Modeling residual cholesteatoma growth rate: a meta-analysis of individual participant data

To create a model of residual cholesteatoma growth rate that can be used to estimate optimal timing of follow-up imaging.

Following the recommendations of the PRISMA-IPD guideline and after registering our study protocol on PROSPERO (CRD42023465320), a systematic search was completed in September 2023 and updated in July 2024, conducted in PubMed, Web of Science, Embase, and Cochrane Library. Reference lists of the resulting articles were also searched for potential inclusions. Our search yielded 11,274 articles. Studies with patients of any age and sex diagnosed with residual cholesteatoma were included if data on cholesteatoma size at follow-up were available, collected from imaging results or through second-stage surgery. For statistical analysis, a Monte Carlo simulation approach was applied, and a baseline random-intercept mixed-effects model was fitted.

We included 11 articles reporting individual data from 163 patients. All but one of these articles proved to be high risk of bias. Our simulation models suggest that 95% of cholesteatomas reach or grow past the size of 3 millimetres (threshold for imaging) after 28-35 months and 95% of reach or grow past the size of 5 millimetres (threshold for intraoperative identification) after 47-58 months.

Analyzing previously published data suggests that residual cholesteatoma may only be radiologically detectable on non-EPI DWI MRI after 28-35 months postoperatively in 95% of patients, indicating delayed follow-up for average risk patients compared to current clinical practice. However, this simulation model, based on retrospective data, aims to uncover growth patterns rather than posing as a new guideline.