|Home | About | Journals | Submit | Contact Us | Français|
In patients who underwent neurosurgery for excision of acoustic tumors, we used correlation and multivariate logistic regression analysis to study relationships among nine variables thought to have value in predicting hearing preservation. These variables included auditory perceptual, auditory neurophysiologic, and imaging-related characteristics of acoustic tumor dimensions. The univariate correlations showed two general trends. The first trend demonstrated relationships among imaging-related acoustic tumor dimensions (size or porus acousticus widening) with either postoperative hearing or brainstem auditory evoked potential (BAEP) abnormalities (tumor size and postoperative hearing; porus acousticus widening and postoperative hearing; tumor size and preoperative BAEP abnormalities). The second trend reflected interrelationships among preoperative audiometric variables that were not related to postoperative hearing outcome. Logistic regression analysis of eight independent variables on the presence or absence of postoperative hearing found that porus acousticus widening was the best prognostic indicator for hearing outcome. When the porus acousticus was widened, the odds ratio was 11:1 that hearing would be lost postoperatively. Viewed as a whole, the literature on prognostic variables related to hearing preservation has been limited, almost entirely, to univariate relationships. To determine more accurately which preoperative variables have unique prognostic value, more complex multivariate, analysis procedures will be required. Additionally, standardized criteria for assessment and reporting of auditory test results are also needed.