Comparison of 3 ABR Methods for Diagnosis of Retrocochlear Hearing Impairment

BACKGROUND The purpose of this study was to compare the effectiveness of methods for screening for retrocochlear pathologies based on auditory evoked brainstem responses (ABRs). The study compared the sensitivity, specificity, and effectiveness of these 3 techniques. MATERIAL AND METHODS The methods were: (i) standard ABR utilizing click-evoked responses, (ii) stacked ABR based on derived-band responses, and (iii) ABRs evoked by tone-pips (ABR TP). The methods were tested on patients with retrocochlear pathologies confirmed by MRI-Gd, normal-hearing subjects, and patients with cochlear hearing loss. The system and software used in the tests was NavPro AEP v.6.2.0 (BioLogic - Natus). Prior to testing, all subjects were given comprehensive audiologic and otologic examinations, including MR imaging. Sensitivity and specificity functions and predictive values of methods were determined. RESULTS The stacked ABR method as realized in the NavPro system exhibited high sensitivity but specificity was very low, due to the high variability of stacked ABR amplitudes. The standard ABR method had good specificity, but low sensitivity in cases of small tumors (below 1 cm in diameter). Best sensitivity and specificity was obtained with the ABR TP method. CONCLUSIONS The stacked ABR method allows small acoustic tumors to be detected, but produces high percentage of false positive results. The ABR TP method offers good sensitivity and specificity, and relatively high predictive value. The best option would be to use a two-stage screening, consisting of a standard ABR in the first stage and an ABR TP test in the second.