The diagnostic accuracy of non-imaging screening protocols for vestibular schwannoma in patients with asymmetrical hearing loss and/or unilateral audiovestibular dysfunction: a diagnostic review and meta-analysis.

Abstract

Currently, all patients presenting with asymmetrical sensorineural hearing loss and/or unilateral audiovestibular dysfunction (i.e. tinnitus, dizziness) undergo MRI, leading to a substantial amount of MRIs with negative findings as the incidence of vestibular schwannoma (VS) in this screening population varies between 1% and 4.7% (i.e. more than 95% of MRIs are negative for VS). The aim was to assess the diagnostic accuracy of different non-imaging screening protocols that can be used prior to MRI to select patients at high risk of VS. Diagnostic review and meta-analysis. We systematically searched MEDLINE, Embase and The Cochrane Library as from inception up to 28 July 2016. We included studies that compared non-imaging screening protocols to MRI as gold reference standard. Methodological quality was assessed by two independent reviewers using the Quality Assessment of Diagnostic Accuracy Studies tool. Data necessary to complete 2 × 2 tables were obtained, and patient, study, screening and imaging characteristics were extracted. We calculated sensitivity and specificity of all tests and obtained pooled estimates using a bivariate random effects model. We analysed 12 studies (4969 patients) of poor to moderate quality according to the quality assessment. Most studies tested diagnostic accuracy of multiple screening protocols. Five pure-tone audiometry (PTA) protocols were studied by multiple authors; pooled estimates for sensitivity ranged from 88% [95% CI: 84-91] to 91% [95% CI: 52-99] and specificity from 31% [95% CI: 10-66] to 58% [95% CI: 49-65]. Due to heterogeneity, we were unable to pool other tests. In five studies testing auditory brainstem response, sensitivity values ranged from 37% [95% CI: 23-52] to 100% [95% CI: 40-100] and specificity from 57% to 96% [95% CI: 87-100]. Two authors studied PTA shape as a screening test. Presenting symptoms, electronystagmography, caloric irrigation and hyperventilation test were assessed by one study each. All reported low diagnostic accuracy. All identified studies had a moderate-to-high risk of bias, and none of the currently available non-imaging screening protocols appear to be accurate in detecting VSs.