, and anterior cingulum. The frontal regions of the brain are known to be involved in cognitive functions and control of emotions. Our study showed significant association between FSIQ scores and MD values in frontal regions such as the orbitofrontal cortex as well as regions around Broca’s area. The frontal inferior opercularis acts indirectly through the motor cortex within the precentral gyrus to control the motor aspects of speech production (25). Interestingly, it is common for patients with neurodegenerative diseases such as Alzheimer’s disease and dementia to have motor speech deficits (26, 27). Hence, GCT survivors might be prone to motor speech problems. Our findings are consistent with previous studies on childhood acute lymphoblastic leukemia survivors with a history of chemotherapy and/or radiotherapy, which found that lower volume of the caudate nucleus, was associated with significantly worse verbal fluency (28). In addition, our study demonstrated the significant association of the precentral gyrus with the IQ scores of GCT survivors, suggestive of the involvement of the precentral gyrus in cognitive tasks. The precentral gyrus is the site of the primary motor cortex and is traditionally implicated in voluntary movement control. However, more recent studies suggested that the primary motor cortex not only plays a role in stimulus-response compatibility, plasticity, motor sequence learning and memory as well as learning of sensorimotor associations, but is also engaged in motor imagery, spatial transformations and working memory tasks (29–32). Therefore, our study provides additional evidence to support the involvement of the primary motor cortex in higher cognitive tasks.

Most of the microstructural changes that were associated with functional impairment, as indicated by the poor KPS score, were in white matter regions. These regions included the pyramidal tracts within the internal capsule or fiber tracts such as the corona radiata above the basal ganglia, which had elevated MD but reduced MK and FA. Our findings were consistent with studies on children with cerebral palsy, which found significant correlation between motor function scores and the FA of sensory and motor pathways such as the corticospinal tract, thalamic radiation, or the posterior limb of the internal capsule (33). Functional impairments that were associated with microstructural changes in grey matter regions were mostly confined around the basal ganglia, which have important roles in motor control.

Historically, radiation damage was thought to affect brain white matter rather than the cortex itself (12, 34). Nevertheless, a recent study demonstrated that brain tumor survivors with a history of irradiation had cortical atrophy that was dependent on the radiation dose (35). Seibert et al. showed that cortical atrophy in patients after brain radiotherapy was significantly associated with radiation dose in the entorhinal and inferior parietal regions, which are areas responsible for memory and executive functions (36). They also demonstrated that these cerebral cortex regions seemed to be more vulnerable to dose-dependent radiation atrophy. Our study corroborates the findings by Seibert et al. We also found significant correlations between grey matter microstructural changes and cognitive function in GCT survivors with history of brain radiotherapy. Our study is first to demonstrate significant associations between white and grey matter microstructural changes in vulnerable brain regions of GCT survivors in association with cognitive impairments and/or functional deficits.

Our study had several limitations. First, the number of study subjects was small. Nevertheless, our study was the first to perform DKI measurements in association with cognitive and physical functioning assessments in a cohort of GCT survivors. In addition, our study participants were predominantly male (80%). However, predilection to the male gender is a well-known characteristic of GCT (37) and previous studies also demonstrated similar male: female ratio (38, 39). Second, due to the retrospective nature of the study, we did not have information on the physical or neurocognitive functions in our cohort at the time of tumor diagnosis. Therefore, it is hard to ascertain if the cognitive or functional impairments are due to individual or combined effects of the tumor, chemotherapy, or radiotherapy. Nevertheless, radiation-induced brain damage has long been recognized in pediatric cancer patients, more recent studies have shown that chemotherapy can also lead to neurotoxicity (40). The microstructural brain changes and cognitive or functional impairments that were found in the current study are likely due to the combined effects of all three factors. Lastly, the measurement of the diffusion metrics of GM regions could be confounded by the partial volume effect from free fluid such as cerebral spinal fluid, for which could not be accounted in our statistical analyses.

In conclusion, our DKI study has successfully demonstrated that impairment in cognitive function and/or deficits in physical functioning were associated with microstructural changes in multiple white and grey matter regions. Proton radiotherapy with reduced irradiation dosage to vulnerable brain regions might lead to improved cognitive and functional outcomes. Diffusion metrics obtained from DKI could potentially be used to identify patients at high risk of cognitive or functional impairment for timely interventions.

The raw data supporting the conclusions of this article will be made available upon request to the authors.

The studies involving human participants were reviewed and approved by The Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (HKU/HA HKW IRB), the Research Ethics Committee of the Kowloon Central/ Kowloon East Cluster (KC/KE CREC), Kowloon West Cluster (KW CEREC) and New Territories West Cluster (NTW CEREC). Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

Conceptualization: WT, EH, TL, PK and GC. Data curation: WT, EH, TL, AL, PI, VV, KC, DF, DHFC, FH, KY, DK, DKLC, CL, MS, LL, PK and GC. Formal analysis: WT, EH, TL, AL, KC, DF, FH, KY, LL, PK and GC. Funding acquisition: WT. Investigation: WT, EH, TL, VV, PK and GC. Methodology: WT, EH, TL, AL, PI and FH, PK and GC. Project administration: WT, EH, TL, PK and GC. Software: EH and KY. Supervision: WT, EH, TL, PK and GC. Validation: DF. Writing – original draft: WT and EH. Writing – review and editing: WT, EH, TL, AL, PI, VV, KC, DF, DHFC, FH, KY, DK, DKLC, CL, MS, LL, PK and GC. All authors contributed to the article and approved the submitted version.

This work was supported by Research Grants Council of the Hong Kong Special Administrative Region, China (No. HKU 17118815). The funder had no role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.