August 2026 Research Update

This months research focused on how imaging supports key decisions in cholesteatoma: confirming whether cholesteatoma is present, describing its location, and deciding whether repeat surgery is needed after prior surgery. Overall, studies suggested that using both a detailed CT scan and diffusion-weighted MRI together can improve diagnostic confidence, while MRI-based monitoring may be an option for some small, symptom-free postoperative recurrences. At the same time, early testing of general-purpose artificial intelligence image readers showed poor accuracy for distinguishing cholesteatoma from other chronic ear disease when used on limited CT images.

In a retrospective study of 85 people (65 confirmed cholesteatoma and 20 controls), reading high-resolution CT together with non-echo-planar diffusion-weighted MRI had the best overall diagnostic performance compared with either test alone. The combined approach reached 93.8% sensitivity and 92.0% specificity, with a reported average location error of 1.8 mm and strong agreement between readers (kappa 0.88). The authors noted that whether this improved accuracy changes surgical outcomes still needs prospective confirmation. [1]

A separate retrospective single-center study tested multimodal large language models (GPT-5 and Gemini 2.5 Pro) as image readers using six key CT images per case in 101 surgical patients. The radiologist achieved 96.0% accuracy, while the models had low accuracy (about 44% to 50%), with very high sensitivity but extremely low specificity (1.9% to 13.2%), meaning they frequently labeled non-cholesteatoma cases as cholesteatoma. Reproducibility was only fair to moderate across repeated runs, supporting the conclusion that these tools were not suitable as independent second readers for this task in this study setup. [2]

One study explored whether detailed CT measurements of temporal bone anatomy could help predict who might have residual or recurrent disease after cholesteatoma surgery. Seven of ten measurement methods had acceptable agreement between observers, and some anatomic differences were seen between cholesteatoma cases and healthy controls (including mastoid ventilation and tegmen height), but the measurements did not correlate with residual or recurrent disease. This suggests that while some measurements can be made consistently, their value for forecasting recurrence was not shown here. [3]

A small prospective pilot study followed 13 people who had prior cholesteatoma surgery and then developed a small (5 mm or less) recurrence seen on diffusion-weighted MRI but had no ear symptoms. During about 43 months of follow-up with repeat MRI every 12 to 18 months, 77% stayed surgery-free, while 23% later needed surgery because the lesion grew or symptoms developed; surgery-free survival was reported as 100% at 24 months and 91% at 36 months. Some lesions decreased on imaging without surgery (23%), supporting that imaging-guided surveillance was feasible in this selected group, although safety outcomes were not reported and the study was small. [4]

For patients and caregivers, these studies mainly clarify how imaging may be used to support diagnosis and follow-up decisions, rather than introducing new treatments. The strongest practical message is that CT and diffusion-weighted MRI can provide complementary information, and in one study their combined reading improved accuracy compared with either alone. For people who have already had cholesteatoma surgery, one small study suggests that careful MRI monitoring may allow some symptom-free patients with very small recurrences to avoid immediate repeat surgery, but it remains uncertain how broadly this applies and what the full safety trade-offs are. Finally, early testing suggests that current general-purpose artificial intelligence models should be viewed as background research in this area, because they were not reliable as stand-alone interpreters of limited CT image sets in the study reported.

Taken together, these retrospective studies suggest that less invasive viewing approaches through the ear canal (endoscopic surgery, including robot-assisted endoscopy) generally achieved disease control similar to traditional microscope-based surgery in selected cholesteatoma cases. Some studies also reported practical advantages such as shorter operating time, shorter hospital stay, quicker skin healing of the ear canal, and in one study better hearing results, but the authors noted that differences in reconstruction choices and non-randomized designs limit how confidently we can compare approaches. [5] [6] [7]

In people with limited cholesteatoma, transcanal endoscopic ear surgery (TEES) and microscopic ear surgery (MES) had similar rates of residual or recurrent disease and similar major complication rates. TEES had better postoperative hearing measures in this study and was linked with shorter surgery, anesthesia time, and hospital stay, although the authors cautioned that differences in when and how hearing reconstruction was done could partly explain the results. [5]

For external auditory canal cholesteatoma (cholesteatoma in the ear canal), TEES had a shorter average operative time and faster time to complete skin healing of the canal than MES. In this report, no recurrence or canal narrowing was described after TEES, while MES had some recurrences, canal narrowing or closure, and one case of temporary facial weakness. [6]

In a 40-person adult study with 3-year hearing tests and diffusion-weighted MRI follow-up, robot-assisted endoscopic cholesteatoma surgery had complete eardrum healing in all cases and no robot-related complications reported. Hearing outcomes were broadly similar between robot-assisted and microscope surgery, and residual cholesteatoma was found at similar rates in both groups. [7]

For patients and caregivers, this set of studies mainly adds background about how different surgical tools and viewpoints compare rather than proving one best option for everyone. The evidence here comes from moderate-sized, retrospective comparisons, so differences between groups (such as who was selected for which technique and how hearing reconstruction was timed) could influence results, and larger prospective studies are still needed. What is clearer is that endoscopic approaches, including robot-assisted endoscopy in one study, were reported to be workable and had broadly similar disease-control results to microscope surgery in the studied settings, with some reports of shorter procedures and quicker healing.

One retrospective cohort study focused on matching the operation to disease stage for external auditory canal cholesteatoma, using a CT-based staging system to guide surgical planning. Overall, recurrence was uncommon in this series, and reported postoperative issues such as infection or canal narrowing were typically addressed with follow-up care or, when needed, revision surgery. [8]

Among 59 patients (61 ears), recurrence occurred in 2 ears (3.28%), and both were in Stage II and were reported as cured after revision surgery. Postoperative ear canal infection occurred in 3 ears (4.92%) and restenosis (re-narrowing) in 2 ears (3.28%), and these were reported as resolved after management. [8]

The study also described longer-term shape changes of the ear canal after surgery, with 3 ears (4.92%) developing a wide canal cavity and one of these also having narrowing at the opening. One patient developed mild facial weakness triggered by an acute outer-ear infection and recovered after revision canal surgery, with satisfactory canal shape reported afterward. [8]

This evidence is most relevant for people told their cholesteatoma is located in the external ear canal, since the study addresses that specific type and how surgeons may plan the operation based on imaging stage. Because the report comes from one center and is retrospective, it cannot show that staging is better than other planning methods, but it does give a clearer picture of expected issues that may sometimes need follow-up treatment or revision surgery. What remains uncertain is how well these results generalize to other hospitals, surgeons, and patient groups.

Across these studies, recurrence and prevention were tied less to a single factor and more to where disease tends to persist and how completely difficult spaces are addressed during surgery. Evidence here comes from retrospective, single-center studies, so the patterns are useful for understanding risk but cannot prove cause and effect. [9] [10]

In a 14-patient series of people who developed recurrent cholesteatoma after canal wall up or canal wall down surgery with obliteration, recurrence most often involved the epitympanum (the upper part of the middle ear). All preoperative CT scans in these recurrent cases showed incomplete obliteration of the epitympanum, and the recurrent disease sometimes extended into the mesotympanum or mastoid. [9]

A separate 200-patient study of middle ear surgery focused on complications rather than recurrence, but it helps frame overall postoperative risk. Having cholesteatoma, longer operations, revision surgery, and age 60 or older were independent risk factors for experiencing at least one complication, and the authors reported their prediction model had good discrimination. [10]

For patients and caregivers, these results add context on why follow-up after cholesteatoma surgery may remain important even when surgery includes filling (obliteration): recurrent disease in this small series clustered in a specific hard-to-reach area and coincided with incomplete filling seen on scans. Separately, the complication study suggests that people having surgery for cholesteatoma, especially revision operations or longer procedures and older patients, may fall into groups with higher overall postoperative risk, but it does not address which choices reduce recurrence. What remains uncertain is how often these patterns apply across different hospitals, surgeons, and surgical techniques, and whether changing specific surgical steps reliably lowers recurrence rates. [9] [10]

These papers point in the same general direction: cholesteatoma is linked to measurable inflammation-related signals, and some of these signals track with either tissue growth behavior or how advanced the disease looks at surgery. The evidence is still exploratory, with one lab-focused study and one single-center patient study, so it is not yet clear which markers are reliable for routine use. [11] [12]

In a lab study using cholesteatoma tissue and cultured cholesteatoma skin cells (keratinocytes), both PAI-1 and TGF-beta1 were increased in cholesteatoma samples, with higher levels reported in pediatric cases. In the cell model, PAI-1 increased keratinocyte growth and weakened the growth-slowing effect of TGF-beta1, and this was linked to activation of ERK signaling (a cell growth pathway). [11]

In a cross-sectional study of 72 people having surgery for cholesteatoma, several preoperative blood inflammation markers were compared with a standard cholesteatoma staging system. The Systemic Immune-Inflammation Index (SII), a combined measure derived from platelet, neutrophil, and lymphocyte counts, showed the strongest relationship with higher stage, and erythrocyte sedimentation rate (ESR) also increased with stage. Neutrophil-to-lymphocyte ratio (NLR) and C-reactive protein (CRP) trended upward but were not statistically significant in this sample. [12]

For patients and caregivers, this month’s biology research mainly adds background about why cholesteatoma can behave like an inflammatory, fast-growing condition. One study suggests specific tissue signals (like PAI-1) may help drive cholesteatoma cell growth in the lab, while another suggests certain routine blood test combinations (especially SII, and also ESR) may rise in people with more advanced surgical stage. What remains uncertain is whether these markers can consistently predict severity, outcomes, or recurrence across different hospitals and over time, since the clinical study was single-center and cross-sectional and the other study was largely laboratory-based. [11] [12]

These reports describe unusual, very large ear and mastoid lesions that looked concerning because of their size, bone destruction, or extension toward the space around the brain. In all cases, computed tomography (CT) and magnetic resonance imaging (MRI) were central for mapping the problem and narrowing the diagnosis, and in two cases follow-up MRI was used to check results after surgery, but each report is just one person so it is unclear how often these patterns occur. [13] [14] [15]

One 49-year-old woman had an exceptionally large congenital mastoid cholesteatoma extending into the supratentorial intracranial space, with imaging showing a massive lesion and near-complete compression of a major venous sinus. Surgeons removed it through a parietooccipital craniotomy, and MRI at 2 months showed no remaining or returning lesion, although a small epidural hematoma and small subdural collection were noted after surgery and resolved. The authors also noted uncertainty about the exact origin, discussing that a giant intradiploic epidermoid with secondary mastoid extension could not be fully excluded. [13]

Another report described a 51-year-old woman with a giant mastoid cortex osteoma alongside cholesteatoma that was confined to the middle ear and mastoid, even though the eardrum was preserved. High-resolution CT and a diffusion-focused MRI sequence were used before surgery, and a canal-wall-down radical mastoidectomy with osteoma removal led to a dry, stable cavity at 3 months with no residual or recurrent cholesteatoma reported at that time. The authors proposed two separate contributing mechanisms, including ear canal obstruction from the osteoma and eustachian tube dysfunction leading to an attic retraction pocket. [14]

A third case involved a middle ear cholesterol granuloma that had a mushroom-like extension into the middle cranial fossa. After transmastoid removal of the middle ear portion and sealing the bony defect, the intracranial part was observed rather than directly removed, and it disappeared on MRI at 4 months with no recurrence at 21 months and no reported worsening of hearing or neurologic problems. The intracranial portion was not confirmed by tissue testing in this report. [15]

For patients and caregivers, these papers mainly provide background context about rare, complex presentations that can look alarming on scans because of size or extension toward the skull or intracranial spaces. They show how imaging and tissue testing (when obtained) were used to clarify what the lesion was and how follow-up scans were used to look for remaining disease, but they cannot predict outcomes for most people because each report describes only one patient and follow-up was short in two of the three cases. Uncertainty remains about how often these situations happen, which symptoms best signal them, and how well results hold up over longer periods. [13] [14] [15]

Across this research, the main pattern was that different surgical approaches to place a bone conduction hearing implant can lead to similar longer-term results, even if the operation itself is shorter with one technique. The evidence here comes from a retrospective (look-back) comparison rather than a randomized trial, so it cannot prove one approach is better. [16]

In 101 people receiving a Ponto bone conduction implant, most often for chronic middle ear disease (chronic otitis media) and including some after cholesteatoma surgery, the punch-only minimally invasive approach had a shorter median operative time (30 minutes) than the linear incision approach (40 minutes). No major complications during surgery were reported in either group. [16]

Skin reactions around the implant site were numerically more common with the linear incision approach (18.2%) than with the punch-only approach (8.7%), but this difference was not statistically significant. Over follow-up up to 48 months, implant losses and revision-free survival were similar overall, with one implant loss in the linear incision group linked to cholesteatoma recurrence. [16]

For patients and caregivers, this study mainly helps set expectations that bone hearing implant placement can be done with different skin approaches, and in this one-center experience the less invasive approach was faster while longer-term implant outcomes looked broadly similar. Because the comparison was not randomized and the groups were treated in different time periods, it remains uncertain whether the differences seen (including skin reactions or long-term survival) would hold in other hospitals or specifically for people after cholesteatoma surgery. [16]