Referral for Ophthalmology Evaluation and Visual Sequelae in Children With Primary Brain Tumors

IMPORTANCE Visual impairment in children with brain tumors has received limited attention, as mostpediatricneuro-oncologyclinicaltrialsneitherrequireophthalmologicevaluationonenrollment nor monitor effects of treatment on visual function during and after treatment. OBJECTIVE To investigate ophthalmology referral patterns for children with primary brain tumors, the prevalence of visual sequelae, and the association between tumor characteristics and vision-related diagnoses. DESIGN, SETTING, AND PARTICIPANTS This retrospective cohort study included 141 children with primary brain tumors treated at Loma Linda University Children’s Hospital and Eye Institute, a university-based tertiary referral center, between January 2013 and September 2017. Data analysis was completed in March 2019. INTERVENTION Comprehensive ophthalmologic evaluation for children with primary brain tumors. MAIN OUTCOMES AND MEASURES Percentage of patients with ophthalmology evaluation, prevalence of abnormal ophthalmic findings, and their association with tumor characteristics. total 141 children (73 [52%] male; median age, 7 [0-18] years) with primary brain enrolled in this study. Seventy-three patients (41 [52%] male; median [range] age, 8 [0-17] years) never had formal ophthalmologic evaluation. Sixty-eight patients (32 [48%] male; median [range] age, 7 [0-18] years) evaluated by 1 of 4 board-certified, fellowship-trained overall survival for patients who had eye examination was not significantly different from those who did not (mean [SD] survival, 78.3% [6.2%] vs 84.9% [4.7%]). Median (range) time from to initial ophthalmologic evaluation was 9 (0-94) months. Only 10 of 68 children (15%)


Introduction
Brain tumors are the most common solid tumor and the second leading cause of cancer death in individuals aged 19 years and younger in the United States and Canada. 1 Childhood brain tumor incidence varies by country from 1.12 to 5.14 cases per 100 000 persons, with the highest incidence in the United States. 1 With improvements in diagnosis and treatments, 5-year overall survival is greater than 60%. 1,2 Children with optic pathway and suprasellar tumors often have visual symptoms as a presenting feature, 3 but tumors in other areas of the brain may also lead to permanent visual impairment, 4,5 even without symptoms. One study 5 at 2 large tertiary referral centers of 139 patients with posterior fossa tumors found symptoms in less than 50% of the cohort and reported that tumors with more aggressive growth patterns had significantly worse visual outcomes, including 17% with visual acuity less than or equal to 20/40. Another study 6 of 182 patients with posterior fossa neoplasms found only 27% with ophthalmologic symptoms, and while esotropia was more common (29%), visual acuity was less than or equal to 20/40 in 10% of patients.
Brain tumors can alter the normal neuroanatomical structures of the visual system, leading to visual impairment and dysfunction. 4,5 Visual impairment in childhood is associated with lifelong effects for children and their families 4 and may affect self-perception, childhood development, education, driving eligibility, employment, and quality of life. 7,8 As the survival of patients with brain tumors improves, untreated visual sequelae could be an important factor affecting quality of life.
Previous studies have focused largely on the effects of the tumor and/or treatment on vision and alignment 9,10 with little attention to referral rate for ophthalmologic evaluation and the prevalence and association of visual impairment with tumor characteristics. This study adds to previous reports and recapitulates that prevention of permanent vision loss requires referral for ophthalmologic evaluation of children enrolled in clinical trials for brain tumors.

Methods
This study was approved by the institutional review board at Loma Linda University. A waiver of informed consent was granted by the board given the retrospective nature of the study and minimal risk to patient care. Medical records of all children (age 0-18 years) with primary brain tumors from January 2013 to September 2017 at Loma Linda University Medical Center and Children's Hospital were reviewed. Codes from the International Classification of Diseases, Ninth Revision (ICD-9) (191.0-when available: date of examination, chief concern(s), visual acuity, confrontation visual fields, slitlamp examination findings, dilated fundus examination, sensorimotor examination, cycloplegic refraction, visit assessment, and treatment plan. Visual acuity and confrontation visual field examinations were performed by a certified ophthalmology technician or a certified orthoptist in addition to a board-certified pediatric and/or neuro-ophthalmologist. Humphrey visual field testing, optical coherence tomography (OCT) of the retinal nerve fiber layer (RNFL), and color fundus photographs were obtained from cooperative patients.

Statistical Analysis
Bivariate Pearson correlation coefficient (2-tailed), Spearman correlation coefficient, χ 2 test, and t test were used to determine differences in frequency of each visual sequela in different groups identified by patient demographic characteristics, tumor characteristics, and treatments. Logistic regression with confidence intervals was calculated to determine likelihood of outcomes. Statistical analysis was performed using SPSS statistical software version 25 (IBM). Five-year overall survival was calculated using the Kaplan-Meier method with log-rank test to calculate differences between groups. Median survival time was determined using the reverse Kaplan-Meier survival method. A 2-sided P < .05 was considered statistically significant. The data analysis was completed in March 2019.

Results
A total of 141 patients (73 [52%] male; median [range] age, 7 [0-18] years) with primary brain tumors were included (Table). There were 100 individuals who were newly diagnosed and 41 who were seen for follow-up evaluation (brain tumor was diagnosed prior to study enrollment). The most common tumor type was glioma, whereas other tumor types (eg, ependymoma and atypical teratoid rhabdoid  Visual field defects were found by confrontation in 13 patients (19%) and by Humphrey visual field testing in 5 (7%); only 3 (4%) had symptoms. A total of 7 patients (10%) were able to perform Humphrey visual field testing. Visual field defect was the most common visual impairment among patients without visual symptoms (9 of 58 patients [15%]; ρ = 0.33; P = .007), although no association with any single tumor type or location was found.

Discussion
In the studied cohort, fewer than 50% of children with primary brain tumors were referred for ophthalmologic evaluation. To our knowledge, this is the first study to report referral rates for ophthalmologic evaluation at a university-based, tertiary care center and provide comprehensive data on visual impairment. Ophthalmologic symptoms were similarly uncommon among those with or without ophthalmologic evaluation whose median age, sex distribution, overall survival, and follow-up time with treating clinicians were comparable (Table). When matched with previous, similarly sized cohorts, ophthalmologic symptoms in our population were less frequent, which likely limited referral rate for ophthalmologic evaluation.

JAMA Network Open | Pediatrics
Another possible explanation for low referral rates in our cohort is the difference in tumor characteristics between those with and without ophthalmologic evaluation. Fewer patients were referred for ophthalmologic evaluation with less aggressive tumor pathology; however, most patients with tumor location in the temporal, parietal, and occipital lobes were not referred for ophthalmologic evaluation (79%). Although these patients were under the care of the same neurooncologists, neurologists, and neurosurgeons who referred the patients for ophthalmologic evaluation, there is no standard guideline for referral for ophthalmologic evaluation that accounts for both tumor grade and location.
The prevalence of specific diagnoses in our cohort is comparable to results found in similarly sized cohorts. A study 9 of 92 children with brain tumors found 15.2% with undiagnosed visual field defects, which is similar to the current report (19%). Description of papilledema in our cohort (35%) was also similar to rates ranging from 13% to 34% reported in previous studies. 3,11,12 We note the exception is prevalence of strabismus (60%), which was nearly triple that found in previous reports (6%-29%). 3,11,12 With more than 40% of patients with tumor location in the posterior fossa without ophthalmologic evaluation, the actual prevalence of strabismus may be higher. children with brain tumors reported a 38% incidence of visual difficulties at any point in their course of treatment. 8 Although our report of 90% visual impairment seems high in comparison, previous reports of visual impairment in pediatric brain tumors have omitted publication of referral rates for ophthalmologic evaluation. Review of the current large, multicenter, cooperative clinical trials for pediatric brain tumors confirms no standard to refer for ophthalmologic evaluation. 14 Given a low percentage of patients with visual symptoms in our cohort without ophthalmologic evaluation, tumor management may have been prioritized over ophthalmologic evaluation despite evidence that treatment may affect vision. 15 Additionally, the absence of a standard to refer for ophthalmologic evaluation may explain underestimated prevalence of visual impairment in previous reports.
Both afferent and efferent visual impairment in childhood may result in lifelong effects for the child and the family, 4 and may also affect self-perception, childhood development, education, driving eligibility, employment, and quality of life, 7,8 especially as the overall survival of patients with brain tumors continues to improve. Our report highlights that neither clinicians nor caregivers can rely on a patient's report of visual symptoms, especially patients aged 3 years or younger. A referral for ophthalmologic evaluation is recommended for all patients enrolled in clinical trials to treat brain tumors, which includes age-appropriate visual acuity, visual field, and ocular motility evaluation; persistent papilledema eventually had tumor progression; 1 also had worsening homonymous hemianopia associated with tumor recurrence.

Limitations
This study has limitations. We studied the visual sequelae of children with primary brain tumors who were referred for ophthalmologic evaluation without a standard protocol. There may also be a few patients with ophthalmologic evaluation outside our institution whose visual impairments have not been evaluated in this cohort, although we estimate this to be less than 5% given complexity of coexisting neurologic conditions in this population whose insurance constraints require in-network referral for pediatric ophthalmologic evaluation. Additionally, the population studied was less than half of the total cohort, which subjects the conclusions in this study regarding the prevalence of visual sequelae in children with primary brain tumors to confounding and referral among other biases. A large, prospective study and/or required ophthalmologic evaluation for all children enrolled in clinical trials to treat brain tumors may further clarify associations of visual sequelae and tumor characteristics.

Conclusions
In this study, more than 50% of children with primary brain tumors were not referred for ophthalmic evaluation. The time for a multidisciplinary, comprehensive approach to the diagnosis and management of visual impairment in pediatric patients with brain tumors is now. Given significant advances in technology, OCT analysis of the RNFL and ganglion cell complex in children as young as 3 years routinely complement a thorough pediatric and/or neuro-ophthalmology evaluation at our institution. These and other tests alert the patient, their family, and other health care professionals to details regarding microscopic, neuroanatomical changes in the retina and optic nerve before loss of function in some cases. As such, cure may be defined as prevention of permanent vision loss.