Cholesteatoma: 2026 so far

Signals strengthened, but reviews stress heterogeneous, non-high-level evidence.

In January 2026 to September 2026, several reports linked adding mastoid/epitympanic obliteration (including S53P4 bioactive glass) to fewer recurrences or recidivism compared with non-obliteration approaches. However, the main comparative data were retrospective (looking back at past cases). Later reviews also emphasized that the evidence is mixed and varies a lot between studies, with overall limited quality. One small series of recurrent disease suggested that when recurrence happened after obliteration, it often involved the epitympanum (the upper part of the middle ear) and lined up with incomplete epitympanic obliteration on CT, which points to a possible technical weak spot rather than a guarantee of prevention. Safety summaries for S53P4 bioactive glass reported low pooled rates of wound infection/extrusion and stable inner-ear hearing, while also noting postoperative drainage can still occur. This means obliteration is repeatedly studied as a way to lower recurrence risk, but the size of the benefit and who benefits most are still uncertain in the available evidence.

Research covered in: January 2026, April 2026, September 2026, August 2026, March 2026

More head-to-head comparisons reported similar control in selected cases.

From February through September, several observational comparisons (non-randomized studies comparing groups) reported that transcanal endoscopic ear surgery (using a camera through the ear canal) and, in one study, robot-assisted endoscopic surgery achieved disease control that was broadly similar to microscope-based surgery in selected cholesteatoma cases. Some reports also found shorter operations and quicker canal skin healing. Short-term patient experience was included in some evidence: one pediatric comparison linked endoscopic transcanal surgery to less pain and higher quality-of-life scores at one month than a behind-the-ear microscopic approach. At the same time, feasibility series for more extensive disease still reported recurrence and occasional nerve-related symptoms, showing that the choice of technique does not remove the need for follow-up. In practice, this means endoscopic vs microscopic approaches are increasingly compared using hearing, recovery experience, and recurrence/residual outcomes, but most comparisons are still non-randomized and depend heavily on the specific situation.

Research covered in: February 2026, March 2026, April 2026, July 2026, August 2026

More attention to patient-impact outcomes beyond disease removal alone.

In the months covered, studies paid more attention to recovery outcomes that affect daily life, such as having a dry ear, quality of life, taste disturbance, and hearing results after reconstruction. A meta-analysis (a study that combines results across studies) across middle ear operations found taste disturbance was less common when the chorda tympani nerve (a taste nerve that runs through the middle ear) was preserved, and taste symptoms often decreased over time. It also noted that persistent symptoms could reduce oral satisfaction and quality of life. Several cohorts highlighted that hearing improvement varies and depends on surgical and anatomic details. For example, stapes preservation (keeping a small middle-ear bone) during titanium ossiculoplasty (rebuilding the hearing bones with titanium) was linked to better air-bone gap closure (a measure of conductive hearing loss). One cohort also reported that many patients with suboptimal postoperative hearing met criteria for revision surgery, an implantable hearing device, or both when formally reassessed. This supports that follow-up often covers more than recurrence alone, because hearing, taste, and other symptoms may lead to additional procedures or supportive options.

Research covered in: January 2026, February 2026, June 2026, July 2026, September 2026, April 2026

Evidence converged that detection can be delayed, supporting longer surveillance.

A meta-analysis and modeling work suggested that after surgery, diffusion-weighted MRI detection of residual/recurrent cholesteatoma often becomes more informative after the first year. The signal can extend into multi-year timeframes, including an estimate that many residual lesions may reach a detectable size after roughly 28-35 months. In a small prospective pilot (a small forward-looking study), selected patients with very small (<=5 mm), symptom-free recurrences seen on diffusion MRI were followed with repeat MRI every 12-18 months. Most stayed surgery-free during follow-up, although some later needed surgery because of growth or symptoms. Separately, a cadaver study (done on donated bodies) reported that one bone-conduction implant created MRI distortion that made the implanted-side temporal bone hard to evaluate, including diffusion imaging used for surveillance. This points to imaging follow-up as a long-timeline process, but it may not work the same way for everyone (including people with implants), and evidence for MRI-only monitoring is still limited to small cohorts.

Research covered in: March 2026, June 2026, August 2026, May 2026

Combined imaging improved confidence; AI ranged from helpful to unreliable.

In January 2026 to September 2026, diagnostic imaging evidence moved forward in two areas: how well standard scans perform and how experimental computer tools might help. One retrospective study reported that reading high-resolution CT together with non-echo-planar diffusion-weighted MRI improved diagnostic performance compared with either test alone. A separate prospective study found high-resolution CT closely matched surgical findings for several common erosion sites. AI findings were mixed: single-center CT deep-learning models (AI trained to recognize patterns in images) reported high performance and improved clinician accuracy with model support in one study. In contrast, general-purpose multimodal large language models performed poorly as stand-alone readers on limited CT image sets, with very low specificity. In practical terms, CT plus diffusion MRI is increasingly supported for more confident identification and localization, while AI help is still inconsistent and not reliably ready to act as an independent interpreter in the reported setups.

Research covered in: August 2026, June 2026, May 2026