Treatment age and neurocognitive outcomes following proton beam radiotherapy for pediatric low- and intermediate-grade glioma

Abstract

The relationship between age and neurocognitive functioning following proton beam radiotherapy (PRT) in low- and intermediate-grade gliomas (LIGG) has yet to be examined. Eighteen LIGG patients treated with PRT were prospectively enrolled and received annual neurocognitive evaluations of perceptual/verbal reasoning, working memory, and processing speed post-radiotherapy. The median age at diagnosis was 8.2 years (range=1.0–14.7) and the median age at PRT was 9.9 years (range=4.2–17.0). Patient’s neurocognitive performance did not change on any measure following PRT (p≥0.142). We did not observe significant changes in cognitive function over time among a small group of LIGG patients treated with PRT.

Introduction

Treatment for low- and intermediate-grade gliomas (LIGG) may require radiotherapy, which is associated with increased cognitive risk.¹ The established association between radiotherapy and cognitive risk in pediatric brain tumor patients is primarily based on the use of photon radiotherapy (XRT).² Despite the development of conformal techniques, XRT is limited in its ability to avoid non-target tissues, increasing cognitive risk.³ In contrast, proton beam radiotherapy (PRT) allows for steep dose gradients and has superior dose profiles compared to XRT,⁴ sparing healthy tissue. Diagnosis and treatment with XRT at a younger age increases cognitive risk for patients with pediatric LIGG.⁵ To minimize neurocognitive late effects, radiotherapy is often delayed or avoided in younger patients, and chemotherapy is used as the preferred initial treatment.⁶

In the current study, we prospectively assessed neurocognitive functioning following PRT in pediatric patients with LIGG. We sought to evaluate if younger age remained a significant risk factor following PRT treatment. Additionally, we explored the influence of established predictors of neurocognitive functioning in pediatric brain tumor patients, including hydrocephalus requiring a ventriculo-peritoneal (VP) shunt⁷ and maternal education.⁸

Methods

Patients

This study reports on outcomes from a prospective, longitudinal study of neurocognitive functioning in pediatric brain tumor patients. Following approval from the Institutional Review Board, eligible participants were identified by medical chart review and consecutively enrolled. Informed written consent and assent was obtained prior to participation. Participants received annual neurocognitive testing following surgery.

Eligible participants, primarily English-speaking pediatric brain tumor patients between the ages of 3–18 years (inclusive), were enrolled following surgery. In the current analysis, we included patients diagnosed with a low or intermediate grade glioma requiring PRT treatment, including juvenile pilocytic astrocytoma (n=10), grade II astrocytoma (n=4), infiltrating glioma (n=2), and single cases of anaplastic ganglioglioma and fibrillary astrocytoma.

Between 2012–2019, we enrolled 211 patients on the study, with a 90% retention rate. Patients who declined enrollment did not differ from enrolled participants by age at treatment, sex, or race (n=53; data not shown, p>0.05). Overall, 87 patients were diagnosed with LIGG and 29 of those patients received PRT treatment. We analyzed data for 18 patients who had completed two or more neurocognitive evaluations at the time of the present analysis.

Measures

Participants were administered the age-appropriate version of the Wechsler Scales of Intelligence.^9–11 The Full-Scale IQ (FSIQ) score provides a measure of global intellectual functioning. Index scores include: Verbal Comprehension (VCI), Perceptual Reasoning (PRI), Working Memory (WMI), and Processing Speed (PSI).

Statistical Analyses

Demographic and clinical characteristics were compared by treatment group using Chi Square, Fisher exact test, or independent t-tests, as appropriate. General linear mixed models compared change in neurocognitive scores over time between treatment groups. Due to our small sample size and limited power, we only included established predictors in the literature to reduce the number of analyses. The base model included a fixed effect for time as well as a random intercept and slope. We then built individual models for the five outcomes using individual predictors, including age at treatment (< 9 vs. ≥9), VP shunt treatment (yes/no), and maternal education (maternal 4-year degree).

Results

Longitudinal Neurocognitive Performance

Base linear-mixed models, without covariates, revealed no change in neurocognitive performance over time on the VCI (baseline=94.03, +0.26 points/year, p=0.71), PRI (baseline=92.17, +0.67 points/year, p=0.14), PSI (baseline=85.80, +.38 points/year, p=0.61), WMI (baseline=94.72, +0.34 points/year, p=0.44), and FSIQ (baseline=90.01, +0.75 points/year, p=0.22).

Age at Treatment

Demographic and clinical characteristics of the sample by age at treatment are provided in Table 1. On average, VCI and WMI scores were 19.1 and 21.2 points higher among patients who were treated with radiotherapy ≥9 years of age compared with those treated prior to 9 years of age after adjusting for time since RT (without interaction term), respectively (p<0.05). Table 2 displays changes in cognitive functioning over time by treatment age group (≥9 vs. <9). On average, the PSI decreased by 1.01 points per year for patients who started PRT at <9 years and increased by 1.17 points per year for patients who were ≥9 years at the start of PRT (p=0.17).

TABLE 1.

Demographic and clinical characteristics by age group at treatment

	Total		< 9		≥ 9
	n	%	n	%	n	%	p
Sex							0.332
Male	7	38.9	4	57.1	3	27.3
Female	11	61.1	3	42.9	8	72.7
Race/Ethnicity							0.245
Hispanic/Latino	4	22.2	3	42.9	1	9.1
Other	14	77.8	4	57.1	10	90.9
Maternal Ed.							0.637
< 4 year degree	11	61.1	5	71.4	6	54.5
4 year degree	7	38.9	2	28.6	5	45.5
Tumor Location							0.119
Supratentoriala	14	77.8	7	100	7	63.6
Infratentorial	4	22.2	0	0.0	4	36.4
Number of tumor directed surgeries							0.316
One	11	61.1	6	85.7	5	45.5
Two	6	33.3	1	14.3	5	45.5
Three	1	5.6	0	0.0	1	9.1
Shunt							>0.999
No	13	72.2	5	71.4	8	72.7
Yes	5	27.8	2	28.6	3	27.3
Chemotherapy							0.141
No	12	66.7	3	42.9	9	81.8
Yes	6	33.3	4	57.1	2	18.2
	Median	Min-Max	Median	Min-Max	Median	Min-Max
Maximum tumor diameter (cm)	4.2	1.5 – 7.4	5.0	1.5 – 6.7	4.0	3.1 – 7.4	0.177
Age at diagnosis (years)	8.2	1.0 – 14.7	4.9	1.0 – 8.4	11.4	3.8 – 14.7	0.012
Age at radiation (years)	9.9	4.2 – 17.0	8.2	4.2 – 8.9	14.2	9.2 – 17.0	0.003
Radiation dose (Gy)	50.4	45.0 – 55.8	50.4	48.6 – 55.8	50.4	45.0 – 54.0	>0.999
Treatment to final evaluation (years)	3.8	1.2 – 7.6	3.3	1.5 – 7.6	3.9	1.2 – 6.1	>0.999
Baseline to final evaluation (years)	4.0	1.0 – 7.5	3.1	1.5 – 7.5	4.1	1.0 – 6.0	0.929
Number of evaluations	4.0	2 – 7	4	2 – 7	5	2– 7	0.614

Note. Abbreviations: n = sample size, Ed = Education, Min = Minimum, Max = Maximum. Group comparisons based on Chi Square, Fisher exact test, or independent t-test, as appropriate.

^aPrimary supratentorial tumor location included, optic (n = 5), hypothalamic (n = 5), temporal (n = 3), and single cases of left lateral ventricle, pineal gland, and basal ganglia

TABLE 2.

Neurocognitive functioning over time by age at radiotherapy group

Neurocognitive Measure	Effect	Estimate	SE	p value
Verbal Comprehension Index (n = 18)	Intercept	81.79	5.13	<.0001
	Time Since Radiation (years)	0.69	1.10	0.540
	Age group (≥ 9 years)a	20.01	6.55	0.007
	Time * age group	−0.71	1.40	0.620
Perceptual Reasoning Index (n = 18)	Intercept	87.43	6.38	<.0001
	Time Since Radiation (years)	1.11	0.76	0.157
	Age group (≥ 9 years)a	7.79	8.16	0.355
	Time * age group	−0.70	0.96	0.469
Processing Speed Index (n = 16)	Intercept	86.45	8.24	<.0001
	Time Since Radiation (years)	−1.01	1.17	0.419
	Age group (≥ 9 years)a	−0.67	10.42	0.950
	Time * age group	2.18	1.43	0.168
Working Memory Index (n = 16)	Intercept	80.85	6.75	<.0001
	Time Since Radiation (years)	0.40	0.73	0.590
	Age group (≥ 9 years)a	21.39	8.36	0.020
	Time * age group	−0.09	0.92	0.919
Full Scale Intelligence Quotient (n = 18)	Intercept	81.00	6.68	<.0001
	Time Since Radiation (years)	0.83	0.98	0.416
	Age group (≥ 9 years)a	14.76	8.54	0.103
	Time * age group	−0.17	1.25	0.897

Note. Abbreviations: SE = standard error. Neurocognitive scores captured by performance on the Wechsler Intelligence Scales. Results based on general linear mixed models with a fixed effect for time, age group, and interaction term, as well as a random intercept and slope. Model for Working Memory Index did not include a random slope because the models failed to converge on a solution due to small sample size.

^aBivariate variable based on age at radiotherapy treatment. Less than 9 years of age at radiotherapy used as the reference level.

Shunt

On average, PRI and PSI scores were 20.1 and 23.6 points lower among patients with hydrocephalus requiring a shunt compared to those without after adjusting for time since RT (without interaction term), respectively (p<0.05).

Maternal Education

On average, VCI, PRI, PSI, WMI, and FSIQ scores were 20.3 (SE=6.8), 17.7 (SE=6.8), 26.3 (SE=8.4), 18.4 (SE=8.2), and 28.6 (SE=5.9) points higher among patients with mother’s who have a college degree, respectively (p<0.05). The change in scores over time did not significantly depend on maternal education level for any of these outcome measures (p≥0.138).

Discussion

Neurocognitive functioning did not change following PRT for patients diagnosed with an LIGG. Younger patients demonstrated poorer performance at baseline on measures of verbal reasoning and working memory, and these differences remained over time. Age at treatment did not appear to alter neurocognitive trajectories for measures of reasoning or working memory.

The current findings differ from prior studies of focal XRT¹² and craniospinal PRT,¹³ which demonstrated slowed development of reasoning and working memory skills following treatment. The tissue sparing properties of PRT may have lessened the increased risk traditionally associated with younger age at treatment. Although change in neurocognitive scores over time did not differ by age, there were significant baseline differences based on age at treatment.

Performance on processing speed measures declined one-point per year on average for children younger than 9 years of age. The susceptibility of speeded processing skills to the effects of radiotherapy are well documented. Radiotherapy can interfere with the development of endothelial and glial cells and is associated with immune mediated inflammatory processes that disrupt myelination and white matter development.¹⁴

Patients with VP shunts displayed IQ scores that were approximately one standard deviation below patients without a VP shunt. Consistent with prior findings,¹⁵ hydrocephalus can have a negative effect on cognitive functioning. The neurocognitive deficits observed in patients requiring a VP shunt were evident prior to radiotherapy, suggesting that that there is an immediate brain insult rather than an interaction with radiotherapy.

Beyond medical risk, maternal education was the strongest predictor of neurocognitive outcomes in our sample. Patient’s with mothers who had at-least a 4-year college degree displayed stronger neurocognitive performance across domains. Sociodemographic factors are known to be important determinants of neurocognitive functioning in the general population and in those with a pediatric brain tumor.^{16, 17}

There are several study limitations. The analyses were conducted with a small sample, and some neurocognitive domains and time points were represented with partial data. Because of the small sample size, we were unable to assess predictors such as tumor location, radiation dosage, or chemotherapy treatment.

To conclude, we observed stable neurocognitive functioning several years following PRT for patients diagnosed with an LIGG. Age at treatment did not appear to influence neurocognitive trajectories for measures of reasoning or working memory, in contrast to prior studies utilizing photon radiotherapy. Even so, baseline differences observed in younger patients suggest there are non-radiotherapy factors associated with age that exert an early and persistent influence on neurocognitive development.

Abbreviation Full term

PRT

Proton beam radiotherapy

LIGG

Low- and intermediate-grade gliomas

XRT

Photon radiotherapy

FSIQ

Full scale IQ

VCI

Verbal Comprehension Index

PRI

Perceptual Reasoning Index

WMI

Working Memory Index

PSI

Processing Speed Index

VP

Ventriculo-peritoneal