Neuropsychological outcomes of children with Optic Pathway Glioma

Optic Pathway Glioma (OPG) is a relatively common brain tumour in childhood; however, there is scarce understanding of neuropsychological sequelae in these survivors. In this study, 12 children with diagnosis of OPG before 6 years of age received a comprehensive standardised assessment of visual perception, general intelligence and academic achievement, using adjustments to visual materials of the tests, to examine the extent of concurrent impairment in these functional domains. Information about vision, clinical and socio-demographic factors were extracted from medical records to assess the associations of neuropsychological outcomes with clinical and socio-demographic factors. Children with OPG exhibited high within-patient variability and moderate group-level impairment compared to test norms. Visual perception was the most impaired domain, while scholastic progression was age-appropriate overall. For cognition, core verbal and visuo-spatial reasoning skills were intact, whereas deficits were found in working memory and processing speed. Visual function was associated with tasks that rely on visual input. Children with OPG are at moderate risk of neuropsychological impairment, especially for visual perception and cognitive proficiency. Future research should elucidate further the relative contribution of vision loss and neurofibromatosis type 1 co-diagnosis within a large sample.

www.nature.com/scientificreports www.nature.com/scientificreports/ Visuo-perceptual skills were assessed using the Developmental Test of Visual Perception, either 2 nd edition (DTVP-II; 51 for ages 4-10 years) or Adolescent and Adult edition (DTVP-A; 52 for ages 11-74 years). These tests consist of 8 and 6 subtests respectively and provide two analogous indices: Motor-Reduced Perception and Visual-Motor Integration. Tasks of both indices rely on vision, but only Visual-Motor Integration requires fine motor skills mainly with untimed performance. A composite index of General Vision Perception was obtained.
Cognitive abilities were evaluated with the Wechsler Intelligence Scale for Children, 4 th UK edition (WISC-IV UK53 ; for ages 6-16 years). It consists of 10 core subtests that provide four indices of broader abilities: Verbal Comprehension, Perceptual Reasoning, Working Memory and Processing Speed. Verbal Comprehension and Working Memory involve verbal-auditory input and output; Perceptual Reasoning and Processing Speed rely on vision, with Processing Speed requiring fine motor skills and timed performance. Two composite indices were obtained: Full-Scale IQ which is based on all four individual indices, and General Ability index which is based on Verbal Comprehension and Perceptual Reasoning.
Scholastic attainment was assessed using the Wechsler Individual Achievement Test, 2 nd UK edition (WIAT-II UK54 ; for ages 4-16 years). It consists of 10 core subtests that provide four indices of broader academic abilities: Reading, Mathematics, Written Language and Oral Language. All indices rely on vision, but only Mathematics and Written Language require fine motor response. A Total Composite index was obtained.
The neuropsychological assessment was conducted in a quiet area at the child's home over one to three sessions, within four consecutive weeks. All children wore their glasses, if prescribed. For subtests that required vision, children with moderate or severe visual impairment were allowed to hold print material close for tasks that required a fine motor response (regardless of the timing) and to use their own magnifier device for motor-free subtests. No adaptation was used with children with monocular/binocular normal vision or mild visual impairment (best corrected VA ≥ 0.50 logMAR score) to minimise disruption of standardised procedures.
Neuropsychological performance was scored following the test manuals and raw scores were transformed into standard scores (μ = 100, σ = 15) for each index. Scores were classified as: clinical impairment (below -2 SD), deficit (−1 to −2 SD), average (−1 to +1 SD), above-average (+1 to +2 SD) and gifted (above +2 SD). Socio-economic status (SES) was estimated through postcodes using the Income Deprivation Affecting Children Index rank 2015 55 . Values range from 1 to 32,482, with lower scores indicating more disadvantaged areas.
Sample biases were assessed comparing eligible patients who did and did not participate in terms of age, sex and NF1 comorbidity through independent-sample t-tests and chi-square tests. Chi-square tests with Fisher's exact correction were used to assess differences between sporadic and syndromic OPGs on socio-demographic and clinical characteristics. All individual and composite indices were analysed using one-sample t-tests to assess deviations from the test norms. For comparison with previous research 14 , at the cognitive level, paired-sample and independent-sample t-tests were conducted with Verbal Comprehension, Perceptual Reasoning and Full-Scale IQ to investigate discrepancies across indices and differences between sporadic and syndromic cases for these measures. Confidence intervals at 95% level were computed and effect sizes were estimated as Cohen's d 58 .
Pearson correlations were conducted to examine the associations across the three domains. Finally, several socio-demographic and clinical factors in relation to neuropsychological outcomes were considered. Associations with sex (1: female, 2: male), tumour type (1: sporadic, 2: syndromic) and visual function (1: poor, requiring visual aids, 2: good, not requiring visual aids) were examined using point-biserial correlations. The relationship with visual acuity (VA in the best eye) and SES was evaluated using Spearman's ρ correlations. Pearson's r correlations were conducted to examine the impact of age at diagnosis and time post treatment on OPG outcomes. All correlation analyses were conducted on composite indices and Bonferroni correction was applied (α = 0.05/4 = 0.0125). Further correlations using individual indices were explored with the same alpha level of 0.0125.

Visual function.
Neuropsychological outcomes: extent of impairment. Two children did not complete the cognitive and scholastic tests due to fatigue at sustaining multiple sessions. Table 1 summarises individual and group results for the neuropsychological assessments. Figure 2 reports group means and comparisons to test norms.
For visual perception, a significant difference from test norms was found for both individual indices Motor-Reduced Perception and Visual-Motor Integration, and for the composite index General Visual Perception. (2020) 10:3344 | https://doi.org/10.1038/s41598-020-59896-2 www.nature.com/scientificreports www.nature.com/scientificreports/ For intellectual function, there was a significant difference from test norms for Full-Scale IQ, but not General Ability. Scores for Working Memory and Processing Speed were significantly depressed compared to test norms. No significant differences were found between Verbal Comprehension and Perceptual Reasoning (t 9 = 0.58, P = 0.577). No significant differences were found between syndromic and sporadic cases for either Full-Scale IQ (82. 6  For scholastic achievement, a deviation from test norms was found for two abilities, Written Language and Oral Language, but not the Total Composite index. Neuropsychological outcomes: associations across domains. Table 2 shows associations between all the neuropsychological indices. Across domains, strong correlations for composite indices were found between scholastic attainment and cognitive functioning (both Full-Scale IQ and General Ability; r ≥ 0.85, P ≤ 0.002), whereas no significant correlations were found for visual perception. On individual indices, one medium-to-strong correlation was found between the perception index of Visual-Motor Integration and the visuo-spatial cognitive index of Processing Speed (r = 0.68, P = 0.031), but none between perception and academic attainment. Several moderate-to-strong correlations were found between individual indices of cognitive and scholastic abilities (r ≥ 0.65, P ≤ 0.043). Table 3 shows the relationship of neuropsychological outcomes with clinical and socio-demographic factors. Visual function was significantly associated with Perceptual Reasoning, Processing Speed and Written Language (all P < 0.05), demonstrating worse performance in children with poor vision requiring low vision aids compared to those with sufficiently intact vision (although www.nature.com/scientificreports www.nature.com/scientificreports/ these did not survive Bonferroni correction). The other clinical and socio-demographic factors did not correlate significantly with neuropsychological outcomes. The direction of the effects was mixed for all variables except for the NF1 status, indicating that children with sporadic OPG consistently obtained lower scores on all indices (Supplementary Table S2).

Discussion
This is the first study documenting low-level visual acuity, high-level visual perception, verbal and visuo-spatial intellectual function and real-life scholastic abilities in children with OPG managed with observation or chemotherapy.      www.nature.com/scientificreports www.nature.com/scientificreports/ Analyses of individual profiles showed visual perception was the most affected area, with half of the sample underperforming. For intellectual function and academic attainment, most children performed on average for each domain. Neuropsychological outcomes displayed a high degree of heterogeneity and complex relationships with visual impairment and NF1 co-diagnosis. Of note, three children (two sporadic cases) performed at or above average for some indices, whereas two children with moderate or severe visual impairment and co-diagnosis of NF1 (treated with chemotherapy) performed below average on almost all abilities. This result corroborates findings from a previous case series describing cognitive deterioration after chemotherapy only in 3 children with NF1-associated OPG, who also had very poor vision at diagnosis 18 . At the group level, none of the indices were clinically impaired (≤ −2 SD), but all mean scores were in the lower range of average performance or just below     www.nature.com/scientificreports www.nature.com/scientificreports/ −1SD. The underperformance in comparison to the test norm μ = 100 was statistically significant for more than half of the indices (2/4 composite and 6/10 individual indices).
Visual perception was found to be a marked area of weakness in paediatric OPG patients as group performance of all measures was significantly below the test mean. Difficulties with processing visual-perceptual stimuli, even without motoric response and after low vision adjustment, raise questions about the reliability of visual acuity tests in these patients 59 , when children are required to analyse the physical properties of a stimulus to recognise and name a shape or letter without meaningful context. Visual-perceptual processing and visual-motor integration were altered in children with low vision 60 , and weaknesses in these areas are also well established among NF1 children 6,7,31 . Our small sample did not afford the relative contribution of visual function and NF1 status to be elucidated. However, these results indicate that visual perception should be incorporated into clinical assessments of children with OPG.
Amongst composite indices of intellectual function, the mean score of General Ability did not differ from test norms, whereas the mean score of Full-Scale IQ was significantly below. General Ability comprises only of Verbal Comprehension and Perceptual Reasoning, both of which were preserved. In contrast, Full-Scale IQ also includes Working Memory and Processing Speed, which showed significant weaknesses in our sample. Therefore, the discrepancy between average General Ability and statistically-reduced Full-Scale IQ demonstrates the weaknesses in Working Memory and Processing Speed, which reflects the proficiency and efficiency of cognitive processing 61 . A similar pattern of results was found in a heterogeneous group of young brain tumour survivors 62 , where all children received radiation therapy and all cognitive measures were below the test mean. Our study provides promising results for non-irradiated OPG patients as it demonstrates that core verbal and visuo-spatial reasoning abilities remain intact in these children. Nonetheless, the deficit shown on general intelligence when Working Memory and Processing Speed are considered indicates that these two cognitive domains require clinical attention as they are critical for intellectual development [63][64][65] and scholastic progression 66,67 .
At the cognitive level, there was no discrepancy between Verbal Comprehension and Perceptual Reasoning, and both mean scores were close to the test norm. Previous research by Lacaze and colleagues 14 reported a significant difference between intact verbal IQ and depressed performance IQ (analogous indices in the WISC-III), but their results were confounded by the presence of irradiated children in their sample and the lack of control over visual input during neuropsychological assessment. On the contrary, our study with non-irradiated OPG patients demonstrated similar development of Verbal Comprehension and Perceptual Reasoning when children use visual magnifiers. This ecological approach, which is in line with the recommendations of the Brain Tumour Quality of Survival Group, International Society of Paediatric Oncology (Europe) 24 , showed that when visual stimuli are made accessible, visuo-perceptual reasoning in children with OPG develops adequately for the child's age and is similar to their verbal comprehension skills. In addition, compared to Lacaze and colleagues 14 , our study showed no differences between sporadic and syndromic cases not only on verbal IQ measure, but on visual-spatial and total IQ scores. Although these novel results could arise from the adaptation of visual input, subgroup mean scores support a tendency towards better performance in children with sporadic OPGs compared to those with NF1 co-diagnosis. Of note, these analyses were conducted only for replication purpose, systematic investigation with larger samples is warranted in future.
Our results revealed relatively preserved scholastic attainment compared to other domains, with some children performing above average on some measures. Children with OPG had appropriate scholastic attainment based on the Total Composite index and this was also confirmed on the individual indices of Reading and Mathematics. Mild impairment was found only in Written Language and Oral Language. Underperformance on Written Language might be attributable to visual-motor and handwriting difficulties, but also spelling deficits that can result from either low vision 68,69 or NF1 syndrome 70,71 . The impairment in Oral Language might reflect difficulties at processing fine details of visual stimuli necessary for many of these tasks.
Across domains, significant associations between composite indices of visual perception and cognitive and scholastic abilities were not found. However, some moderate-to-strong correlations emerged for individual indices across domains that rely on vision and/or fine motor response, although these did not reach statistical significance (for example, between Motor-Reduced Perception and Perceptual Reasoning and between Visual-Motor Integration and Written Language), probably because of the small sample size. Therefore, associations/dissociations across these three domains should be explored further in larger studies before excluding the potential utility of visual rehabilitation programs for these patients.
Finally, among the clinical and socio-demographic factors, visual function showed medium-to-strong positive correlations with some cognitive and scholastic individual indices that were assessed with visual tasks. Consistently, VA in the best eye was negatively correlated with Processing Speed. This demonstrates that, even if children with poor vision were allowed low vision aids, their performance was significantly poorer in comparison to the children with relatively intact vision. It has been proposed that early visual loss may either hamper neuropsychological acquisition causing long-term detrimental effects, or facilitate timely neural re-organisation and behavioural adaptation resulting in adequate development 12 . Our results support the first proposition as children with poor visual function obtained lower scores on neuropsychological tests despite the use of low visual aids. This is consistent with results in adulthood reporting poor neurocognitive outcomes in low-grade glioma survivors with bilateral blindness compared to those without visual impairment 12 . Notably, NF1 demonstrated a consistent effect on all neuropsychological abilities, with syndromic cases underperforming in comparison to sporadic cases, although this was statistically significant only for Reading. This is consistent with findings by Lacaze and colleagues 14 , who also compared children with and without NF1 treated with front-line chemotherapy. While NF1 research tends to exclude patients with OPG due to brain abnormality or visual impairment, the results of this study suggests that overall deficits associated with OPG might be driven by NF1 co-diagnosis and therefore these children require more research and clinical attention, although there is a desperate need for reliable and standardised tools to evaluate visually impaired children 72 .