1000/1000
Hot
Most Recent
Radiotherapy for head and neck cancers exposes small parts of the brain to radiation, resulting in radiation-induced changes in cerebral tissue. Implementation of neurocognitive assessment with advanced MRI examination in monitoring brain microstructural and functional changes of head and neck cancer patients could detect cognitive changes early. With suitable intervention, further deleterious effects on the patient’s cognition can be prevented.
First Author, Year | Average MoCA Post-RT (Range, Bonus Point) | Pre-MRI Findings | Post-MRI Findings | Study Limitations |
---|---|---|---|---|
Ma Q, 2016 [38] | 24.2 (22–27) | 45 altered FC compared to untreated NPC group | Heterogeneous treatment protocol, combined both non-irradiated and irradiated subjects, varied sample size, lack of new and larger sample, and between-subject variance | |
Qiu Y, 2017 [29] | NR | Functional network connectivity for NPC patients pre- and post-RT shared similar connectivity | Weaker intra-network connectivity with lower mean connectivity correlation than baseline | Heterogeneous treatment protocols, between-subject variance |
Ma Q, 2017 [39] | 24.2 (22–27) | Altered FC between cerebellar seeds and relative brain clusters | Heterogeneous treatment protocol, combined both non-irradiated and irradiated subjects, varied sample size, lack of new and larger sample, and between-subject variance | |
Guo Z, 2018 [40] | <26 | No differences in cerebral volume of pre-NPC to controls | Decrease in brain macrostructural volume | Combined both non-irradiated and irradiated subjects, short time interval, and varied sample size |
Lv X, 2018 [41] | NR | No significant differences in volumes of hippocampus and hippocampal subfields between groups | Significant volume reductions in bilateral hippocampus and hippocampal subfields | Combined both non-irradiated and irradiated subjects and varied sample size |
Ren WT, 2019 [36] | 27 (24–29) | No significant changes in regional cerebral and connectivity before RT | Reduced regional cerebral and neural network functions | Comparison to healthy controls and small sample size, short time interval |
Wu G, 2020 [37] | <26 (<12 years education and >65 years age) |
Baseline of kurtosis and diffusivity does not show significant difference | Significantly lower kurtosis and diffusivity of white matter | Heterogeneous treatment protocols, comparison between different marker groups, and between subject-variance |
First Author, Year | Score | Functional Connectivity or Volume | Significant Relationships and Prediction Details | Summary |
---|---|---|---|---|
Functional connectivity | ||||
Ma Q, 2016 [38] | MoCA | Vermis and hippocampus | r = 0.4440, p = 0.00043 |
↓ FC ↓ MoCA score |
Attention | r = 0.4282, p = 0.00072 |
↓ FC ↓ Attention score | ||
MoCA | Cerebellum lobule VI and dIPFC | r = −0.4343, p = 0.00059 |
↑ FC ↓ MoCA score | |
Precuneus and dFC | r = 0.4622, p = 0.00023 |
↓ FC ↓ MoCA score | ||
Cuneus and middle occipital lobe | r = 0.4282, p = 0.00071 |
↓ FC ↓ MoCA score | ||
Anterior insula and cuneus | r = 0.4569, p = 0.00028 |
↓ FC ↓ MoCA score | ||
Qiu Y, 2017 [29] | MoCA | Left anterior cingulate cortex within the default mode network (DMN) | No significant correlation | |
Right insular within salience network (SN) | No significant correlation | |||
Bilateral executive control network (ECN) | No significant correlation | |||
Ma Q, 2017 [39] | MoCA | Right cerebellar lobule VIIb and right fusiform gyrus | r = −0.34, p = 0.008 |
↑ FC ↓ MoCA score |
Attention | r = −0.41, p = 0.002 |
↑ FC ↓ Attention score | ||
MoCA | Left cerebellar lobule VIII and right crus I | r = −0.30, p = 0.021 |
↑ FC ↓ MoCA score | |
Attention | r = −0.32, p = 0.001 |
↑ FC ↓ Attention score | ||
Attention | Left cerebellar lobule VIII and right MFG | r = −0.27, p = 0.040 |
↑ FC ↓ Attention score | |
Ren WT, 2019 [36] | MoCA | Default mode network (DMN) | No significant correlation | |
Volume | ||||
Guo Z, 2018 [40] | MoCA | Ventricular | bβvolume = −4.63 × 10−4, p = 0.007 |
↓ Volume ↓ MoCA score |
Lv X, 2018 [41] | MoCA | Left hippocampus | bβvolume = 0.010, p = 0.017 |
↓ Volume ↓ MoCA score |
Right Hippocampal | bβvolume = 0.013, p = 0.002 |
↓ Volume ↓ MoCA score | ||
Left Subiculum | bβvolume = 0.061, p = 0.018 |
↓ Volume ↓ MoCA score | ||
Left Granule cell layer (GCL) | bβvolume = 0.102, p = 0.011 |
↓ Volume ↓ MoCA score | ||
Right Granule cell layer (GCL) | bβvolume = 0.158, p = 0.022 |
↓ Volume ↓ MoCA score | ||
Right molecular layer (ML) | bβvolume = 0.285, p = 0.002 |
↓ Volume ↓ MoCA score | ||
Kurtosis | ||||
Wu G, 2020 [37] | MoCA | Hippocampal | r = 0.76, p < 0.05 | Kurtosis mean-1 best in predicting MoCA scores decline |
First Author, Year | Dose-Dependent Changes |
---|---|
Ma Q, 2016 [38] | Functional connectivity pattern in NPC treated patients was significantly impaired compared to NPC untreated with changes shown in cerebellum, sensorimotor, and cingulo-opercular. |
Qiu Y, 2017 [38] | Changes in right insular functional connectivity were negatively correlated with dose of right temporal lobe. |
Ma Q, 2017 [39] | Altered cerebral-cerebral functional connectivity within dorsal attention, default, and frontoparietal networks shown in NPC treated patients. |
Guo Z, 2018 [40] | Significantly decrease volume in bilateral temporal lobe with increased mean dose to this region. |
Lv X, 2018 [41] | Volume deficits in the bilateral hippocampus, bilateral granule cell layer, and right molecular layer negatively correlates with the mean dose to ipsilateral hippocampus. |
Ren WT, 2019 [36] | Decreased connectivity in multiple cerebellar-cerebellar regions mainly in the default-mode networks likely because of radiation dose. |
Wu G, 2020 [37] | Significant radiation-induced changes in both white and gray matter of the temporal lobes due to the high radiation dose received. |
Studies indicate a cut-off point of 26 in MoCA assessments to define cognitive impairments. Though changes in MoCA scores were associated with MRI outcomes, the cut-off may be too stringent and not optimal among minorities [43] and certain health condition populations [44]. Additionally, the cut-off is also too high for cognitively normal older adults, even those who are highly educated [45]. Nevertheless, the use of MoCA is shown to be efficient in screening for mild cognitive impairment among the Chinese population [46][47] with the Cantonese Chinese MoCA being a consistent and reliable instrument [48]. Therefore, it is crucial to use age [45], education [43][45], and race or ethnicity [43][49] in correcting the cut-off scores to avoid misdiagnosis of cognitive decline. A lower MoCA cut-off score 23/30 yielded an overall better diagnostic accuracy with a lower false positive rate and excellent sensitivity (96%) and specificity (95%), thus, is recommended as the new MoCA cut-off score [49][50].
The neurocognitive assessment has shown the likelihood to be associated with MRI outcome following head and neck cancer radiotherapy, especially for the temporal region. Focus is given to the region due to its proximity to the target volume and would inevitably be incorporated into the treatment field, which exceeds the tolerance limit. Changes in functional connectivity (FC) and brain volume were significantly correlated with MoCA scores in most studies [38][37][40][41]. From the findings, the change of correlation from negative to positive may implicate that the RT process might have impaired the anticorrelation between the two networks of NPC patients. The impaired anticorrelation between the dorsal attention and default networks may suggest deficits in cognitive and attention processing of NPC patients after RT [39]. The correlation may also be inferred to be due to the radiation-induced cognitive impairment of domains such as short-term memory, visual memory, language ability, attention, and executive function [38][39]. In terms of volume, longitudinal changes in MoCA scores were associated with the longitudinal changes in total grey matter and bilateral temporal and ventricular volumes.
Radiation-induced changes were also observed throughout the studies investigated. The early changes are closely related to vascular damage shown by vessel dilation, endothelial cells loss, nuclei enlargement, vessel wall thickening, increased vessel permeability, and decreased vessel density and length [19][51]. Resultant functional connectivity and brain volumes from irradiation were observed in multiple cerebellar regions. This was shown with the altered correlation between brain networks observed in NPC patients following RT, which may imply deficits in cognitive and attention processing [38][39]. In addition, the demonstrated differences in the FC pattern also suggests that radiation-induced changes may not be bound to the exposed area only, but other encephalic regions such as the cerebellum, sensorimotor, and cingulo-opercular areas [38]. This shows that the incidental radiation received by the brain during treatment of HNC could contribute to cognitive impairment [52]. The findings suggest that early microstructural injury of the temporal lobe has a direct contributory relation to the delayed neurocognitive decline with lower MoCA scores shown post-radiotherapy [37]. Specifically, an increased radiation dose to the temporal lobes and cerebellum were significantly associated with worse memory performance and motor coordination, respectively [52]. In addition, a higher radiation dose (30 Gy) induced earlier and more severe histological changes than a lower dose that were reflected with changes in diffusivity and perfusion [53][54][55]. Nevertheless, in the study done by Zer et al. [56], no significant correlation was shown to suggest the risk of treatment parameters, such as chemotherapy regimen or radiation dose, to greater cognitive decline.
Radiation-induced atrophy was also demonstrated in the bilateral hippocampus, bilateral GCL, and right molecular layer [17][41][57], suggesting the atrophy of the subfields is primarily induced by radiation that might be associated with early radiation effects on vascular injury, reduced molecular layer volume, and disruption of neuronal structure and synaptic integrity [41]. The elevated volume losses in these areas were associated with a rapid cognitive function decline evaluated by MoCA in irradiated patients [41], indicating dose-dependent atrophy. Additionally, altered FC within the default-mode and salience networks also indicates high-order cognition impairment, especially memory and attention [29]. According to Wen et al. [58], limiting the dose delivered to 0.5-cm3 temporal lobe volume (D0.5cc) to less than 65.06 Gy may be advisable during IMRT for NPC patients, as it decreases the risk of temporal lobe injury (TLI) in older patients with advanced tumour stage. Thus, the implementation of NTCP modelling could potentially predict TLI and allow individualised follow-up management. Therefore, a clinically appropriate and safe dose is crucial in protecting these vulnerable regions.