Value of echo-planar diffusion-weighted MRI in the detection of secondary and postoperative relapsing/residual cholesteatoma.

Purpose: To assess the utility of echo-planar diffusion-weighted magnetic resonance imaging (EPI-DWI) in the diagnosis of secondary cholesteatoma of the middle ear and in the differential diagnosis between residual/relapsing cholesteatoma and non-cholesteatomatous tissues (scar, granulation and inflammatory tissue) after conservative mastoidectomy.

Materials And Methods: Twenty-four patients, who had previously undergone clinical and CT investigation, were prospectively examined by standard and EPI-DWI magnetic resonance imaging (MRI) after conservative mastoidectomy. Secondary cholesteatoma was suspected in 5 patients and residual/relapsing cholesteatoma in 19 patients. Two radiologists, blinded to patient's identity, clinical data and CT findings, reached consensus on the presence of tissue consistent with cholesteatoma in conventional CT and areas of altered signal in EPI-DWI in the petrous bone. All patients underwent mastoidectomy, second time of tympanoplasty or review surgery within 15 days from MR investigation. Sensitivity, specificity and negative and positive predictive values were evaluated separately for standard and EPI-DWI MRI.

Results: In EPI-DWI sequences, 11/12 patients with cholesteatoma showed an area of hyperintense signal, whereas patients with non-cholesteatomatous tissue showed no pathologic signal in the petrous bone. In the single case of cholesteatoma undetected on EPI-DWI a cholesteatomatous pearl approx. 2 mm in diameter was visible in the surgical cavity. Sensitivity, specificity, positive and negative predictive values were 92%, 100%, 100%, 92% for EPI-DWI MRI and 92%, 25%, 55%,75% for standard MRI, respectively.

Conclusions: EPI-DWI sequences are useful in the diagnosis of secondary cholesteatoma and in the differential diagnosis between residual/relapsing cholesteatoma and non-cholesteatomatous tissues after conservative mastoidectomy. However, the usefulness of EPI-DWI sequences as a screening test after conservative mastoidectomy requires further assessment.