Diffusion-weighted magnetic resonance imaging for residual and recurrent cholesteatoma: a systematic review and meta-analysis

BackgroundDiagnosis and management of recurrent or residual cholesteatoma can be problematic. Diffusion-weighted imaging magnetic resonance imaging (MRI) sequences have been used for follow-up of such lesions. More recent non-echoplanar imaging (non-EPI) sequences are thought to be superior to older echoplanar imaging (EPI) sequences. Objective of reviewEvaluate whether diffusion-weighted magnetic resonance imaging is useful in the diagnosis of recurrent or residual cholesteatoma. Type of reviewSystematic review and meta-analysis. Search strategyMEDLINE, EMBASE, CINAHL, Web of Science and the Cochrane Database were searched, with no limits on date or language. Study SelectionAdults or children who had previously undergone tympanomastoid surgery by any method with confirmation of recurrence/residual disease by second-look/revision surgery. Evaluation methodsTwo reviewers independently reviewed studies. Data extracted on 11 domains and rechecked. Data synthesisStatistical analysis with SPSS. ResultsA total of 575 studies were identified of which 27 met the inclusion criteria. These covered 727 patient episodes. For EPI studies: sensitivity (sd) 71.82 (24.5), specificity (sd) 89.36 (13.4), PPV (sd) 93.36 (8.1) and NPV (sd) 73.36 (15.8). For non-EPI studies: sensitivity 89.79 (12.1), specificity (sd) 94.57 (5.8), PPV (sd) 96.50 (4.2) and NPV 80.46 (20.2). Improved sensitivity of non-EPI sequences reached significance (P = 0.02). ConclusionsDiffusion-weighted MRI is both sensitive and specific for the detection of recurrent or residual cholesteatoma following ear surgery. Non-EPI techniques are superior to EPI techniques.