There is a clear clinical need to reliably detect residual cholesteatoma after canal wall up mastoid surgery. Ideally, this would be achieved through non-invasive radiological means rather than second-look surgery, thus preventing morbidity in those patients in whom no residual disease is found.

We describe a case in which non-echo-planar, diffusion-weighted magnetic resonance imaging sequences were used pre-operatively, and compared with subsequent surgical findings. This case highlights both the potential of this increasingly popular magnetic resonance technique and also its current limitations.

Various magnetic resonance sequencing types have been employed to try to reliably detect residual cholesteatoma, each with varying success. Non-echo-planar, fast-spin echo, diffusion-weighted sequences currently appear to be the most reliable at detecting even the smallest pearl of cholesteatoma, down to 2 mm in diameter. In our unit, a propeller, diffusion-weighted image sequence is employed on a GE Signa scanner. However, both this case study and other reports show that the accuracy of the technique is not 100 per cent. This begs the question of how much one can rely on the findings of such techniques when deciding whether second-look surgery is indicated. Scan-negative patients will require continued follow up as, at the time of imaging, residual disease may not have reached a detectable size.