Objectives: To objectify the effects of retrocochlear disease on distortion-product otoacoustic emissions (DPOAEs) by developing a computer-based software strategy for classifying DPOAE patterns as cochlear or noncochlear and to evaluate the sensitivities of these techniques in a large series of patients with unilateral acoustic neuromas.
Study design: Development of a novel, software-based method of DPOAE analysis, which was evaluated with data obtained from a retrospective review of the results from audiometric tests performed in a series of patients.
Methods: A computer-based software strategy was developed, using frequency-specific data from normal-hearing adults, for the purpose of distinguishing cochlear from noncochlear patterns of hearing loss, by determining the discrepancies between DPOAEs and behavioral audiometry. Preoperative pure-tone thresholds and DPOAEs from 97 patients with surgically confirmed acoustic neuroma were compared using an objective method and a standard, subjective technique that was considered to be the gold standard. The effects of bilateral hearing losses, such as noise-induced hearing loss and presbycusis, were accounted for during the analysis to isolate the effects of the tumors on hearing thresholds and DPOAEs.
Results: Overall, 55 (57%) of the tumor ears were assigned to the cochlear group (i.e., DPOAEs consistent with hearing thresholds), 40 (41%) to the noncochlear group (i.e., DPOAEs inconsistent with hearing thresholds), and 2 (2%) to an indeterminate group, using the subjective technique for classifying DPOAEs. There was no significant difference in the categorization of the patients with acoustic neuroma when employing the objective strategy. The objective algorithm, when modified to maximize the number of noncochlear identifications, led to assignments of 36 (37%) to the cochlear, 57 (59%) to the noncochlear, and 4 (4%) to the indeterminate categories.
Conclusions: Subjective analysis of a large series patients with acoustic neuromas showed that the majority of ears with tumors demonstrated cochlear (57%), rather than non-cochlear (41%), patterns of DPOAEs. The computerized, software-based algorithm developed for differentiating cochlear from noncochlear patterns of DPOAEs in patients with retrocochlear disease had a maximum sensitivity of 59%. This value was significantly higher than that reported in previous studies.