Article Text


Use of MRI and audiological tests in presymptomatic diagnosis of type 2 neurofibromatosis (NF2)

Statistics from

Editor—We have previously reported on a large clinical study of patients with neurofibromatosis 2 (NF2)1and described the usefulness of audiological screening in early detection of VS in NF2.2 Although DNA diagnosis is possible by linkage analysis and by detecting the NF2 family specific mutation, this is not possible in all cases. Only 40% of sporadic patients (new mutations) have an identifiable NF2 mutation3 and as 50% of NF2 patients have no family history,4 up to 30% of those at risk of NF2 will not be able to have a DNA presymptomatic test. While it is possible to calculate residual risks of NF2 from age at onset of symptoms curves,5 a more useful measure would be the level of risk reduction from a normal cranial MRI scan. A database recording information on affected subjects with NF2 was set up in Manchester in 1989. Patients were actively sought across the UK from paediatricians, neurosurgeons, ENT surgeons, geneticists, and dermatologists. We have used this database to determine the age at which MRI detects asymptomatic tumours in subjects with NF2 and the value of audiological screening. We have also created life curves for risk reduction in subjects at 50% risk of NF2.

Patient details from hospital notes and proformas have been entered into a database at St Mary's Hospital Manchester since 1989. A total of 344 patients fulfilling our published criteria for NF2 are recorded (table 1).1 3 5 Information on all known tumours, age at presentation and diagnosis, and many different symptoms are included. We have analysed details of patients who have been detected as having VS on a screening MRI scan without any known central nervous system pathology, signs, or symptoms. It has been our practice to carry out a baseline MRI scan between 12 and 16 years of age, although some centres have screened earlier. A review was also made of subjects diagnosed on the basis of presymptomatic DNA tests. A separate analysis was undertaken on those cases who had sequential audiological screening with pure tone audiogram and ABR before MRI scanning. Standard MRI screening usually involved T1 and T2 weighted images before, and T1 weighted images after gadolinium enhancement with 3 mm cuts through the internal auditory meati.

Table 1

Diagnostic criteria for NF2

A cumulative age at onset curve was derived from the age at presymptomatic diagnosis on MRI scan (fig 1) and for residual risk of NF2 (fig 2). An adjustment was made for large tumours detected on initial prevalence scan. This allowed for an annual growth rate of 2 mm (derived from our unpublished data and those of M Baser/V Mautner) with a minimum detection size of 2-3 mm. Thus, a 20 mm tumour could have been detected up to an average of eight or nine years earlier. Age of onset curves were also derived from offspring in NF2 families to determine the proportion who had developed symptoms by each five year age group. Residual risks of NF2 were based on Bayesian calculations,3 so when 67% of subjects would be expected to be symptomatic or detected on MRI the residual risk of NF2 would be 25%.

Figure 1

Cumulative chance of detecting NF2 using MRI based on prospective screening in 35 at risk subjects.

Figure 2

Residual risk of NF2 if MRI negative based on screening 35 at risk subjects.

MRI screening is useful for reducing the residual risk of someone at initial 50% risk of inheriting NF2. ABR screening still has a place in early detection of vestibular schwannoma, although MRI screening is the method of choice.

Forty three first degree relatives of NF2 patients have been diagnosed with NF2 presymptomatically. Twenty one patients were diagnosed initially with a presymptomatic DNA test using linkage or mutation analysis. Fourteen of these have subsequently been found to have VS on MRI scan. The remaining seven children have not had an initial MRI scan. Twenty one further patients were diagnosed on the basis of an MRI scan before having symptoms. These scans were carried out on initial assessment when the patient was first known to be at risk. Thus, some scans were only performed when the patients were in their twenties to forties and the tumour size at initial diagnosis therefore varied from 1-20 mm in diameter. Mean age at initial positive scan was 20.5 years (range 8-45 years, median 18 years). All but two patients were the offspring of a known NF2 patient. One father of a severely affected NF2 patient who claimed not to have symptoms had bilateral 15 and 10 mm VS on initial MRI scan aged 45 years. A further 53 year old male with an affected brother has had a normal MRI scan, but his son aged 19 years had a unilateral VS on scan aged 19 years. The family have a deletion in the NF2 gene removing one exonic cosmid (J Dumanski, personal communication). Only three patients had an initial normal MRI scan. The 45 year old father noted above was excluded from analysis as he may represent a mosaic and was not at a priori 50% risk. Of the 34 patients detected presymptomatically, DNA presymptomatic testing would be possible in 24 owing to the presence of a known mutation or a pedigree structure suitable for linkage analysis at time of diagnosis. However, in 10 cases (30%) no testing would have been possible if the patient presented at the time of writing. Age of onset of symptoms in 110 offspring of NF2 patients is presented in fig 3, while the residual risk for NF2 in someone asymptomatic at 50% initial risk is presented in fig 4.

Figure 3

Age at onset of symptoms in 110 offspring of NF2 patients.

Figure 4

Residual risk of NF2 based on age at onset of symptoms in 110 offspring of NF2 patients.

Fourteen patients had records of both pure tone audiograms and ABR testing within three months of the initial positive MRI scan (11 from Manchester). Twelve pure tone audiograms were within normal limits with no threshold above 20 dB. Two audiograms showed a unilateral high frequency hearing loss. Five patients with bilateral intracanalicular VS (1-8 mm diameter) and a further patient with a l mm unilateral VS showed no abnormality on audiogram or ABR testing. ABR tests carried out one year later on four bilateral cases showed delayed latencies on both sides in 3/4 while intracanalicular diameter remained <9 mm. In three further patients with bilateral VS on scan, the smaller tumour did not cause an abnormality to the ABR, although intracanalicular diameter did not exceed 5 mm. In one of these tumours, six years after initial scan the tumour still showed no delay in latencies or wave morphology although the tumour was impinging on the CP angle. In all the remaining five patients there was abnormal wave morphology or delay of latencies bilaterally consistent with MRI findings. Of the 27 tumours from 14 patients, 19 have shown abnormal ABR (six one year after MRI detection of VS) with a range of 2 mm intracanalicular diameter to 15 mm intracranial diameter. Median diameter at initial ABR abnormality was 6 mm, with only one VS extending into the cerebellopontine angle not being detected.

The UK NF2 database is to our knowledge the largest clinical database of NF2 in the world. While ascertainment for the whole UK is not complete, the living diagnostic prevalence of 1 in 270 000 comes close to the 1 in 200 000 we previously calculated from our own population in the north west.4 Therefore, this patient series approximates to a population based series of patients presenting with NF2. We have detected VS on MRI scans in 35 patients without NF2 symptoms or signs. The age curves derived from this group of patients (figs 1 and 2) will be useful in counselling subjects at 50% risk of the disease, particularly if no DNA test is available in their family. As mutation detection can take up to six months, some degree of risk reduction will already be possible before a test becomes available in a family. These curves will, if anything, underestimate the degree of risk reduction. This is because scans are not undertaken until 10 years of age and, as can be seen from fig 3, 10% of NF2 offspring present symptomatically before 10 years of age. Our corrected curves assume an arbitrary rate of growth of 2 mm per year. Many of the families in which presymptomatic diagnosis was made are at the extreme mild end of the scale with missense or splice site mutations and large deletions.6-9 All the patients in table 2 in which a mutation had been found had these types of mutation. Nonetheless, faced with a patient without obvious features of NF2 in which the family history of NF2 is mild, the curves are very useful. Most patients with severe NF2 do not have children3 and therefore most offspring are at risk of the milder forms of the disease. Indeed, many children of sporadic patients with NF2 may be at less than 50% initial risk owing to gonosomal mosaicism in their parent.3

Table 2

Audiological features in 14 NF2 patients presenting with presymptomatic VS on MRI scan

The value of early detection of VS is still open to some debate. However, the only realistic chance of hearing preservation surgery is when the tumour is small10 11 and tumours may not present symptomatically until 20 mm in diameter. ABR testing, while not detecting some of the very small intracanalicular tumours, still detected the great majority of tumours at a size when hearing preservation surgery would not have been compromised. However, the larger size at detection would leave less leeway for observation of the growth rate of the tumour. Where MRI scans are less accessible, ABR testing has more of a role than CT scanning (see case 11) and could serve as a back up to MRI in the interval years after normal scans.

It is clear that NF2 patients fare better if they are managed in specialist centres where surgery and occasionally other measures such as radiosurgery can be offered at the appropriate time.12In experienced hands, hearing can be preserved with removal of VS and rehabilitation is now possible with auditory brain stem implants and on occasion with cochlear implants.10 11

There is still clearly a place for MRI and audiological screening in at risk relatives as 30% of the patients in this study would not have been able to have predictive DNA tests. While the proportion of families with mutations can be boosted by a comprehensive deletion finding strategy,13 these are not widely available for NF2 or many other diseases (BRCA1/2) and still fail to detect mosaic mutations. A child who presents with NF2 with no family history is likely to represent a new mutation of theNF2 gene. If there is no suggestive family history, the risk to parent and sibs is still low (probably <1%). The cases presented here of two parents being affected after a child is to our knowledge the first such report. There is also no objective evidence that VS disease is worse in females than males.14All four of the female sibs in table 2 had smaller VS at initial scan than their male counterparts.

Children of affected patients should be considered to be at 50% risk of NF2 and screening for NF2 can start at birth.5Cataracts can affect vision in early life and other tumour implications are present in the first 10 years of life. Formal screening for VS should start at 10 years, as it is rare for tumours to occur before that time. Audiological tests including ABR are still a useful adjunct to MRI and can occur annually from 10 years of age. As surgery would only be contemplated for tumours at about 6 mm and tumour growth averages 2 mm per year, three yearly screening with MRI in a subject with no tumours is probably sufficient. The initial MRI scan could be at around 12 years of age. Once tumours are present screening should probably be annual. Spinal tumours are found very frequently on MRI, as many as 90% having evidence of these in some studies.15However, only 25-30% require a spinal operation for a symptomatic tumour.1 Nonetheless, a full annual neurological examination is probably a wise precaution. In most families it is now possible to develop a genetic test so screening can be targeted on affected subjects only. Uptake of such testing in childhood and adult life is high.16 Using the data from this paper even without DNA tests the residual chance of having NF2 in a subject at 50% risk can be reduced to around 30%, 10%, and 5% if MRI is negative at 15, 25, and 35 years respectively.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.