Bibliogr. s. e57

Purpose: The aim of the study was to analyse the value of 2 different diffusion-weighted imaging (DWI) techniques (echo-planar imaging [EPI] and on-echo-planar imaging [non-EPI]) in the diagnosis of cholesteatoma. Material and methods: Our material consisted of 32 subjects suspected of cholesteatoma, who underwent magnetic resonance imaging of the temporal bone using both EPI and non-EPI DWI. Two independent readers retrospectively analysed magnetic resonance images. Intra- and interobserver agreements as well sensitivity, specificity, and negative (NPV) and positive (PPV) predictive values of both DWI sequences were assessed. Results: Using non-EPI DWI all cholesteatomas were correctly diagnosed by both readers with no false negative nor inconclusive cases and with only one false positive result. Non-EPI DWI revealed high interobserver agreement (k = 1) and high correlation with histopathological results (r = 0.895). EPI DWI misdiagnosed 27-31% of cholesteatomas (false negative results), showing also significantly low interobserver agreement (k = 0.373) and low correlation with histopathological results (r = 0.328 for reader 1 and r = 0.267 for reader 2). Non-EPI DWI revealed very high sensitivity (100%), specificity (83.3%), NPV (100%), and PPV (96.3%) in comparison to EPI DWI, which showed lower sensitivity (69.2%), specificity (66.6-83.3%), NPV (33.3-38.4%), and PPV (90.0-94.7%). Conclusions: Non-EPI DWI with high sensitivity, specificity, and interobserver agreement is a very reliable technique in detecting middle ear cholesteatoma regardless of the pre- or postoperative state of the ear, and it should entirely replace EPI DWI in clinical practice.