PURPOSE
To assess, in a large pediatric population affected by neurofibromatosis type 1 (NF1), the prevalence, incidence, age of onset, and genotype correlation of the main NF1-related ocular signs, including optic pathway glioma (OPG), Lisch nodules (LNs), choroidal abnormalities (CAs), and retinal vascular abnormalities (RVAs).
METHODS
NF1 patients ≤16 years old followed at our institution between 2010 and 2022 were included. Presence of NF1-related ocular signs was assessed at baseline and during follow-up evaluations through slit lamp observation (LNs), near-infrared imaging (CAs and RVAs), and neuroimaging revision (OPG). Patients were categorized according to their genetic variant.
RESULTS
A total of 237 patients were enrolled. Among those, 204 underwent at least 1 follow-up and genetic test was available for 210. Prevalence of OPG, LNs, CAs, and RVAs at baseline was, respectively, 20.7%, 43.5%, 46.8%, and 6.8%. Their incidence during follow-up was 6.4%, 22.4%, 21.4%, and 5.4%, respectively, and the mean age at onset was 6.3±3.6, 7.1±3.0, 6.4±3.0, and 6.6±2.9 years. Patients with truncating mutations presented a higher number of ocular signs than those with non-truncating mutations (1.7±1.0 vs 0.9±0.9, P = .0019).
CONCLUSIONS
Data on prevalence and incidence of NF1-related ocular signs in pediatric patients evidences that the development of these signs seems negligible after the age of 7. LNs and CAs seem to develop independently and, therefore, can be considered as two separate diagnostic criteria. Truncating mutations correlate with a higher number of NF1-related ocular signs phenotype. Note: Publication of this article is sponsored by the American Ophthalmological Society.
N eurofibromatosis type 1 (NF1) is a genetic inherited autosomal dominant disease with a worldwide incidence of approximately 1 in 3000 live births. It is caused by autosomal dominant loss-of-function mutations of NF1 gene (Neurofibromin 1; OMIM 613113) located on chromosome 17q11.2 and encoding for neurofibromin, a negative regulator protein of the Ras/MAPK pathway involved in the control of cell growth and proliferation. The inactivation of neurofibromin leads to the hyperactivation of Ras and its downstream mediators and contributes to tumor formation.
Clinical manifestations are various even among individuals of the same family. Main features include cafè-au-lait macules (CALMs), intertriginous freckling, cutaneous and plexiform neurofibromas and distinctive osseous lesions, and some ocular signs, such as optic pathway glioma (OPG), Lisch nodules (LNs), choroidal abnormalities (CAs), and retinal vascular abnormalities (RVAs). Eye evaluation in children with NF1 not only represents an essential tool to diagnose this condition, because ocular findings might help reaching a full clinical diagnosis, but it is also fundamental to the screening and management of many NF1-related ocular manifestations and complications.
OPG is a relative frequent complication of NF1 occurring in up to 20% of patients. It is a pilocytic astrocytoma, a low-grade neoplasia that may arise from every segment of the optic pathway, from optic nerve to optic radiation, with possible involvement of the hypothalamus. It has commonly an indolent course and only in one-third of subjects OPG manifests itself with vision loss or other ophthalmologic, endocrine, and neurologic symptoms, thus requiring specific treatment with different chemotherapy regimens. Although the diagnosis of OPG requires brain magnetic resonance (MR), peripapillary retinal nerve fiber layer (pRNFL) analysis conducted by means of optic coherence tomography (OCT) has been demonstrated to be an objective and noninvasive tool that helps the diagnosis, monitoring, and management of NF1-related OPG, showing significant correlation with visual acuity ( Figure 1 ).
LNs are melanocytic hamartomas of the iris identified at slitlamp examination as regular, smooth, dome-shaped superficial nodules yellow to brown in color and up to 2 mm of diameter ( Figure 2 ).
CAs are described as bright patchy areas at near-infrared (NIR) reflectance imaging, representing choroidal ovoid bodies consisting of hyperplastic Schwann cells, melanocytes, and ganglion cells ( Figure 3 ). They have been recently added as a new clinical diagnostic criterion of NF1, which is fulfilled when at least 2 CAs are present, as an alternative to the presence of 2 or more LNs ( Table 1 ).
| A- At least 2 of the following in an individual who does not have a parent diagnosed with NF1 B- At least 1 of the following in an individual who does have a parent diagnosed with NF1 |
|
|
|
|
|
|
|
Retinal vascular abnormalities represent another characteristic sign of NF1, although not a formal NF1 diagnostic criterion. RVAs are described as small tortuous retinal vessels with a “spiral/corkscrew” aspect, originating from small tributaries of retinal veins and, as for CAs, visible using NIR reflectance imaging ( Figure 4 ).
Ocular signs, with the exception of RVAs and together with other typical manifestations of NF1, are part of the diagnostic criteria for NF1, first proposed in 1988 by the National Institutes of Health and recently revised ( Table 1 ). NF1 can be clinically diagnosed when at least 2 of the diagnostic criteria (one in an individual with a parent diagnosed with NF1) are fulfilled, but this might be challenging in certain cases because the complete penetrance is reached only by the age of 8 years in 95% of patients, and clinical spectrum might be incomplete in younger children.
In this context, genetic testing acquires great importance, contributing to confirm the diagnosis in patients who do not yet meet the clinical diagnostic criteria or to differentiate NF1 from other conditions presenting common cutaneous features such as Legius syndrome. Nevertheless, molecular testing to identify causative mutations in NF1 is quite complex, mainly because of the large size of the NF1 gene, composed of 60 exons and spanning more than 300 kb of genomic DNA, to the relative lack of mutation hotspots, to the considerable number of splicing mutations and highly homologous NF1 pseudogenes existing in the human genome. A combination of complementary and genomic DNA (cDNA and gDNA) sequence analysis with multiplex ligation-dependent probe amplification (MLPA) has been recommended for molecular confirmation, providing a mutation detection rate of approximately 95%.
Given the high phenotype variability of the disease, a great amount of genotype-phenotype correlation studies has been carried out, confirming that the type of mutation in the NF1 gene does not account for the expressed features of the condition. Except for large deletions of the NF1 gene, representing less than 10% of cases and associated with a more severe clinical phenotype, no clear genotype-phenotype correlations have been established as regards the majority of mutations, that is, intragenic variants, with few reported exceptions. , Furthermore, only few studies analyzed correlations between NF1-related genotype and the ocular signs of NF1.
Therefore, because the prevalence and incidence of ocular features in pediatric populations of NF1 patients, as well as their age of onset, are not well defined, our study aimed primarily to identify, in real life pediatric population, the prevalence of NF1-related ocular signs such as OPG, LNs, CAs, and RVAs, and their incidence during follow-up, focusing on the mean age of onset of each of these signs, to guide the adequate timing of ophthalmologic assessment in children with confirmed or suspected NF1 diagnosis. Furthermore, because CAs have been recently added to the diagnostic criteria for NF1 in association with or as an alternative to LNs, we analyzed the correlation of the development of these two signs.
The second aim of this study was to describe the genetic spectrum of mutations in the same population, searching for possible correlations between ocular features and genetic variants.
METHODS
We conducted an observational cross-sectional study on a large population of subjects followed at our institution from 2010 to 2022 for the presence of confirmed or suspected neurofibromatosis type 1 ( Figure 5 ).
In particular, our population consisted of 850 patients: 35 (4.1%) affected by neurofibromatosis type 2 and 815 (95.9%) with a confirmed or suspected NF1. Among patients examined for NF1, 445 (54.6%) were NF1-affected subjects (male-female ratio 215:230, mean age 15.6 ± 16.8 years) and 370 (45.4%) were suspected cases (male-female ratio 185:185, mean age 22.3 ± 17.8 years). Patients considered as suspected were the parents or siblings of affected children, or patients with only one sign typical of NF1 (mostly CALMs) in absence of any other ocular finding that may have fulfilled the diagnostic criteria during the observation period. We divided the group of patients affected by NF1 in two subgroups according to their age at first evaluation: 150 subjects were >16 years old and 295 were ≤16 years old, therefore classified as pediatric patients.
This study was focused on the pediatric population, and conducted in accordance with the tenets of the Declaration of Helsinki and approved by the Ethics Committee of Padova University Hospital (protocol code 74n/AO/20). Informed consent was obtained from the legal guardian of each enrolled infant. Included subjects were children aged ≤16 years with a diagnosis of NF1 according to the revised diagnostic criteria for neurofibromatosis type 1 of the International Consensus Group on Neurofibromatosis Diagnostic Criteria.
Exclusion criteria were diagnosis of segmental neurofibromatosis, history of any other ophthalmologic disease unrelated to NF1 that may have affected the choroid and/or retina (e.g., uveitis, retinopathy of prematurity, maculopathy, and congenital ocular malformations) or inadequate fundus visualization (e.g., congenital cataract or other media opacities). Patients with poor-quality imaging acquisition were also excluded. After baseline evaluation of all included subjects, follow-up visits were conducted at different time points according to age and to the presence or absence of ocular abnormalities. , At each evaluation, patients underwent a complete ophthalmologic examination including visual acuity assessment using age-appropriate visual function tests, , slit lamp biomicroscopy examination focused at LN detection, and fundus examination using indirect ophthalmoscopy.
The presence or absence of NF1-related CAs and RVAs was evaluated using Spectralis HRA + OCT (Heidelberg Engineering) in NIR reflectance modality, capturing for each eye an image of the posterior pole using a 50° lens centered onto the fovea, as previously reported. Furthermore, the presence or absence of NF1-related OPG was obtained for each patient at each visit. We also collected, when available, data about the molecular NF1 genetic testing of the enrolled patients. Over the years, different methods of molecular analysis of the NF1 gene have been employed, including mutation screening approaches (high resolution melt analysis), complementary DNA (cDNA) analysis, and a next generation sequencing (NGS)–based sequencing protocol analyzing NF1, SPRED1 , and other genes associated with CALMs.
The nomenclature for NF1 variants was based on the reference isoform NM_000267.3. Screening for whole gene deletions was carried out through fluorescence in situ hybridization analysis or MLPA analysis, which also allowed the detection of single/multiexon intragenic deletions/duplications (SALSA MLPA kits P081/P082; MRC Holland). When DNA was available from other family members, segregation analysis was carried out through bidirectional Sanger sequencing or MLPA analysis depending on the variation. Patients with pathogenetic (P) or likely pathogenetic (LP) variants, according to the American College of Medical Genetics and Genomics (ACMG) criteria, were divided into four different groups according to the type of their genetic variant: frameshift and nonsense mutations, missense and inframe mutations, splicing and startloss mutations, other mutations (deletions and translocations).
Frameshift and nonsense were considered truncating mutations, whereas missense and inframe were considered non-truncating mutations. , This kind of separation was chosen to avoid subcategories with a small number of patients and to homogenize groups for patient’s age, according to previous classification. Then, we searched for correlation between genotype and NF1-related ocular manifestations (OPG, LNs, CAs, and RVAs) presented by each subject at the age of 8 years, when the penetrance of the disease is complete in 95% of cases, or at the last available eye examination for children ≤8 years old.
STATISTICAL ANALYSIS
Studied parameters have been summarized according to the methods provided in the descriptive statistics: mean value and standard deviation (SD) for the quantitative continuous parameters and frequency, absolute and relative (percentage), for the qualitative ones. Prevalence and incidence rates of ocular signs have been expressed as percentage of number of cases over people at risk with its 95% CI. Overall and class of age-specific rates have been computed. Probability of developing specific ocular signs during the study period was estimated by means of logistic regression models, which also provided odds ratios (ORs) and their 95% CIs. Probability functions have been graphically reported. Mean age of patients at the time of developing ocular signs has been computed and compared with that of patients never showing the same specific sign during follow-up by means of two-sided Student t test for independent samples. Correlation between CAs and LNs has been investigated in terms of patient’s age when developing each sign. Specific analyses have been performed to correlate the presence of ocular signs and genetic variant type. Descriptive frequency distributions have been provided by variant types and truncating/non-truncating mutations. Comparisons of the qualitative variables among groups were made by Fisher exact test, whereas for quantitative variables an ANOVA model was used. Direct comparisons between groups were adjusted by a Tukey-Kramer test for multiple comparisons. All the analyses have been made using SAS 9.14 statistical software (SAS Institute). Statistical tests have been interpreted as significant when P < .05.
RESULTS
POPULATION
A total of 237 patients were eventually included in this study: 115 (48.5%) were male and 122 (51.5%) female. As regards ethnicity (White, Black, Hispanic, Asian), our population mainly consisted of White subjects, except for 8 Black patients (3.4%) and 3 Hispanic patients (1.3%). Mean age at first evaluation was 7.2±4.2 years. Thirty-three patients (mean age 9.9±4.7 years) underwent only one examination, whereas for the other 204 (mean age 6.7±3.9 years) at least one follow-up visit was available. Of these, 63 patients had 2 observations, 51 had 3, 30 had 4, 22 had 5, 12 had 6, 12 had 7, 6 had 8, 5 had 9, and 2 had 10; only 1 patient had 11 observations.
Among the 204 patients presenting at least one follow-up visit, the mean time between the first and the last evaluation was 2.8±1.6 years. Considering NF1-related ocular signs, that is, OPG, LNs, CAs, RVAs, at their first examination 78 patients (32.9%) presented with no sign, 67 patients (28.3%) had at least 1 sign (15 presented OPG, 22 LNs, 29 CAs, 1 RVAs), 67 patients (28.3%) had 2 signs (6 presented OPG and LNs, 8 OPG and CAs, 47 LNs and CAs, 3 LNs and RVAs, 3 CAs and RVAs), 22 patients (9.2%) had 3 signs (16 presented OPG, LNs and CAs, 5 LNs, CAs and RVAs, 1 OPG, LNs and RVAs), and only 3 subjects (1.3%) showed all studied signs.
PREVALENCE OF NF1-RELATED OCULAR SIGNS
The prevalence of NF1-related ocular signs at the first evaluation is reported in Table 2 . For each ocular sign, all positive subjects were significantly older than those without the same specific sign, except for OPG. Dividing our subjects according to their age at first evaluation into three subgroups (≤5 years, 5-10 years, and >10 years), we found that the prevalence of OPG was 19.5% (16 of 82 patients), 20.0% (17 of 85 patients), and 22.9% (16 of 70 patients), respectively ( P = .8632) ( Figure 6 , A). The estimated risk of developing OPG during the study period, that is, up to 16 years of age, increased by 2.4% every year (OR 1.024, 95% CI 0.950-1.104) ( Figure 7 , A). According to the age subgroup categories, LNs were present in 17 patients (20.7%) of the younger group, in 42 patients (49.4%) of the middle group and in 44 patients (62.9%) of the older group ( P = .0001) ( Figure 6 , B).
| NF1-Related Ocular Sign | No. of Patients | Prevalence, % (95% CI) | Age, y, Mean±SD | P Value | |
|---|---|---|---|---|---|
| OPG | Present | 49 | 20.7 (15.3-26.0) | 7.5±3.9 | .5348 |
| Absent | 188 | 7.1±4.3 | |||
| LNs | Present | 103 | 43.5 (36.9-50.0) | 9.0±3.7 | <.0001 |
| Absent | 134 | 5.8±4.0 | |||
| CAs | Present | 111 | 46.8 (40.3-53.4) | 8.7±3.7 | <.0001 |
| Absent | 126 | 5.4±4.1 | |||
| RVAs | Present | 16 | 6.8 (3.4-10.1) | 10.0±2.9 | .0045 |
| Absent | 221 | 7.0±4.2 | |||
The estimated risk of developing LNs during the study period increased by 22.4% every year (OR 1.224, 95% CI 1.140-1.315) ( Figure 7 , B). CA prevalence in the 3 subgroups was 17.1% (4 patients), 57.7% (49 patients), and 68.6% (48 patients), respectively ( P = .0001) ( Figure 6 , C). The estimated annual risk of developing CAs during the study period was 26.4% (OR 1.264, 95% CI 1.173-1.362) ( Figure 7 , C). RVA prevalence in the 3 subgroups was 0%, 9.4% (8 patients), and 11.4% (8 patients), respectively ( P = .0022) ( Figure 6 , D). The estimated annual risk of developing RVAs during the study period was 19.7% (OR 1.197, 95% CI 1.051-1.364) ( Figure 7 , D). Figures 6 and 7 graphically represent the aforementioned results.
INCIDENCE OF NF1-RELATED OCULAR SIGNS
During the follow-up period, among the 188 subjects not affected by OPG at first evaluation, 12 patients (6.4%) developed OPG at a mean age of 6.3±3.6 years. As regards LNs, incident cases during the observation period were 30 (22.4%), and mean incidence age was 7.1±3.0 years. Analyzing CAs, incident cases were 27 (21.4%) and had a mean age of 6.4±3.0 years. Among the 208 subjects not presenting RVAs at their first evaluation, 12 (5.5%) developed this sign during follow-up, at a mean age of 6.6±2.9 years. Data are summarized in Table 3 .
