Epidemiology

The search of the epidemiology literature was performed to investigate: the incidence of vestibular schwannoma in the general and elderly populations, the size of tumours at diagnosis and how this has changed with time, and the relationship between age and tumour size.

Our search revealed 1022 studies. By following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart, 11 papers were included that met our criteria.^{Reference Gal, Shinn and Huang4–Reference Foley, Shirazi, Maweni, Walsh, McConn Walsh and Javadpour14} The results are shown in Figure 1, and Tables 1 and 2.

Fig. 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (‘PRISMA’) 2009 flow diagram for epidemiology of vestibular schwannoma.

Table 1.

Epidemiology of vestibular schwannoma – incidence rate and percentage of older patients

SEER = Surveillance, Epidemiology, and End Result Program

Table 2.

Epidemiology of vestibular schwannoma – tumour size and relationship with age

*Number obtained from 2010 study. ^†Study included 0.26 per cent neurofibromatosis type 2 patients. VS = vestibular schwannoma

The main reasons for excluding 1011 papers were as follows: the articles did not perform an epidemiological study themselves, but quoted figures from other sources; the main topics of discussion were vestibular schwannoma treatment, risk factors or change in management; the articles compared incidence of vestibular schwannoma between males and females, between people from different races or of different ethnicities, between different socio-economic classes; or the articles reported a regional comparison of vestibular schwannoma incidence within a country. It was not possible to determine the total number of cases across the entire population studied from the data in these papers.

The results for incidence rates of vestibular schwannoma in the general population came from studies conducted in the USA, Denmark and the Netherlands. We found seven papers in this section.^{Reference Gal, Shinn and Huang4–Reference Babu, Sharma, Bagley, Hatef, Friedman and Adamson10} The incidence rates varied from 1.1^{Reference Gal, Shinn and Huang4} to 3.07^{Reference Stepanidis, Kessel, Caye-Thomasen and Stangerup8} per 100 000 population per year. A long-term study carried out by Stangerup and Caye-Thomasen^{Reference Stangerup and Caye-Thomasen5} showed that, after a steady increase, vestibular schwannoma incidence started to decline. The figure quoted by the British Acoustic Neuroma Association is 2 per 100 000 per year.¹⁵ An earlier study, by Tos et al.,^{Reference Tos, Stangerup, Cayé-Thomasen, Tos and Thomsen16} showed an increase in vestibular schwannoma incidence between 1976 and 2001.

Out of these seven papers, five^{Reference Gal, Shinn and Huang4,Reference Stangerup and Caye-Thomasen5,Reference Kleijwegt, Ho, Visser, Godefroy and van der Mey7,Reference Stepanidis, Kessel, Caye-Thomasen and Stangerup8,Reference Babu, Sharma, Bagley, Hatef, Friedman and Adamson10} showed the percentage of patients aged over 65–70 years when they were first diagnosed with vestibular schwannoma. Two papers^{Reference Gal, Shinn and Huang4,Reference Stepanidis, Kessel, Caye-Thomasen and Stangerup8} published the percentage of patients aged over 65 years, giving figures of 21 per cent and 25–30 per cent respectively. Three papers^{Reference Lau, Olivera, Miller, Downes, Danner and van Loveren6,Reference Kleijwegt, Ho, Visser, Godefroy and van der Mey7,Reference Babu, Sharma, Bagley, Hatef, Friedman and Adamson10} showed the percentage of patients aged over 70 years; the respective results were 12 per cent, 15.9 per cent and 14.01 per cent. It appears that most of the patients were younger than 65 years when they were diagnosed with vestibular schwannoma.

A long-term epidemiological study conducted in Denmark^{Reference Stangerup and Caye-Thomasen5} showed that the mean age of vestibular schwannoma diagnosis increased from 49 in 1976 to 58 in 2008, but the number of patients aged under 40 years did not change significantly. This indicates that more vestibular schwannoma cases are diagnosed at an older age, but the percentage of those aged over 70 years is still low compared to the total diagnosed.

An assessment was made of the average vestibular schwannoma size at diagnosis, how this has changed with time, and the relationship between patient age and tumour size at diagnosis. Results were obtained from 10 studies and are shown in Table 2.^{Reference Gal, Shinn and Huang4,Reference Stangerup and Caye-Thomasen5,Reference Kleijwegt, Ho, Visser, Godefroy and van der Mey7–Reference Foley, Shirazi, Maweni, Walsh, McConn Walsh and Javadpour14} The focus was mainly directed on the change in tumour size with time and the age of patients.

It is clear that the tumour size at diagnosis decreased over time. This suggests that an increasing number of people were diagnosed with smaller tumours. This is mainly attributed to improved technology used in investigations for vestibular schwannoma. It was also found that, generally, older people were more likely to have smaller tumours.^{Reference Gal, Shinn and Huang4,Reference Stangerup and Caye-Thomasen5,Reference Babu, Sharma, Bagley, Hatef, Friedman and Adamson10,Reference Stangerup, Tos, Thomsen and Caye-Thomasen12,Reference Harun, Agrawal, Tan, Niparko and Francis13}

A long-term study by Stangerup and Caye-Thomasen^{Reference Stangerup and Caye-Thomasen5} found that, by 2008, the average size of tumours in people aged 70 years or more was 13 mm. A study from the same centre, published in 2010,^{Reference Stangerup, Tos, Thomsen and Caye-Thomasen12} found that the number of intra-meatal tumours and extra-meatal tumours sized 1–10 mm was highest in people aged over 70 years. A study by Harun et al.^{Reference Harun, Agrawal, Tan, Niparko and Francis13} showed that, for each additional year of age at diagnosis, the mean tumour size decreased by 0.244 mm. Papers from Kleijwegt et al.^{Reference Kleijwegt, Ho, Visser, Godefroy and van der Mey7} and Babu et al.^{Reference Babu, Sharma, Bagley, Hatef, Friedman and Adamson10} reported that, irrespective of age, the proportion of smaller tumours was higher than that of larger ones.

In summary, we can see from the published papers that the incidence of vestibular schwannoma has increased with time, and the size of the tumour at diagnosis has decreased with time. The percentage of people first diagnosed with vestibular schwannoma when aged over 70 years was reported as 12–15.9 per cent. In addition, elderly people had smaller sized tumours.

Change in vestibular schwannoma management

As the average patient age and vestibular schwannoma size at initial diagnosis changed, so did the management of this neoplasm. Factors considered when making decisions about treatment included tumour size, patient's age, health and hearing status, and the patient's preference.^{Reference Carlson, Habermann, Wagie, Driscoll, Van Gompel and Jacob9} However, the same authors also mentioned that sometimes the mode of treatment was influenced by the opinion and preference of the doctor(s) who saw these patients.

Generally, younger patients and larger tumours were treated with surgery, while older patients and those with smaller tumours received conservative management.^{Reference Babu, Sharma, Bagley, Hatef, Friedman and Adamson10} Conservative management involved a ‘wait and scan’ method, with serial MRI, to observe any tumour growth, and to check the symptoms and signs, in order to determine whether any change in treatment was required. There is also an argument in favour of active treatment for hearing preservation,^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17} but this has been strongly challenged by Møller et al.^{Reference Møller, Hansen, Miyazaki, Stangerup and Caye-Thomasen18}

Overall, our literature search revealed a decline in the rate of surgery and an increase in conservative management with time, which is shown in Table 3.^{Reference Lau, Olivera, Miller, Downes, Danner and van Loveren6,Reference Kleijwegt, Ho, Visser, Godefroy and van der Mey7,Reference Carlson, Habermann, Wagie, Driscoll, Van Gompel and Jacob9–Reference Patel, Vasan, van Loveren, Downes and Agazzi11,Reference Foley, Shirazi, Maweni, Walsh, McConn Walsh and Javadpour14,Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17,Reference Moffat, Kasbekar, Axon and Lloyd19–Reference Mackeith, Kerr and Milford21} Carlson et al.^{Reference Carlson, Habermann, Wagie, Driscoll, Van Gompel and Jacob9} predicted that, by 2017, conservative management would become the most common method of treatment, and, by 2026, at least 50 per cent of tumours would be at least initially managed using the wait and scan method. The same paper proved (with a statistically significant correlation) that both surgery and radiotherapy were more commonly used for younger patients and larger tumours.

Table 3.

Change in management of vestibular schwannoma with time

VS = vestibular schwannoma

Tumour growth and conservative management outcome

With the changes in vestibular schwannoma epidemiology (an increase in the number of smaller tumours diagnosed; Table 2) and management (more cases of conservative management than active treatment; Table 3), the rate of vestibular schwannoma growth and the factors contributing to growth are two major topics of interest. These two topics can predict which of those tumours that are small at diagnosis will need active treatment during the observation period. This has led to many articles being published on the subject of tumour growth.

Our search in this section identified 1336 records. Twenty-two papers^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17,Reference Moffat, Kasbekar, Axon and Lloyd19,Reference Suryanarayanan, Ramsden, Saeed, Aggarwal, King and Rutherford20,Reference Younes, Montava, Bachelard-Serra, Jaloux, Salburgo and Lavieille22–Reference Klersy, Arlt, Hofer and Meixensberger40} were ultimately included by following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart (Figure 2); these are summarised in Table 4. In addition to these, a systematic review on tumour growth – published during the search period – was identified, along with two other systematic reviews published before 2008; these are summarised separately^{Reference Paldor, Chen and Kaye42,Reference Smouha, Yoo, Mohr and Davis43,Reference Yoshimoto45} (Table 5).

Table 4.

Tumour growth and conservative management outcome

*Only picked up patients who completed conservative management between 2007 and 2013. ^†Two patients lost to follow up, one died, 47.6 per cent of tumours were intra-canalicular. ^‡Excluding those who showed tumour growth but did not have any treatment, and one who died from cerebral oedema. **Number of patients according to tumour size: intra-canalicular = 34; 1–10 mm = 14; 11–20 mm = 10; 21–30 mm = 1. ^§Mean volume doubling time was 4.4 years; any tumour showing volume doubling time of more than 0 years was considered in the growing group. ^#Article involved neurofibromatosis type 2, but data for sporadic tumours were given separately. CT = computed tomography; MRI = magnetic resonance imaging; RT = radiotherapy

Table 5.

Systematic reviews on vestibular schwannoma growth

Fig. 2.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (‘PRISMA’) 2009 flow diagram for growth and conservative management outcome of vestibular schwannoma.

Our review of the included papers revealed a lack of uniformity in measuring tumour growth. While most papers considered an increase in extra-canalicular diameter as growth,^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17,Reference Moffat, Kasbekar, Axon and Lloyd19,Reference Suryanarayanan, Ramsden, Saeed, Aggarwal, King and Rutherford20,Reference Younes, Montava, Bachelard-Serra, Jaloux, Salburgo and Lavieille22–Reference Hughes, Skilbeck, Saeed and Bradford29,Reference Whitehouse, Foroughi, Shone and Hatfield31,Reference Régis, Carron, Park, Soumare, Delsanti and Thomassin32,Reference Bakkouri, Kania, Guichard, Lot, Herman and Huy35,Reference Klersy, Arlt, Hofer and Meixensberger40} some studies used an increase in tumour volume or volume doubling time.^{Reference Varughese, Breivik, Wentzel-Larsen and Lund-Johansen34} Even when diameter was considered, some papers defined growth as 2 mm or more per year,^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17,Reference Moffat, Kasbekar, Axon and Lloyd19,Reference Younes, Montava, Bachelard-Serra, Jaloux, Salburgo and Lavieille22,Reference Martin, Senthil, Chavda, Walsh and Irving23,Reference Fayad, Semaan, Lin, Berliner and Brackmann25,Reference Hughes, Skilbeck, Saeed and Bradford29,Reference Kirchmann, Karnov, Hansen, Dethloff, Stangerup and Caye-Thomasen33,Reference Ferri, Pirodda, Ceroni, Fioravanti, Calbucci and Modugno36} while others defined it as 1 mm or more.^{Reference Suryanarayanan, Ramsden, Saeed, Aggarwal, King and Rutherford20,Reference Hajioff, Raut, Walsh, Bath, Bance and Guha27,Reference Whitehouse, Foroughi, Shone and Hatfield31,Reference Agrawal, Clark, Limb, Niparko and Francis37} The time interval used to measure growth was also variable. While most considered growth per annum,^{Reference Moffat, Kasbekar, Axon and Lloyd19,Reference Suryanarayanan, Ramsden, Saeed, Aggarwal, King and Rutherford20,Reference Younes, Montava, Bachelard-Serra, Jaloux, Salburgo and Lavieille22,Reference Martin, Senthil, Chavda, Walsh and Irving23,Reference Fayad, Semaan, Lin, Berliner and Brackmann25,Reference Hajioff, Raut, Walsh, Bath, Bance and Guha27,Reference Hughes, Skilbeck, Saeed and Bradford29,Reference Whitehouse, Foroughi, Shone and Hatfield31,Reference Kirchmann, Karnov, Hansen, Dethloff, Stangerup and Caye-Thomasen33,Reference Ferri, Pirodda, Ceroni, Fioravanti, Calbucci and Modugno36,Reference Agrawal, Clark, Limb, Niparko and Francis37} others considered the growth between two successive scans,^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17,Reference Reddy, Lewis-Jones, Javadpour, Ryland and Lesser26,Reference Régis, Carron, Park, Soumare, Delsanti and Thomassin32,Reference Eljamel, Hussain and Eljamel38,Reference Daultrey, Rainsbury and Irving39} although imaging is not always conducted at one-year intervals.

The average tumour size at diagnosis varied from 5.1 mm^{Reference Suryanarayanan, Ramsden, Saeed, Aggarwal, King and Rutherford20} to 20.1 mm.^{Reference Reddy, Lewis-Jones, Javadpour, Ryland and Lesser26} No average value was reported in five articles;^{Reference Younes, Montava, Bachelard-Serra, Jaloux, Salburgo and Lavieille22,Reference Hughes, Skilbeck, Saeed and Bradford29,Reference Bakkouri, Kania, Guichard, Lot, Herman and Huy35,Reference Ferri, Pirodda, Ceroni, Fioravanti, Calbucci and Modugno36,Reference Eljamel, Hussain and Eljamel38} two articles^{Reference Lee, Park, Kim and Cho30,Reference Kirchmann, Karnov, Hansen, Dethloff, Stangerup and Caye-Thomasen33} included intra-canalicular tumours only, two articles^{Reference Martin, Senthil, Chavda, Walsh and Irving23,Reference Daultrey, Rainsbury and Irving39} included intra-canalicular tumours and up to 2 cm of extra-canalicular tumours, while one article^{Reference Varughese, Breivik, Wentzel-Larsen and Lund-Johansen34} considered tumour volume. We calculated the weighted average diameter of tumours at diagnosis as 9.32 mm. This is based on data gathered from 12 papers.^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17,Reference Moffat, Kasbekar, Axon and Lloyd19,Reference Suryanarayanan, Ramsden, Saeed, Aggarwal, King and Rutherford20,Reference Escorihuela-García, Llópez-Carratalá, Orts-Alborch and Marco-Algarra24–Reference Oddon, Montava, Salburgo, Collin, Vercasson and Lavieille28,Reference Whitehouse, Foroughi, Shone and Hatfield31,Reference Régis, Carron, Park, Soumare, Delsanti and Thomassin32,Reference Agrawal, Clark, Limb, Niparko and Francis37,Reference Klersy, Arlt, Hofer and Meixensberger40} We must remember that these tumours were treated using a conservative approach.

Great disparity was apparent in the percentage of tumours showing growth during the observation period, ranging from 11.9 per cent^{Reference Daultrey, Rainsbury and Irving39} to 79.8 per cent.^{Reference Varughese, Breivik, Wentzel-Larsen and Lund-Johansen34} This raises a serious question regarding how vestibular schwannoma growth is measured. The weighted mean percentage of tumours showing growth was approximately 33.79 per cent. This was calculated using data obtained from 22 papers.^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17,Reference Moffat, Kasbekar, Axon and Lloyd19,Reference Suryanarayanan, Ramsden, Saeed, Aggarwal, King and Rutherford20,Reference Younes, Montava, Bachelard-Serra, Jaloux, Salburgo and Lavieille22–Reference Klersy, Arlt, Hofer and Meixensberger40}

The growth rate was also variable, ranging from 0.3 mm per year^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17} to 4 mm per year.^{Reference Martin, Senthil, Chavda, Walsh and Irving23} This figure was reported as low as 0.16 mm per year for intra-canalicular tumours.^{Reference Hughes, Skilbeck, Saeed and Bradford29} Data on average growth rates were missing from six articles.^{Reference Lee, Park, Kim and Cho30,Reference Kirchmann, Karnov, Hansen, Dethloff, Stangerup and Caye-Thomasen33,Reference Ferri, Pirodda, Ceroni, Fioravanti, Calbucci and Modugno36–Reference Daultrey, Rainsbury and Irving39}

Lastly, the rate of conservative management success (i.e. the patient did not have to switch treatment to surgery or radiosurgery during the study period) ranged from 26 per cent^{Reference Régis, Carron, Park, Soumare, Delsanti and Thomassin32} to 100 per cent^{Reference Escorihuela-García, Llópez-Carratalá, Orts-Alborch and Marco-Algarra24} (follow-up period not available) or 96.9 per cent^{Reference Klersy, Arlt, Hofer and Meixensberger40} (known follow-up period). These data were not available in two articles.^{Reference Oddon, Montava, Salburgo, Collin, Vercasson and Lavieille28,Reference Daultrey, Rainsbury and Irving39} The weighted mean for this (from 20 articles^{Reference Ferri, Modugno, Pirodda, Fioravanti, Calbucci and Ceroni17,Reference Moffat, Kasbekar, Axon and Lloyd19,Reference Suryanarayanan, Ramsden, Saeed, Aggarwal, King and Rutherford20,Reference Younes, Montava, Bachelard-Serra, Jaloux, Salburgo and Lavieille22–Reference Hajioff, Raut, Walsh, Bath, Bance and Guha27,Reference Hughes, Skilbeck, Saeed and Bradford29–Reference Eljamel, Hussain and Eljamel38,Reference Klersy, Arlt, Hofer and Meixensberger40} ) was 73.66 per cent.

Next, we investigated published studies that had carried out systematic reviews on the growth and success of conservative management. Two papers^{Reference Nikolopoulos, Fortnum, O'Donoghue and Baguley41,Reference Paldor, Chen and Kaye42} were published between 2018 and 2008, and three^{Reference Smouha, Yoo, Mohr and Davis43–Reference Yoshimoto45} were published before that time. Of these five systematic review papers, three^{Reference Paldor, Chen and Kaye42,Reference Smouha, Yoo, Mohr and Davis43,Reference Yoshimoto45} used a weighted average (the number of patients in the included articles was used as weight), one^{Reference Nikolopoulos, Fortnum, O'Donoghue and Baguley41} did not give any average value and another^{Reference Yamakami, Uchino, Kobayashi and Yamaura44} used mean value. In order to avoid any confusion in comparing data with our article, only values from those articles that used weighted averages^{Reference Paldor, Chen and Kaye42,Reference Smouha, Yoo, Mohr and Davis43,Reference Yoshimoto45} are given, in Table 5. These three articles reported the percentage of tumours showing growth as between 43 per cent and 50 per cent. The growth rate was 1.11–1.9 mm per year, and the conservative management success rates were 80 per cent^{Reference Smouha, Yoo, Mohr and Davis43} and 82 per cent.^{Reference Yoshimoto45}

Asymmetrical sensorineural hearing loss and vestibular schwannoma

Here, we tried to determine the chance of detecting vestibular schwannoma from MRI scans that were requested on the basis of asymmetrical SNHL and unilateral tinnitus. A literature search using the key words mentioned earlier revealed 115 studies. By following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart (Figure 3), we included six articles^{Reference Kleijwegt, Ho, Visser, Godefroy and van der Mey7,Reference Wilson, Gandolfi, Ahn, Yu, Huang and Kim46–Reference Suzuki, Hashimoto, Kano and Okitsu50} for our qualitative synthesis. All of these were based on asymmetrical SNHL. One paper was found to be a systematic analysis.^{Reference Egan51} Three out of six articles^{Reference Wilson, Gandolfi, Ahn, Yu, Huang and Kim46,Reference Newton, Shakeel, Flatman, Beattie and Ram49,Reference Suzuki, Hashimoto, Kano and Okitsu50} in the study by Egan^{Reference Egan51} were included, and that is why the data from this article are not included in weighted mean calculation. The results of these articles are shown in Table 6.

Fig. 3.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (‘PRISMA’) flowchart for sensitivity of pure tone audiogram, showing asymmetrical sensorineural hearing loss, for vestibular schwannoma.

Table 6.

Vestibular schwannoma detection rate from MRI scans requested on basis of asymmetrical SNHL

*Scans per diagnosis reflect the number of scans required to obtain one positive result, and yield represents the percentage value of positive results divided by the total number of scans performed. ^†Paper states number of scans, not patients (presumed one scan per patient). MRI = magnetic resonance imaging; SNHL = sensorineural hearing loss

It must be mentioned that there was a lack of consensus on how asymmetrical SNHL should be defined. A guideline from the USA^{Reference Sweeney, Carlson, Shepard, McCracken, Vivas and Neff52} recommended a 10 dB or more difference at two or more contiguous frequencies, or a 15 dB or more difference at one frequency. The article also states that a 15 dB or more difference at 3 kHz was the most reliable guideline. Nine different types of definition for asymmetrical SNHL were found in the thesis by Egan.^{Reference Egan51}

As mentioned earlier, 6 studies^{Reference Kleijwegt, Ho, Visser, Godefroy and van der Mey7,Reference Wilson, Gandolfi, Ahn, Yu, Huang and Kim46–Reference Suzuki, Hashimoto, Kano and Okitsu50} included in our qualitative analysis showed that the number of scans per diagnosis (the number of MRI scans per 1 diagnosis of vestibular schwannoma) ranged from 247 scans^{Reference Wilson, Gandolfi, Ahn, Yu, Huang and Kim46} to 23.3 scans.^{Reference Pan, Huang, Morioka, Hathout and El-Saden47} The diagnostic yield (the percentage of positive scans against the total number of scans performed) ranged from 4.2 per cent^{Reference Pan, Huang, Morioka, Hathout and El-Saden47} to 0.4 per cent.^{Reference Wilson, Gandolfi, Ahn, Yu, Huang and Kim46} The systematic review by Egan showed the average number of scans per diagnosis as 19.53 and the diagnostic yield as 5.1 per cent.^{Reference Egan51}

When calculating the weighted mean using the data from these 6 articles, there were 84.15 scans per diagnosis, with a diagnostic yield of 1.18 per cent. When the total number of patients (n = 4791) and positive scans (n = 149) in these six studies was added, and assuming each MRI scan was requested for one asymmetrical SNHL result in a pure tone audiogram, the positive predictive value for asymmetrical SNHL in the diagnosis of vestibular schwannoma was 3.1 per cent.

In addition to the above six papers, our literature search also identified two abstracts for which we could not access the full papers; the yields quoted in those were 1.3 per cent^{Reference Amiraraghi, Lim and Kontorinis53} and 2.4 per cent.^{Reference Aaron, Gildener-Leapman, Trivedi, Reinstadler, Zimmerman and Casazza54}

‘Yield’ or ‘pick-up rate’ were the terms used by different authors^{Reference Wilson, Gandolfi, Ahn, Yu, Huang and Kim46,Reference Pan, Huang, Morioka, Hathout and El-Saden47,Reference Egan51,Reference Amiraraghi, Lim and Kontorinis53} to describe the effectiveness of MRI, requested for asymmetrical SNHL cases, in diagnosing vestibular schwannoma; we selected the former term for use in our study. Some papers have used the terms ‘sensitivity’ and ‘specificity’ in this context, but those papers were mainly focused either on comparing MRI with other non-imaging diagnostic tools for vestibular schwannoma or comparing the best predictive value of different audiological protocols for asymmetrical SNHL.^{Reference Hentschel, Scholte, Steen, Kunst and Rovers1,Reference Pena, Chew, Landau, Breen, Zevallos and Vrabec55,Reference Rajeev, Babur, Karakitsou and Mehta56}

The results of our literature search for vestibular schwannoma and unilateral tinnitus were sparse. The incidence rate quoted was very low in most of the published articles. One systematic review^{Reference Sweeney, Carlson, Shepard, McCracken, Vivas and Neff52} gave this as less than 1 per cent for patients whose presenting symptom was unilateral tinnitus. Another article^{Reference Hoekstra, Prijs and van Zanten57} found that the incidence of vestibular schwannoma in patients with tinnitus was 0.3 per cent. Choi et al.^{Reference Choi, Sajisevi, Kahmke and Kaylie48} found no vestibular schwannoma in patients with unilateral tinnitus and with no hearing loss or symmetric hearing loss. One article^{Reference Reddy, Lewis-Jones, Javadpour, Ryland and Lesser26} with a small number of patients reported detecting vestibular schwannoma in 4.4 per cent of patients with tinnitus as the only presenting symptom.