Our investigation focuses on a subgroup of patients with NF1 and cerebellar FASI on their MRI scans during childhood. We selected them a priori to investigate cerebellar FASI and their clinical impact. Recently, there has been much interest in the expanding roles of the cerebellum and its important role in non-motor in addition to motor tasks . Our patients represent what is commonly seen in clinical practice since many patients with NF1 have FASI [12, 13, 20], and 23.5–84% of these patients also have FASI in the cerebellum/ brainstem [13, 14, 17, 21, 22].
The diagnosis of NF1 was made early in our cohort. A positive family history of NF1, reported commonly in our patients (48%), and similar to other studies (e.g. 35% , and 54.8% ,), and the frequent presence of café-au-lait macules likely prompted an early referral and diagnosis. Tumors (especially neurofibromas), developmental delay, and learning disabilities occurred commonly in our patients, while headaches, ADHD, and seizures were less frequent clinical features. Their prevalence is similar to prior reports [24, 25]. The prevalence of other features including café-au-lait spots, axillary and inguinal freckling, Lisch nodules, and sphenoid dysplasia were also similar to other studies [17, 26].
Visual signs e.g. decreased visual acuity, abnormal visual fields, and optic discs pallor, detected commonly on examination, were important markers associated with optic pathways gliomas. Indeed, optic pathways gliomas occurred commonly and were mostly present on the initial MRI scans in our study. These anticipated findings have been reported previously and are consistent with impaired optic nerve function in some patients with NF1 and optic pathways gliomas .
There was high prevalence of optic pathways gliomas (44.7%) in our cohort in comparison with several studies (4–25%) [17, 24, 25, 27]. A large study in patients with NF1 reported a significant association between the presence of FASI and optic pathways gliomas (odds ratio: 2.1, 95% CI: 1.2–3.6) , which may account for our finding since all our patients had FASI. However, in a study of 31 children with NF1 who had brain MRI, there was no correlation between FASI (present in 27 patients) and optic pathways gliomas, which similar to our study was noted in 14 (45%) of the patients . In another study investigating the natural history of FASI, optic pathways gliomas were present in 33% of 46 patients with NF1 . The higher prevalence of optic nerve gliomas in our study may also be due to selection bias, since such patients are more likely to be seen in a tertiary hospital and have neuroimaging. Similar referral and selection bias may explain the higher proportion of our patients with NF1 who developed other neoplasms (24%) than the prevalence of neoplasms reported in other studies (4–10.7%) [24, 25].
On the other hand, cerebellar motor signs were absent in our cohort, consistent with a normal cerebellar motor function despite the presence of cerebellar FASI. While clumsiness, lower manual dexterity score, and impaired fine motor skills or coordination has been reported in NF1 patients, the occurrence of frank cerebellar ataxia has not [11, 12, 20, 28, 29], unless other complicating factors are present, such as an expanding posterior fossa or spinal cord tumor or high cervical cord lesions compressing the cord [1, 6].
The nature of FASI is unclear but may be related to increased fluid within the myelin associated with hyperplastic or dysplastic glial proliferation as suggested in a study using newer MRI techniques, e.g. multi-exponential T2 relaxation and diffusion MRI including diffusion tensor imaging and diffusion kurtosis imaging . Developmental anomalies such as hamartomas, dysplasias, and heterotopia would not be anticipated to produce reversible signal abnormalities [13, 21].
Studies documenting the neuroimaging-pathological correlation of FASI are rare. In one study, three pediatric patients with NF1 had autopsy which showed that the pathological correlation of FASI were areas of fluid-filled vacuolar or spongiotic change . There was no evidence of demyelination, inflammation, gliosis, stainable material, or axonal damage. Diffusion tensor imaging on 50 children with NF1 showed higher ADC values not only in FASI but also in normal appearing white matter in patients in comparison to control patients without NF1, reflecting an increase in the magnitude of water molecules diffusion and microstructural damage. The specifically higher diffusivity (measured in eigenvalues) found only in FASI was consistent with microstructural abnormalities caused by decreased axonal packing, intramyelinic edema, vacuolation, or fluid accumulation . Fractional anisotropy (FA) was significantly lower in FASI located in the cerebellar white matter only. In another study of 27 children and young adults with NF1, FA decreased in the cerebellum, thalamus, and basal ganglia but only in NF1 patients whose FASI decreased in number and volume in those regions, suggesting persistent microstructural damage even when FASI disappear . Similar conclusions were reported in 15 children with NF1, where ADC values were high in normal appearing brain and highest in FASI in comparison to healthy controls. The ADC values in the locations where FASI regressed in some patients, were higher than normal appearing brain in these patients, suggesting that macroscopic resolution of FASI on MRI does not necessarily lead to the full resolution of the microstructural abnormalities .
Brain MRI scans in our children with cerebellar FASI showed that FASI rarely occur in isolation and most commonly occur in multiple brain regions in addition to the cerebellum, especially the brainstem and basal ganglia as noted before [15, 17, 19, 22], i.e. FASI showed no predilection to a specific brain region in our patients.
There was a suggestion that age at symptom onset or first clinic visit may correlate with where FASI develops, since FASI that developed in the basal ganglia occurred in patients who were relatively older at age of symptom onset. There was a tendency to develop hypothalamic FASI in patients who were relatively younger at the age of their first clinic visit.
FASI developed below the age of 2 years in different brain locations in a few of our patients, which is not considered to be typical by some authors , and especially as myelination is incomplete [17, 33]. The total or regional number of FASI was independent of the age of symptom onset or age at first clinic visit, consistent with their asymptomatic development.
Prior studies have shown that cognitive deficits were more common in patients with NF1 and thalamic FASI [11, 12], especially if the lesions are well circumscribed . However, the association is controversial since other studies showed no such association with executive/ cognitive function in terms of the presence, number, size, or location of FASI [20, 22, 23]. In our investigation, learning disabilities, reported in 30–65% of children with NF1 , and developmental delay (for which there was only a trend), occurred significantly less commonly in patients with NF1 and thalamic FASI. However, our patients did not have formal neuropsychological testing. In addition, our study design is different from the aforementioned studies [11, 12], since all our patients also had cerebellar FASI as part of the inclusion criteria. We only found a trend for developmental delay to occur more commonly in patients with hypothalamic FASI. Furthermore, there was no significant relationship between any of the clinical symptoms and several combinations of brain locations where FASI were present. A longitudinal study of cognitive function and FASI with long term follow up showed that if patients with NF1 had FASI then their IQ is less during childhood but their IQ increases to average values when FASI disappears in early adulthood . One study reported that patients with cerebellar FASI had significantly lower full scale IQ and verbal IQ scores in comparison with NF1 patients without cerebellar FASI . We were not able to investigate this finding since our patients did not undergo detailed cognitive testing and all had cerebellar FASI. However, several of our patients did not have developmental delay or learning disabilities, suggesting that having cerebellar FASI per se does not necessarily leads to impaired development or learning.
As for clinical signs, visual fields and especially funduscopic abnormalities were associated more commonly with patients who had cerebral FASI. This can be explained through confounding with the presence of optic pathways gliomas on baseline MRI, since cerebral FASI was seen significantly more commonly in patients whose baseline MRI also showed optic pathways gliomas. Strabismus was significantly associated with the presence of FASI in the basal ganglia. We found no studies that directly link strabismus with disorders of the basal ganglia. In one study of 213 children with cerebral palsy, strabismus was reported in 3 of 15 children with dyskinetic cerebral palsy. Abnormalities in the basal ganglia are implicated in dyskinetic cerebral palsy . Our finding may have arisen by chance. Further corroboration is needed before any definitive conclusions can be made.
The total number of FASI at baseline MRI was significantly less in patients with ADHD and more if a first degree relative had NF1 and also more in patients with decreased visual acuity. This latter association may have arisen through the occurrence of optic pathways gliomas, which themselves were associated with higher mean numbers of FASI in the supratentorial regions and more specifically in the cerebrum. The former two associations i.e. ADHD and family history of NF1, may be caused by genetic factors/ predisposition but this is speculative. FASI was not associated with the presence of ADHD in 76 NF1 patients, who were preselected and did not have epilepsy or optic nerve glioma .
The relationships between decreasing maternal age at conception and the increasing number of FASI at baseline (statistically significant) or the increasing presence of learning disabilities and ADHD (statistical trends) are intriguing; albeit, small in magnitude. They raise the question on whether decreasing maternal age at conception per se or through confounding factors such as socioeconomic status or educational achievement , somehow influences the development of FASI, especially in the brainstem region, or adversely affects learning and behavior. It will need further validation and their clinical significance in patients with NF1 is unclear at this point since adverse health outcomes (including cognitive disability and ADHD) and maternal age at conception, whether in younger or older mothers, occurs independently of NF1 [37,38,39]. In one study in patients with NF1, learning disability was not associated with maternal age at conception. The presence of ADHD in the cohort was not reported .
A few FASI showed contrast enhancement, mass effect, or both. Such findings are reported to occur rarely [13, 14, 41, 42], but are not considered to be typical of FASI in NF1 [13, 21]. They are concerning for the development of neoplasms . Mass effect was seen mostly initially and tended to resolve in our patients. On the other hand, contrast enhancement tended to persist on repeat MRI scans. In a few of our patients contrast enhancement was still present 16 months - 11 years of follow up. On only one occasion, a single FASI located in the periventricular region developed a mass effect later and became malignant (patient #50 in Table 8). In two other patients, a cerebellar FASI enlarged and displayed contrast enhancement in each patient. Their biopsies revealed a benign ganglioma and a low grade pilocytic astrocytoma (patients #8 and 41 in Table 8). Tumors in the cerebellum and cerebral hemispheres are uncommon in patients with NF1 , and most are benign [6, 8]. However, they occur more commonly in patients with FASI . In 5 of 46 patients with NF1, FASI enlarged, showed mass effect, and enhanced . Biopsy was performed in one other patient, whose brain MRI with contrast was normal initially but who then developed several FASI and also a large enhancing mass in the splenium of the corpus callosum. When the latter increased further in size, a biopsy revealed grade II astrocytoma. The rest of the patients with enhancing FASI were observed and no further follow up information on their outcome was given. The authors advised a wait-and-watch approach in such otherwise concerning cases, since current medical wisdom dictates that mass effect and enhancement of a lesion are suggestive of a tumor but at the same time tumor regression without treatment is recognized in NF1 [2, 5, 13]. In summary, enlargement in the size of FASI is well documented. However, infrequently a few FASI tend to also develop a transient mass effect and enduring contrast enhancement with time [14, 41, 42]. Sometimes this enduring contrast enhancement resolves , and only rarely does malignancy develops. Therefore, we support the clinical practice of repeating brain MRI in patients with NF1 who develop new clinical symptoms or signs (e.g. those suggestive of raised intracranial pressure or hemiparesis ,) or in patients, whose FASI show mass effect and enhancement, since very rarely they can become malignant. If FASI becomes cystic, as we found in one of our patients, a tumor is highly likely [5, 43]. It is still unknown whether the same neuropathology occurs in the lesions that enhance or show mass effect in comparisons with FASI that do not enhance or show mass effect .
Limitations of our study include incomplete ascertainment of patients, missing data, variable durations of follow up, variable ages at which the MRI scans were done, non-uniform MRI magnetic strength and protocols used in these patients, difficulties associated with measuring the diameter of ill-defined FASI, and parental report of learning disabilities were not formally verified through comprehensive cognitive assessment.