Comparison of the EAONO/JOS, STAMCO and ChOLE cholesteatoma staging systems in the prognostic evaluation of acquired middle ear cholesteatoma in children

Machine learning application in otology

This review presents a comprehensive history of Artificial Intelligence (AI) in the context of the revolutionary application of machine learning (ML) to medical research and clinical utilization, particularly for the benefit of researchers interested in the application of ML in otology. To this end, we discuss the key components of ML-input, output, and algorithms. In particular, some representation algorithms commonly used in medical research are discussed. Subsequently, we review ML applications in otology research, including diagnosis, influential identification, and surgical outcome prediction. In the context of surgical outcome prediction, specific surgical treatments, including cochlear implantation, active middle ear implantation, tympanoplasty, and vestibular schwannoma resection, are considered. Finally, we highlight the obstacles and challenges that need to be overcome in future research.

Pediatric Cholesteatoma Follow-Up: Residual and Recurrence in 239 Cases with Over 5-Year Hindsight

INTRODUCTION

Pediatric cholesteatoma is an aggressive disease which requires long-term evaluation to assess management strategies. The objective was to determine optimal follow-up duration in pediatric cholesteatoma to detect residual and recurrent diseases.

METHODS

This cohort study was set in a tertiary referral center. All consecutive patients with a minimum 5-year follow-up were included. Medical history, initial extension, surgical procedures, and follow-up were collected. The main outcome measure were Kaplan-Meier survival curves of residual and recurrence cumulative incidence.

RESULTS

Totally 239 ears with the first tympanoplasty between 2008 and 2014 were studied including 25% congenital. At first surgery (S1), mean age was 8.4 years and mean follow-up time 7.9 years. Mastoidectomy was performed in 69% and stapes was absent in 38% of cases. Notably, 83% (199 ears) had a second procedure (S2) of which 186 were planned. After S1, maximum cumulated incidence of residual was 45% [95%CI, 38%; 52%] at 74 months, with probability of residual of 39% at 3 years. After S2 (n = 199), maximum cumulated residual incidence was 21% [95%CI, 12%; 32%] at 62 months and 16% at 3 years. Concerning recurrence, maximum cumulated incidence after S1 (n = 239) was reached at 98 months with 21% [95%CI, 12%; 32%], 13% at 3 years and 16% at 5 years. Congenital disease had significantly less residuals after S1 (p = 0.02), but similar recurrence rate (p = 0.66) compared with acquired.

CONCLUSIONS AND RELEVANCE

We recommend MRI follow-up of at least 5 years after the last surgery for residual disease and clinical follow-up of at least 10 years to detect recurrence.

LEVEL OF EVIDENCE

4 Laryngoscope, 2024.

Cholesteatoma Severity Determines the Risk of Recurrent Paediatric Cholesteatoma More Than the Surgical Approach

OBJECTIVE

To evaluate factors that influence the rate of cholesteatoma recurrence (growth of new retraction cholesteatoma) in children.

METHODS

Review of children with primary acquired or congenital cholesteatoma. Severity was classified by extent and EAONO-JOS stage, and surgery by SAMEO-ATO. Primary outcome measure was 5-year recurrence rate using Kaplan-Meier or Cox regression analysis.

RESULTS

Median age was 10.7 years for 408 cholesteatomas from which 64 recurred. Median follow up was 4.6 years (0-13.5 years) with 5-year recurrence rate of 16% and 10-year of 29%. Congenital cholesteatoma (n = 51) had 15% 5-year recurrence. Of 216 pars tensa cholesteatomas, 5-year recurrence was similar at 14%, whereas recurrence from 100 pars flaccida cholesteatomas was more common at 23% (log-rank, p = 0.001). Sub-division of EAONO-JOS Stage 2 showed more recurrence in those with than without mastoid cholesteatoma (22.1% versus 10%), with more in Stage 3 (31.9%; p = 0.0003). Surgery without mastoidectomy, including totally endoscopic ear surgery, had 11% 5-year recurrence. Canal wall-up tympanomastoidectomy (CWU) and canal wall-down/mastoid obliteration both had 23% 5-year recurrence. Multivariate analysis showed increased recurrence for EAONO-JOS Stage 3 (HR 5.1; CI: 1.4-18.5) at risk syndromes (HR 2.88; 1.1-7.5) and age < 7 years (HR 1.9; 1.1-3.3), but not for surgical category or other factors.

CONCLUSION

Young age and more extensive cholesteatoma increase the risk of recurrent cholesteatoma in children. When controlling for these factors, surgical approach does not have a significant effect on this outcome. Other objectives, such as lower post-operative morbidity and better hearing outcome, may prove to be more appropriate parameters for selecting optimal surgical approach in children.

Performance of Non-EPI DW MRI for Pediatric Cholesteatoma Follow-Up

OBJECTIVE

To evaluate the accuracy, sensitivity, and specificity of nonecho planar (non-EPI) diffusion-weighted (DW) magnetic resonance imaging (MRI) to detect residual cholesteatoma in children.

STUDY DESIGN

Retrospective study.

SETTING

Tertiary comprehensive hospital.

METHODS

Children operated on for a first-stage cholesteatoma procedure from 2010 to 2019 were included. MRIs were performed with non-EPI DW sequences. Initial reports were collected, indicating the presence or absence of hyperintensity suggestive of cholesteatoma. Three hundred twenty-three MRIs were correlated with the subsequent surgery (66%) or year-later MRI (21%), or were considered accurate if performed 5 years or more after the last surgery (13%). The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of each imaging procedure for the detection of cholesteatoma were calculated.

RESULTS

Two hundred twenty-four children with mean age of 9 ± 4 years old presented with cholesteatoma. MRIs were performed 27 ± 24 months after surgery. Residual cholesteatoma was diagnosed in 35%. The sensitivity, specificity, PPV, and NPV of MRI were 62%, 86%, 74%, and 78%, respectively. Accuracy, sensitivity, and specificity increased significantly over time (multivariate analysis). The mean delay after last surgery was of 30 ± 2.0 months for accurate MRI (true positive or negative) versus 17 ± 2.0 months for nonaccurate (false positive or negative) MRIs (p < .001).

CONCLUSION

However, long the delay after the last surgery, the sensitivity of non-EPI diffusion sequence MRI in children has limitations for the detection of residual cholesteatoma. Surveillance for residual cholesteatoma should incorporate findings at primary surgery, surgeon experience, a low threshold for second-look procedures, and routine imaging.