Cholesteatoma Localization Using Fused Diffusion-Weighted Images and Thin-Slice T2 Weighted Images

Background Fusion of diffusion-weighted imaging (DWI) to computed tomography (CT) has been touted as a possible technique to improve cholesteatomas localization. This study set out to assess the ability of DWI images fused with thin-slice heavily T2-weighted images to similarly localize surgically-confirmed cholesteatomas. Materials and Methods A retrospective review was completed of consecutive patients that underwent MR temporal bone imaging (9/2011-3/2020) with both DWI and thin-slice T2-weighted imaging. Included patients underwent surgical resection of primary or recidivistic cholesteatoma after preoperative MR imaging. A neuroradiologist, who was blinded to operative and clinical notes, localized each lesion on both DWI and fused DWI-T2 images in 11 anatomic subdivisions of the temporal bone. Surgical confirmation of cholesteatoma location was used as the gold standard for comparison. Results Of 24 included examinations, the average age at time of MR was 48.2 +/- 24.7 years; 12/24 (50.0%) were female. Five of 24 had primary cholesteatoma, while the remainder had recidivistic disease. Sensitivity, specificity, and accuracy of unfused DWI images were 52.1%, 88.9%, and 75.8%, respectively. Sensitivity, specificity, and accuracy of fused DWI-T2 images were 57.1%, 94.8%, and 81.8%, respectively. Conclusion Fused DWI-T2 images outperformed unfused DWI images in the anatomic localization of temporal bone cholesteatomas. This method represents a potential alternative to MR-CT fusion for pre-operative cholesteatoma evaluation, and warrants future investigations. Opposed to MR-CT fusion, this method only necessitates MR sequences and removes the need for additional CT acquisition. Level of Evidence 3 Laryngoscope, 2020