In the present study, the radiologic features of the CNS were analyzed with with CT and/or MRI scanning in a series of 125 patients with neurofibromatosis 1 (NF1) from Northern Finland. The patient series was well representative of this hereditary disorder, and it consisted of patients with a wide age range and marked clinical variability. The patients belonged to a larger group collected for an epidemiologic and clinical study and also included some milder cases discovered in family studies.
The present study, which is the first extensive radiologic analysis of Finnish NF1 patients, confirms that NF1 is, even radiologically, an extremely variable disorder, with the CNS affected in most patients. Although certain lesions are common in NF1, the radiologist may encounter and need to explain unexpected findings, of which FSNF is a good example.
1. Of the 124 patients with imaging of the head, 77% showed pathologic findings. T2 hyperintense lesions seen on MRI were the most common findings, with 77% of the children and 33% of the adults affected. Optic gliomas were seen in 29% of the patients: in 44% of the children and 15% of the adults. Other manifestations of the head included intracranial tumours, bony lesions and plexiform neurofibromas. Spinal changes were detected in 75% of the 76 patients studied: in 52% of the children and 91% of the adults. Postural changes were most common (38% of the patients), followed by CSF space changes (29%) and neurofibromas (28%). Other, less common spinal lesions were seen in 49% of the patients.
T2 hyperintense lesions were most common in children in the age group of 5 to 9 years and usually appeared bilaterally in the pallidal globes. In some cases they could have been used as a diagnostic aid, although all patients in this study had been diagnosed clinically before the MRI scanning. T2 hyperintense lesions were equally common in females and males, but lesions in the globus pallidus and putamen were more common in females. On CT, the corresponding lesions were seen as hypodense areas only in a few patients.
Age at the diagnosis of optic gliomas had dropped from 9.6 years in the patients born in the 1970s to 3.8 years in those born in the 1990s. For the purposes of radiologic diagnosis, both CT and MRI were reliable, but MRI is more accurately able to visualize the lesions of the optic tracts and hydrops of the optic sheet. The intraorbital and/or prechiasmal optic nerve was affected in 94% of the patients, the chiasm and/or hypothalamus in 58% and the posterior parts of the optic system in 14%. Optic gliomas were bilateral in most cases, although prechiasmal gliomas were more common unilaterally. Gliomas were equally common in females and males.
Brain tumours other than optic gliomas were found in 8% of the patients, including astrocytomas (5%). In one family, both the mother and a son were affected. In one patient, a verified astrocytoma disappeared spontaneously.
Spinal lesions were more common in adults than in children, but even young children could be severely affected: two children had severe scoliosis requiring operation, and two had quite prominent dural ectasias of the cervical/lumbar level. A young man had widespred neurofibromas of the spinal nerves and multiple medullary tumours. Medullary affisions were rare. Postural changes were mostly mild and commonest in the thoracic spine. Dural ectasias were most common in the lumbar area. Spinal neurofibromas varied in extent and size, and the largest tumours were found in the lumbar area. An exceptional family with rare familial spinal neurofibromatosis (FSNF) in four adult individuals in two generations were characterized by MRI. In them, the main radiological feature was the presence of bilateral multiple neurofibromas at all levels of the spine. Compression of the cord was seen in five patients, four of them with FSNF.
2. At follow-up, T2 hyperintense lesions (mean follow-up time 2.1 years) remained unchanged in 35% and diminished in 25% of the children. Stable and diminishing lesions, which were more common in older children, were found in 20% and mixed-behaviour lesions in 15% of the children. The number and size of the lesions were larger in children than in adults. In one patient, a glioblastoma appeared at the site of a T2 hyperintense lesion, suggesting that the development of these lesions shoud be followed radiologically, especially when their location is atypical.
During a mean follow-up of 6.4 years for the children and 5.5 years for the adults, the optic gliomas remained stable in 84% of the patients. Regression was seen in three children, and one child showed first progression and then regression of the glioma. Eight patients with normal initial CT findings were found to have an optic glioma on the following MRI, usually with a slight thickening of the optic nerve.
In a boy with spinal lesions, progression of the dural ectasias and scalloping of the posterior parts of the vertebrae were seen during follow-up for 12 years.
The variable appearance of the CNS changes in the different age groups and the follow-up data show that NF1 is not a stable disorder, but that different lesions in different individuals show variable behaviour. The exact time of appearance of the lesions is not known, but they mostly appear in early childhood. Some lesions, such as T2 hyperintense lesions, are transient, while lesions such as optic gliomas tend to remain stable. Malignant development of a hamartomatous CNS lesion may occur unexpectedly at any age, even at the site of a benign T2 hyperintense lesion. The present study confirms that radiologic follow-up of NF-related lesions of the head and spine seems warranted in some cases as part of the clinical follow-up of the patients.
3. Macrocephaly was found in 27% of the present patients, which percentage is of the same magnitude as reported previously. Brain measurements of 14 macrocephalic and 14 age- and sex-matched normocephalic patients indicated that no specific area or structure of the brain could be considered responsible for the increased growth of the brain.
It was observed that all studied macrocephalic children had T2 lesions (mean number 5.6/patient), while 59% of the normocephalic children were affected (2.6 lesions). On the other hand, the children without T2 hyperintense lesions did not have macrocephaly, while 27% of those affected with them had macrocephaly. The positive correlation between the appearance of macrocephaly and T2 hyperintense lesions may signify the presence of a common pathogenic factor in their development. No corresponding association between optic gliomas and macrocephaly was observed.
Because NF1 is a progressive and highly variable disorder, the role of the radiologist is not limited to the diagnosis or follow-up of the patient, but also includes the detection and interpretation of unexpected lesions at any level of the CNS. Radiologic studies with MRI therefore also include elements of screening and early detection – or exclusion – of possible clinically significant lesions. Early screening for the lesions of the head with MRI is useful, and can recommended in most cases for newly diagnosed NF1 patients in order to obtain an overview of the extent of the disease. The presence of T2 hyperintense brain lesions confirms the diagnosis, and the presence and extent of optic gliomas will be revealed. Follow-up studies are necessary in patients with T2 hyperintese lesions at exceptional sites and in patients with optic gliomas with symptoms. As there is a tendency to concentrate the primary studies and follow-up of patients with NF1 into multidisciplinary centers, the role of the radiologist will be crucial in proper planning of follow-up.