Non-optic glioma in adults and children with neurofibromatosis 1

Neurofibromatosis 1 (NF1) is an autosomal dominant disorder with an estimated incidence of 1 in 3000 live births [1]. It is caused by mutations in the NF1 gene, a suppressor of the RAS kinase pathway. As a result of the NF1 protein’s ubiquitous expression and important role in cellular regulation, patients with mutations in the NF1 gene develop e.g., neurofibromas and other neoplasms and pigmentary abnormalities of the skin [2].

Gliomas outside of the optic pathways are one of the most common causes of death in NF1 patients [3, 4]. Non-optic gliomas in NF1 patients are usually located in the brainstem or cerebellum and are almost exclusively reported in children, although non-optic gliomas are known to occur in adults with NF1 as well [5]. Approximately 5% of children with NF1 develop a non-optic glioma, a prevalence more than 100 times greater than in the general population [6]. Non-optic gliomas in NF1 patients are usually low-grade astrocytomas and often have a more benign course than in people without NF1 [7‐9]. Their natural history is only incompletely understood [10]. The occurrence of non-optic gliomas has been found to be associated with the presence of optic gliomas in children with NF1 in some studies [11, 12].

MRI is commonly used to evaluate intracranial tumours in clinical as well as research settings [13]. MRI provides information on the presence, location, and size of the tumour and is a crucial tool in routine clinical care of gliomas in individuals with or without NF1.

In this study, we investigated the natural history, progression, and regression of non-optic gliomas in an unselected cohort of adults and children with NF1.

This study comprises the world’s largest collection of unselected brain MRIs of NF1 patients reported to date. Twenty-four (4.3%) of 562 patients were diagnosed with non-optic gliomas. This overall prevalence is 1.5× to 3× higher than in previously published prospective studies (Table 2) [11, 16‐21].

This difference can be explained in part by the fact that the majority of non-optic glioma patients are asymptomatic (seven of our 24 glioma patients were symptomatic) and could only be detected by offering MRIs to all NF1 patients, as was done in our study.

The median age at first scan in this study is 21.2 years -- much higher than the usual age at diagnosis of non-optic gliomas in NF1 patients (Table 2).

Since our study uses age at first scan with glioma and not age at glioma diagnosis to infer prevalence, cases in younger age groups might be underrepresented. Additionally, the majority of adult NF1 patients with gliomas are asymptomatic and some symptoms of brain tumours are also common in NF1 patients without brain tumours (e.g., headache). Even though it is known that NF1 patients have a lifelong elevated risk for tumour development, brain MRIs are not currently recommended for tumour surveillance in NF1 patients of any age as part of regular clinical care [7, 22].

The largest group of adult NF1 patients with non-optic gliomas was described by Gutmann et al. [5] Only three of 15 tumours for which histology was available were Grade 1, compared to nine of nine in our study. This may be explained by Gutmann et al. only reporting symptomatic tumours, which are more likely to be high-grade than asymptomatic tumours. Other authors reported three adult cases, two of which were asymptomatic [23].

Intracranial gliomas in NF1 patients are usually associated with a better outcome than those diagnosed in the non-NF1 setting [24, 25]. Most of the tumours biopsied in this study were PAs -- the most common histological type of intracranial tumour [26]. The 5-year survival rate of NF1 patients with PA was estimated to be 85% [25]. However, this was calculated for patients who underwent surgery for their gliomas, so the true 5-year survival rate is likely much higher than 85%. In further support of this, none of our brain tumour patients with or without biopsy-proven PA died during a cumulative 134 patient years of follow-up.

While malignant transformation of low-grade non-optic gliomas has been described in NF1 patients [27, 28], this appears to be infrequent [29]. We did not see any cases of malignant transformation in this study.

The majority of non-optic gliomas in children with NF1 reported in the literature have been located in the posterior fossa [9, 30]. We observed the same thing in our study. We also found that the majority of non-optic gliomas in adults with NF1 were located in the posterior fossa. Previous reports suggested that these tumours usually occur in other locations in adults [5], but previous studies of non-optic gliomas in adults with NF1 have been small and focused on symptomatic patients. Most of the patients in our study were asymptomatic.

There were five patients with multiple non-optic gliomas. These lesions could either be multiple primary tumours, or multifocal gliomas. Gliomas in NF1 patients were shown to be multifocal [25]; however, it is impossible to distinguish between multifocal and multiple primary tumours without performing molecular analyses. The chance of having 1 non-optic glioma according to our results is 4.3%; therefore we expected to find one individual with two primary non-optic gliomas and 0 individuals with three primary non-optic gliomas by chance in a series of this size. We found two patients with two gliomas and three patients with three gliomas each. Even though these numbers are small, they support the possibility that a patient is more likely to develop additional non-optic tumours if he or she has already developed one. There is a correlation between the occurrence of optic and non-optic gliomas in people with NF1 [11]; however, a correlation between the development of non-optic gliomas has not previously been suggested. There are case reports of NF1 patients with more than one non-optic glioma, and all of these patients were symptomatic [31]. Only two of the five patients in our study showed symptoms, so NF1 patients can have multiple brain tumours and remain asymptomatic. A publication assessing prognostic factors for NF1 patients with brain tumours found that having multiple tumours is not a risk factor for death [32]; however, it did not distinguish between optic and non-optic tumours.

Tumours in adults with NF1 are said to be higher-grade and carry a worse prognosis than those in children with NF1 [32, 33]; however, we did not find any higher-grade tumours in adults. This difference probably reflects the fact that adults were only included in previous studies if they had aggressive tumours, whereas our study included individuals independent of the presence of symptoms.

In order to compare tumour volumes of younger and older non-optic glioma patients, we separated the cohort into two groups, using the median age at diagnosis (21.2 years) as cut-off. Tumours in patients diagnosed before 21.2 years had a significantly greater volume than those in patients diagnosed later (Mann–Whitney U test, p = 0.02). This is counterintuitive if most tumours in adults arose in childhood or adolescence and tumours only grow in young patients. However, a similar observation has been made for plexiform neurofibromas in people with NF1 -- the most rapid growth occurs during childhood but the tumour volume is inversely correlated with age [34]. In addition to having larger tumours, seven of the nine individuals requiring treatment of their non-optic tumour were diagnosed before 21.2 years of age. One possible explanation for the differences in tumour volume between children and adults is that large tumours in childhood require treatment or have fatal consequences, thus only leaving patients with small indolent tumours to be identified later.

One of the limitations of our study is that patients referred to the NF clinic are probably not representative of the NF1 patient population as a whole. Patients with milder manifestations may be less likely to be referred than those more severely affected. In addition, patients with e.g., brain tumours are more likely to get follow-up scans than tumour-free patients. Another limitation is that even though all parents were offered brain MRIs for their children, parents of asymptomatic children might be more likely to refuse MRI, which requires sedation, than parents of symptomatic children. Therefore, this study is probably biased towards patients with a more severe phenotype. It is important to note that there were patients seen in the NF outpatient department who had higher-grade tumours but were not included in this study.

There is currently no recommended head MRI screening protocol for patients with NF1, whether or not they have non-optic gliomas. We screen NF1 patients without non-optic gliomas every 2 years, patients with asymptomatic gliomas every 12 months, and patients with symptomatic gliomas every 6 months. A much larger longitudinal cohort study would be necessary to determine if this is the optimal approach.

Non-optic gliomas may occur in children as well as adults with NF1, and these tumours should be considered as a possible cause of neurological symptoms that develop in NF1 patients of any age.