Cerebellopontine angle (CPA) tumors are the most common neoplasms in the posterior fossa, accounting for 5–10% of all primary intracranial tumors. Most posterior fossa tumors are benign meningiomas, which account for 13–25% of all primary intracranial tumors. [1] However, when looking specifically at the CPA, roughly 80% are vestibular schwannoma. [2] Meningiomas and epidermoids account for 10% and 6%, respectively with lipomas, vascular malformations, hemangiomas and lower cranial nerve schwannomas accounting for the remainder. [2] [3]

A vestibular schwannoma classically presents with unilateral or asymmetrical sensorineural hearing loss (SNHL), often in the high frequencies and accompanied by poor speech discrimination. [4] Speech audiometry however, is not routinely performed in audiology departments in the UK. Diagnostic imaging is, therefore, offered to these patients to rule out vestibular schwannomas as a cause of their asymmetrical hearing loss. Asymmetrical tinnitus, asymmetrical bilateral hearing loss, true vertigo, disequilibrium and trigeminal nerve distribution paresthesia are also indications for diagnostic imaging. [5]

Magnetic resonance imaging of the internal acoustic meatus (MRI IAM) scan with gadolinium enhancement is the gold standard to investigate lesions in the internal acoustic meatus and CPA. [6] Whilst it is possible to identify these lesions on computed tomography (CT) scans, in over 50% of patients [7] this modality often only raises suspicion by demonstrating an area of smooth bony erosion. Furthermore, establishing exact anatomical location and distinguishing between normal brain parenchyma and tumor tissue is more accurate on MRI scan than CT scan. [8]

During the imaging process the cerebral hemispheres, cerebellum, brain stem and more superficial structures are all analyzed. It is, therefore, inevitable that either no abnormality will be detected on imaging or lesions not attributable to the symptoms being investigated may be identified. One systematic review by Morris et al. looking at incidental findings on MRI scan of healthy individuals showed a 0.7% prevalence of a neoplastic incidental brain finding and 2.0% a non-neoplastic one (excluding microvascular disease). Specifically, 0.03% of patients without symptoms were found to have vestibular schwannomas. [9] Wong et al. looked into findings on MRI scans performed in one district general hospital and described that incidental findings were present on 46.7% of patients. [10] In our region, patients with a CPA tumor are referred to a tertiary centre for joint neurosurgical and otolaryngology care. Given this, it seems pragmatic to determine if the detection rate of tumors and incidental findings are different depending on where investigations are performed. Ultimately, after a normal or abnormal scan, the question then arises of how to proceed; one study suggests with appropriate prior counseling, over 50% of patients requiring MRI IAM do not require ENT follow-up. [11]

The aim of this paper is to determine the detection rate of incidental findings on MRI IAM at both a district general hospital (DGH) and a tertiary hospital (TH) and to establish the pickup rate of vestibular schwannomas in these two types of hospital.

Data were collected retrospectively from two trusts; one tertiary centre and one district general hospital. A list of all MRI IAM scans performed for audiovestibular symptoms was acquired from the hospitals' imaging viewing and reporting systems for a two-year period (January 1, 2008-December 31, 2009). All reports were reviewed and indications for scanning and findings recorded. Any missing indications were extracted from clinic letters. Only the first scans for new symptoms were included; all monitoring scans of previously diagnosed tumors were excluded. Statistical analysis was performed using Graph Pad Prism 6 (Graph Pad Software, San Diego). The study gained approval from the local audit and governance departments in each hospital.

Definitions
Non-vestibular schwannoma incidental findings were defined as diagnoses that did not explain the patients' current symptoms, nor were documented in their medical history. Incidental findings were grouped into several cohorts comprising ischemia, sinus disease and mastoid disease. Patients labeled with ischemia had either small vessel arterial disease or evidence of previous cerebrovascular accident. Those with sinus disease either had evidence of sinusitis (thickened mucosa) or the presence of a mucocele. Mastoid disease was applicable in the presence of bony destruction of the mastoid air cells or the presence of fluid. Findings that required intervention or referral to another specialty were considered clinically significant. [10]

District General Hospital: The number of scans taken place in the DGH during the study period was 657 in which 339 patients were female and 318 were male. The mean age of patients scanned was 59.8 years (range 11–95 years). No abnormality was detected following 476 (72.5%) scans. Incidental findings were found in 170 scans (25.8%). The most frequent incidental findings were evidence of ischemia in 51 scans, sinus disease in 28 scans and mastoid disease in 19 scans. Vascular loops, meningioma and arachnoid cyst were also seen. (Figure 1) Vestibular schwannoma was diagnosed on 11 scans (1.7%). (Table 1) All scans were authorized by a consultant radiologist.

Tertiary Hospital: The number of scans taken place in the TH during the study period was 1517 in which 851 patients were female and 666 were male. The mean age of patients was 53.4 years (range 16.1–95.1 years). No abnormality was detected following 911 (60%) scans. Incidental findings were found in 585 scans (38.6%). The most frequent incidental findings were ischemia in 292 scans, sinus disease in 85 scans and mastoid disease in 85 scans. Vascular loops, meningioma, arachnoid cyst and other space occupying lesions were also seen. (Figure 1) Vestibular schwannoma was diagnosed on 21 scans (1.4%). All scans were authorized by a consultant neuroradiologist.

Comparison
When the number of patients at the DGH and TH were compared, there was no significant difference in gender (chi-square test p = 0.06), but there was a significant difference in age (median 61.0 years at the DGH vs 54.4 years at the TH, unpaired T-test p<000.1. (Figure 2) Age was demonstrated to be normally distributed. MRI IAM scans at the TH showed significantly more incidental findings when compared to findings at the DGH (38.6% vs 25.8%, chi-square test p < 0.0001. (Figure 3)

Incidental findings were associated with a higher average age (52.5 vs 60.3 years, Mann-Whitney U test p < 0.0001, (Figure 4) but not with gender (chi-square test p = 0.924). Comparison of the rates of diagnosis of vestibular schwannomas between DGH and TH showed no significant difference (11 vs 21, chi-square test p = 0.6). Caution should be exercised in comparison of indication for scans between units due to the large amount of missing data from the DGH.

In this study, the detection rate for incidental findings on MRI IAM scan for unilateral audiovestibular symptoms was found in over a third of all scans at two different hospitals (34.7%). Vestibular schwannoma detection rate was 1.47%. The results of this study demonstrates that a tertiary hospital (TH), when compared to a district general hospital (DGH), accounts for a greater pick up rate of incidental findings but similar rates in detecting vestibular schwannomas .

The majority of incidental findings in both groups is small vessel/ischemic type change. It does not seem to be a clinically important difference. If this group of findings is ignored, the pickup rates of other potentially more significant findings are comparable between the two units. As there is not a universal protocol to report the scans, it is at the radiologist's discretion to mention less clinically significant incidental findings. All scans were reviewed by a consultant radiologist at the DGH and neuroradiologist at the TH. Both hospitals used 1.5 Tesla scanners and comparable protocols (1 mm slices through the IAM and 5 mm for rest of brain at the DGH and 1.5 mm slices through the IAM and 5 mm for rest of brain at the TH). The increased reporting of incidental findings may reflect the specialist presence at the TH.

A recent review of 200 MRI IAM scans in a population undergoing investigation for audiovestibular symptoms had an incidental finding rate of 47.5%, [12] while another study by Wong et al. found 46.7% (n = 396) [10] which are both higher than our report (34.7%). The population in our study included a younger age range than in other studies, (11–95 vs 53–92 [10] and 17–82) [12] years and thus our rates of ischemia were lower (15.8% vs 26.5% [10] and 33%) [12] which may account for this difference.

As reported previously, our results showed that abnormalities on MRI IAM scan are associated with increased age, the majority of which were ischemic or small vessel disease changes. [9] [13] The TH involved in this study does not provide pediatric services so we would expect this population to have a higher mean and median age, however, this is not the case.

With current practice, the pickup rate of vestibular schwannomas on MRI scan is less than 2% (TH 1.4%, DGH 1.7%). This rate showed no significant difference between the two hospitals in this study. There are no universal applied guidelines on whom to scan when investigating patients with audiovestibular symptoms. However, Obholzer et al. showed using a scan criterion of 15 dB or more inter-aural difference in two adjacent frequencies are used, if the mean threshold in the better ear was <30 dB, and 20 dB if greater than 30 dB on pure tone audiogram, sensitivity is increased to 97% and number of scans requested for these symptoms is reduced by 44%. [5] The consistent vestibular schwannoma pickup rate across the two hospitals are similar to previously reported diagnostic rates [14]and these values will be useful when counseling patients about the likelihood of diagnosing vestibular schwannoma when undertaking an MRI scan. Importantly for patients, the type of unit performing MRI scans does not seem to determine the likelihood of clinically significant pathology being detected.

The other factor to consider is whether it is appropriate to discuss all incidental findings with patients especially as many are asymptomatic or considered normal changes for age. In cases where carotid aneurysm or glioma was found, a neurosurgical referral was made. Advice exists for the follow-up and management of meningiomata and arachnoid cysts. [15] Knowledge of the management of incidental findings is important for two reasons; firstly, due to the pickup rates of vestibular schwannomas and incidental findings, we are more likely to make an incidental finding on imaging than to diagnose vestibular schwannomas. Secondly, incidental findings are more common in older patients and demographics are shifting, reflecting an older society. We recommend that units carrying out MRI IAM scans should have a policy outlining the follow-up and management of the most common or serious findings.

In conclusion, this study demonstrates the detection rates of incidental findings and vestibular schwannomas in a tertiary hospital and a district general hospital in the UK. The study compares pickup rate of clinically significant and insignificant incidental findings at different types of hospital and discusses their nature and management. Whilst we have shown there is a disparity in either the true incidence or reported rate of clinically insignificant incidental findings, importantly there is no difference in detection rates of clinically significant ones. We believe that the results from this report are most useful when counseling patients before and after receiving diagnostic imaging.

Dr Sarah Dawes, House Officer, Department of Otolaryngology, Queen Elizabeth Hospital, Birmingham, UK.

1. Park BJ, Kim HK, Sade B, et al. Epidemiology. In: Lee JH, editor. Me-ningiomas: Diagnosis, treatment, and outcome. London: Springer 2008. p11.    
2. Springborg JB, Poulsgaard L, Thomsen J. Nonvestibular Schwannoma Tumors in the Cerebellopontine Angle: A Structured Approach and Management Guidelines. Skull Base 2008;18(4):217–7.   [CrossRef]   [Pubmed]    
3. Lalwani AK. Meningiomas, epidermoids, and other nonacoustic tumors of the cerebellopontine angle. Otolaryngol Clin North Am 1992;25(3):707–28.   [Pubmed]    
4. Tos M. Natural history of vestibular schwannoma. In: Gleeson MJ, Browning G, Burton MJ, Clarke RC, Hibbert K, Jones N, et al. (eds.) Scott-brown's otorhinolaryngology: Head and neck surgery. 7th ed. London, UK: CRC Press 2008. P. 3957–66.    
5. Obholzer RJ, Rea PA, Harcourt JP. Magnetic resonance imaging screening for vestibular schwannoma: Analysis of published protocols. Journal of Laryngology and Otology 2004;118(5):329–2.   [CrossRef]   [Pubmed]    
6. Sidman JD, Carrasco VN, Whaley RA, Pillsbury HC 3rd. Gadolinium; The New Gold Standard for Diagnosing Cerebellopontine Angle Tumors. Arch Otolaryngol Head & Neck Surg 1989;115(10):1244–7.   [CrossRef]   [Pubmed]    
7. Curati WL, Graif M, Kingsley DP, King T, Scholtz CL, Steiner RE. MRI in acoustic neuroma: a review of 35 patients. Neuroradiology 1986;28(3):208–14.   [CrossRef]   [Pubmed]    
8. Régis JM, Roche PH. Modern management of acoustic neuroma. Basal: S Karger Publishing 2008.p. 123.    
9. Morris Z, Whiteley WN, Longstreth WT Jr, et al. Incidental findings on brain magnetic resonance imaging: systematic review and meta-analysis. BMJ 2009;339:b3016.   [CrossRef]   [Pubmed]    
10. Wong BY, Capper R. Incidence of vestibular schwannoma and incidental findings on the magnetic resonance imaging and computed scans on patients from a direct referral audiology clinic. J Laryngol Otol 2012;126(7):658–2.   [CrossRef]   [Pubmed]    
11. Powell HR. Choa DI. Should all patients referred for magnetic resonance imaging scans of their internal auditory meatus be followed up in ENT clinics? Eur Arch Otorhinolaryngol 2010;267(9):1361–6.   [CrossRef]   [Pubmed]    
12. Papanikolaou V, Khan MH, Keogh IJ. Incidental findings on MRI scans of patients presenting with audiovestibular symptoms. BMC Ear Nose Throat Disord 2010;10:6.   [CrossRef]   [Pubmed]    
13. Van Dijk EJ, Prins ND, Vrooman HA, Hofman A, Koudstaal PJ, Breteler MM. Progression of Cerebral Small Vessel Disease in Relation to Risk Factors and Cognitive Consequences: Rotterdam Scan Study. Stroke 2008;39(10):2712–9.   [CrossRef]   [Pubmed]    
14. Vandervelde C, Connor SE. Diagnostic yield of MRI for audiovestibular dysfunction using contemporary referral criteria: correlation with presenting symptoms and impact on clinical management. Clin Radiol 2009;64(2):156–3.   [CrossRef]   [Pubmed]    
15. Pradilla G, Jallo G. Arachnoid cysts: case series and review of the literature. Neurosurg Focus 2007;22(2):E7.   [CrossRef]   [Pubmed]