Auditory brainstem response versus magnetic resonance imaging for the evaluation of asymmetric sensorineural hearing loss.

OBJECTIVES/HYPOTHESIS: Auditory brainstem response (ABR) testing and magnetic resonance imaging (MRI) are compared for the evaluation of patients with asymmetric sensorineural hearing loss (SNHL). MRI with gadolinium administration is the current gold standard for identifying retrocochlear lesions causing asymmetric SNHL. The study seeks to determine the sensitivity and specificity of ABR in screening for possible retrocochlear pathology. Most important among SNHL etiologies are neoplastic lesions such as vestibular schwannomas, cerebellopontine angle (CPA) tumors, as well as multiple sclerosis, stroke, or other rare nonneoplastic causes. The study results will allow the author to recommend a screening algorithm for patients with asymmetric SNHL. STUDY
DESIGN: The study is a multi-institutional, institutional review board approved, prospective, nonrandomized comparison of ABR and MRI for the evaluation of patients with asymmetric SNHL.
METHODS: Three hundred twelve patients (between the ages of 18 and 87) with asymmetric SNHL completed the study. Asymmetric SNHL was defined as 15 dB or greater asymmetry in two or more frequencies or 15% or more asymmetry in speech discrimination scores (SDS). These patients prospectively underwent both ABR and MRI. The ABR and MRI were interpreted independently in a blinded fashion. In addition to the ABR and MRI results, a variety of clinical and demographic data were collected.
RESULTS: Thirty-one (9.94%) patients of the study population of 312 were found on MRI to have lesions causing their SNHL. Of the 31 patients with causative lesions on MRI there were 24 vestibular schwannomas, 2 glomus jugulare tumors, 2 ectatic basilar arteries with brainstem compression, 1 petrous apex cholesterol granuloma, 1 case of possible demyelinating disease, and 1 parietal lobe mass. Twenty-two of the 31 patients had abnormal ABRs, whereas 9 patients (7 with small vestibular schwannomas) had normal ABRs. This gives an overall false-negative rate for ABR of 29%. The false-positive rate was found to be 76.84%. Sensitivity of ABR as a screening test was 71%, and specificity was 74%.
CONCLUSIONS: Ten percent of patients with asymmetric SNHL (by this study's criteria) are likely to have causative lesions found on MRI. Although the recently reported annual incidence of vestibular schwannoma in the general population is 0.00124%, for patients with asymmetric SNHL in this study, the incidence was 7.7% (nearly 4 orders of magnitude higher). ABR has been demonstrated to have low sensitivity and specificity in the evaluation of these patients and cannot be relied on as a screening test for patients with asymmetric SNHL. Keeping the use of MRI conditional on the results of ABR will annually result in missed or delayed diagnosis of causative lesions in 29 patients per 1,000 screened. The author recommends abandoning ABR as a screening test for asymmetric SNHL and adoption of a focused MRI protocol as the screening test of choice (within certain guidelines).