Obliteration of the Epitympanum and Mastoid Results in Reduced Cholesteatoma Recidivism: Long-term Outcomes and Technique

Abstract

Objective:

To compare long-term rates of cholesteatoma recidivism between cases undergoing tympanomastoidectomy with obliteration and without obliteration.

Study design:

Retrospective case-control.

Setting:

Tertiary referral center.

Patients:

Adults and older children with pars flaccida cholesteatoma.

Intervention:

Tympanomastoidectomy with or without obliteration of the epitympanum and mastoid.

Main outcome measure:

Recidivistic cholesteatoma (ie, residual and/or recurrent cholesteatoma) determined by otoscopy, revision surgery, or imaging.

Methods:

Retrospective chart review of pars flaccida cholesteatoma surgeries by a single surgeon between 2015 and 2023. Demographic, clinical, and surgical information, inclusive of cholesteatoma type and EAONO/JOS staging.

Results:

There were 60 cases with obliteration and 63 cases without obliteration that were similar in age, sex, hearing, and cholesteatoma type and stage. Cases with obliteration had statistically significantly lower rates of recurrent (5% vs. 17%, P=0.0447) and residual (3.3% vs. 16%, P=0.0303) cholesteatoma compared with those without obliteration. The overall rate of recidivistic disease was 23.8% in cases without obliteration and 6.8% in cases with obliteration (P=0.0116). Using Kaplan-Meier analysis, recidivistic cholesteatoma–free survival rates at 5 years for the obliteration and non-obliteration groups were 91% and 63%, respectively (P=0.0367, log-rank test). No other clinical or surgical factors influenced recidivism.

Conclusions:

Epitympanum and mastoid obliteration during CWU and CWD tympanomastoidectomy for pars flaccida cholesteatoma is associated with reduced rates of recurrent and residual disease when compared with cases without obliteration.

Introduction

The management of cholesteatoma primarily involves surgical intervention, with canal-wall-up (CWU) and canal-wall-down (CWD) tympanomastoidectomy being the principal approaches. Despite surgical removal, cholesteatoma is notorious for its high rates of recurrence and residual disease, which can be as high as 60% in some series. The wide range of reported outcomes reflects variations in follow-up duration and detection methods among studies.^1–3 To mitigate recidivism following CWU and CWD procedures, the mastoid obliteration technique has gained popularity. Two recent systematic reviews of retrospective case series have reported promising results regarding the efficacy of mastoid obliteration in reducing recurrent and residual cholesteatoma rates. In a review by van der Toom and colleagues of 1534 pooled patients who underwent CWU and CWD with mastoid obliteration, rates of recurrent and residual disease were found to be 4.6% and 5.4%, respectively. In cases of CWU tympanoplasty with mastoid obliteration, these rates were even lower, at 0.28% and 4.2%, respectively.⁴ Ille’s and colleagues included 11 studies in their systematic review, with 8 studies in the quantitative analysis, encompassing a total of 1847 patients. Their pooled data indicated a reduced odds ratio (OR) for recurrent and residual cholesteatoma in obliterated cases. In a subgroup analysis comparing CWU without obliteration to CWU with obliteration, which included 165 cases from 3 series, the OR was 0.48, favoring obliteration.⁵ Both systematic reviews highlighted significant variability among the included studies, underscoring the necessity for further research and standardization in data collection, surgical indications, and outcome reporting.

The primary aim of this study is to compare long-term rates of residual and recurrent cholesteatoma after CWU and CWD tympanomastoidectomy between cases undergoing obliteration of the epitympanum and mastoid versus those that did not undergo obliteration. The secondary goal is to evaluate demographic, clinical, and surgical technique variables as predictors for recurrent and residual cholesteatoma.

Methods

This study includes consecutive cases of middle ear cholesteatoma undergoing primary and revision surgery by a single surgeon at our institution between 2015 and 2023. The technique of epitympanum and mastoid obliteration was introduced in our center in 2018 and performed routinely when appropriate afterward, allowing us to compare earlier cases without obliteration and later cases undergoing obliteration. Inclusion criteria were (1) acquired cholesteatoma originating solely from the pars flaccida, or from both pars flaccida and pars tensa retraction; (2) adult and pediatric patients aged at least 5 years; and (3) postoperative follow-up of at least 12 months. Exclusion criteria included (1) congenital cholesteatoma and nonretraction cholesteatoma of the pars tensa; (2) children under 5 years of age; (3) incomplete clinical, surgical, or postoperative data; and (4) in patients with bilateral cholesteatomas, one randomly selected ear was excluded. At our institution, relevant data are collected uniformly before and during surgery using an electronic database that includes demographic information, clinical presentation, cholesteatoma type and staging at presentation, audiometry, surgery type, and surgical findings. This study was approved by our Internal Review Board (IRB# 20091049).

Audiometry

The average of the 4-frequency pure-tone threshold (PTA) values for air conduction (AC) and bone conduction (BC) at 0.5, 1, 2, and 3 kHz (or 4 kHz when 3 kHz was unavailable) was obtained from the patients’ electronic medical records. For each patient, the air-bone gap (ABG) was calculated by subtracting the BC PTA from the AC PTA at each interval. ABG change was calculated by subtracting the last postoperative ABG from the preoperative ABG: a result above 0 indicates an improvement in ABG, while a result below 0 indicates a deterioration of the ABG.

Disease control

The primary outcome was disease recidivism, defined as either residual and/or recurrent cholesteatoma. Residual cholesteatoma (ie, incomplete removal) was defined as the presence of cholesteatoma in the middle ear or mastoid cavity identified by second-look surgery and/or imaging [ie, MRI with diffusion-weighted imaging (DWI-MRI)], and in cases of CWD without obliteration, by direct inspection (ie, otoscopy). Recurrent cholesteatoma was defined as the reformation of a retraction pocket with the accumulation of keratin debris in the epitympanum or mesotympanum. We investigated demographic, clinical, and surgical factors that might be associated with cholesteatoma recidivism, including sex, age at surgery, side, race and ethnicity, EAONO/JOS cholesteatoma type and stage,⁶ PTA/ABG, and the degree of mastoid development as assessed by computed tomography. Surgical variables included the type of surgery (CWD or CWU tympanomastoidectomy), surgical findings (eg, number of affected sites in the temporal bone, including ossicular erosion), and whether epitympanic and mastoid obliteration were performed.

Postoperative surveillance and outcome

Postoperative surveillance was individualized and typically included a combination of oto-microscopy, imaging, second-look surgery, and audiometry. All patients undergoing CWU tympanomastoidectomy without obliteration had a second-look surgery between 9 and 13 months after the primary surgery. In CWD without obliteration, postoperative surveillance for recidivism was done by oto-microscopy. All patients undergoing CWU or CWD tympanomastoidectomy with obliteration had either a second-look surgery and/or DWI-MRI study. To increase the sensitivity of detecting recidivistic cholesteatoma, the MRI was obtained at least 16 months after the primary surgery.⁷ Follow-up time was defined as the period from the primary procedure to the date of identification of recidivistic disease or to the last follow-up date for cases with no residual/recurrent cholesteatoma.

Surgery

Surgeries were primarily performed using oto-microscopy and were adapted based on the location and extent of the disease. Surgical approaches included attico-antrostomy, CWU, or CWD tympanomastoidectomy.⁸ CWU typically included epitympanectomy with removal of the diseased incus and malleus head, along with a posterior tympanotomy. Endoscopes were employed to explore and remove disease from difficult-to-visualize areas, such as the supratubal recess and the sinus tympani. Defects in the attic or ear canal wall were always reconstructed using cartilage grafts. In CWD, the posterior bony external auditory canal was removed down to the fallopian canal, and the mastoid tip was excised when it was extensively pneumatized. In addition, CWD procedures were often combined with meatoplasty. Since the introduction of the obliteration technique in our center in 2018, the epitympanum and mastoid cavity were obliterated in cases of CWU, using either drilled cortical bone dust (ie, bone pate) or active bioglass (Bonalive, Bonalive Inc.). The filler material was compacted to fill the air spaces of the epitympanum and mastoid completely. A palisade of cartilage grafts was created to block the epitympanic space and the posterior tympanotomy, preventing the spillage of obliteration material into the mesotympanum. In CWD, the obliteration material was used to reduce the size of the cavity, covered with a pedicled flap with axial blood supply from the temporoparietal fascia or periosteum to prevent spillage and facilitate healing. All cases were performed under general anesthesia on an outpatient basis.

Statistical analyses

The JMP Pro 18.0.1 statistical software package (SAS Institute Inc.) was utilized for statistical analyses. Continuous variables were expressed as means (with SD), medians (with interquartile ranges), and analyzed using either the Student t test or the Wilcoxon-Mann-Whitney test, as appropriate. Categorical variables were expressed as frequencies (percentages) and analyzed using the χ² test or the Fisher exact test. Comparisons were made between cases undergoing obliteration and those without obliteration. A P value of <0.05 was considered statistically significant. Survival curves were constructed using the Kaplan-Meier method and compared using the log-rank test.

Results

A total of 262 consecutive surgeries for cholesteatoma were reviewed. Of these, 123 cases were included for analysis. The exclusion of 139 patients was due to the following reasons: nonretraction cholesteatoma (73 cases), congenital cholesteatoma, and cases in children under 5 years where the distinction between congenital and acquired cholesteatoma could not be made (19 cases), obliteration cases without postoperative imaging or second-look surgery (18 cases), incomplete surgical or postoperative data (21 cases), and contralateral ear in bilateral cases (8 cases).

The table provides demographic, clinical, and surgical data for the total cohort of 123 patients, stratified by obliteration status. There was no observed gender or side predilection. The median (range) and mean (SD) age of the cohort were 41 (5 to 80) and 40.95 (SD, 19.7) years, respectively. Among the cohort, 114 (92.7%) were primary surgeries, and 9 (7.3%) were revision surgeries. The median (range) and mean (SD) follow-up duration for the total cohort were 34 (12 to 216) and 42.01 (30.5) months, respectively. Recidivistic disease (ie, residual and recurrent cholesteatoma) was noted in 19 out of 123 patients (15.5%), with recurrent cholesteatoma in 14 (11.38%) cases and residual cholesteatoma in 12 (9.75%) cases; 6 patients had both residual and recurrent disease simultaneously: 5 in the non-obliteration group and 1 in the obliteration group.

Epitympanum and mastoid obliteration were performed in 60 cases, while obliteration was not performed in 63 cases. The “obliteration” and “no-obliteration” groups did not differ in terms of demographic and clinical data, follow-up, or audiometry results, including cholesteatoma staging. The table presents outcome comparisons based on obliteration status. Cases with obliteration had statistically significantly lower rates of recurrent (5% vs. 17%, P=0.0447) and residual (3.3% vs. 16%, P=0.0303) cholesteatoma compared with those without obliteration. The overall rate of recidivistic disease was 23.8% in cases without obliteration and 6.8% in cases with obliteration (P=0.0116).

Using Kaplan-Meier analysis (Fig. 1), recidivistic cholesteatoma–free survival rates at 5 years for the obliteration and non-obliteration groups were 91% and 63%, respectively (P=0.0367, log-rank test). In univariate analyses, factors such as sex, age at surgery, side, race, ethnicity, type of surgery (ie, primary versus revision), length of follow-up, canal wall status (ie, CWU versus CWD), mastoid pneumatization (ie, extensive, partial, and sclerotic), and middle ear mucosa status (ie, normal, edema, and granulation) did not significantly affect the rate of recidivism. We also analyzed the rate of recidivistic disease based on ossicular involvement (ie, none, malleus-incus only, and malleus-incus-stapes) and the number of involved sites in the temporal bone (ie, 1 through 5), and these variables did not impact recidivism.

Figure 1.

Kaplan-Meier survival plot of recidivistic cholesteatoma stratified by obliteration using the log-rank test.

A subgroup analysis of CWU and CWD cases was conducted. CWD procedures were performed in 27 patients: 10 without obliteration and 17 with obliteration. Among these, only 1 case of recidivistic disease (recurrent cholesteatoma in the middle ear) occurred in the obliteration group. No statistically significant differences were observed between the obliteration and non-obliteration groups regarding rates of recurrent, residual, or recidivistic disease.

CWU procedures were performed in 96 patients: 43 with obliteration and 53 without. Recurrent cholesteatoma was diagnosed in 11 patients (21%) without obliteration and in 2 patients (5%) with obliteration (P=0.0334). Residual disease was identified in 10 patients (19%) without obliteration and in 2 patients (5%) with obliteration (P=0.059). Recidivistic disease was significantly more common in the non-obliteration group (28%, 15 cases) compared with the obliteration group (7%, 3 cases) (P=0.0087).

Complications

Postoperative prolonged (>2 mo) or recurrent otorrhea was recorded in 7 patients: 4 in the obliteration group and 3 in the non-obliteration group. All these cases were managed with local debridement and topical or systemic antibiotics, and none required revision surgery. Two of the 4 cases in the “obliteration” group underwent CWD tympanomastoidectomy, where the bone pate used for obliteration was covered only with a free graft of temporalis fascia. Spillage of the bone pate occurred, resulting in granulations and partial loss of the mastoid cavity reconstruction. These 2 cases occurred early in the series, prompting the adoption of an axial pedicled vascularized flap to cover the obliteration material in the mastoid cavity. The other 2 prolonged otorrhea cases in obliterated cases were in CWU surgery, characterized by persistent granulations of the reconstructed tympanum. One of the “non-obliterated” CWD cases corresponded to an infection of the meatoplasty incision with chondritis, which required systemic antibiotics and local wound care. There were no other complications.

Hearing outcomes

All cases included in this series had a postoperative audiogram obtained at least 1 year after surgery. No significant differences were found in PTA thresholds, ABG, and ABG change between the 2 groups (Table 1). One patient had preoperative nonfunctional hearing (≥65 dB BC PTA) and was excluded from hearing outcome analyses. No cases exhibited a change in BC PTA >10 dB. The median ABG change was only 1.875 dB, with a range of (−10) to 21.25 dB, and several patients experienced worse hearing despite disease control. Figure 2 shows postoperative ABG stratified by bins.

Table 1.

Demographic, clinical, surgical, and audiometric data of 123 patients undergoing surgery for pars flaccida cholesteatoma, and by obliteration status

Variable	Total	No-Obliteration	Obliteration	P
N	123	63	60
Male:Female	69:54	38:25	31:29	0.3672
Ethnicity
NH:Ha	69:54	37:26	32:28	0.5886
Race, n (%)
White	94 (76.4)	48 (76.2)	46 (77)	0.9432
Black	14 (11.4)	7 (11.1)	7 (12)
Native American, n (%)	2 (1.6)	1 (1.6)	1 (1.7)
Asian, n (%)	1 (0.8)	1 (1.6)	0 (0)
More than 1 race, n (%)	8 (6.4)	4 (6.3)	4 (6.7)
Not reported, n (%)	4 (3.2)	2 (3.2)	2 (3.3)
Age (y)
Mean (SD)	40.95 (19.7)	39.14 (20.6)	41.8 (19.5)	0.2957
Median	41	37	42.5
IQR	26-58	22-58	28-58
Range	5-80	5-78	5-80
Side (Right:Left)	65:58	31:32	34:26	0.4713
Surgery type
Primary:Revision	114:9	59:4	55:5	0.7395
	92.7%/7.3%	93.6%/6.4%	91.7%/8.3%
CWU:CWD	96:27	53:10	43:17	0.1273
	78%/22%	84%/16%	72%/28%
Follow-up (mo)
Mean (SD)	42.01 (30.5)	45.13 (38.7)	38.75 (18.3)	0.8178
Median	34	34	33.5
IQR	24-53	19-61	24.5-50.25
Range	12-216	12-216	13-69
EAONO/JOS stage, n (%)
I	8 (6.5)	5 (8)	3 (5)	0.3024
II	99 (80.5)	53 (84)	47 (78)
III	16 (13)	5 (8)	10 (17)
IV	0	0	0
Mastoid pneumatization, n (%)
Sclerotic	27 (22.5)	16 (26)	11 (18.3)	0.3380
Partial	57 (45)	31 (49)	26 (43.3)
Extensive	26 (21.7)	12 (19)	14 (23.3)
Not reported	13 (10.8)	4 (6)	9 (15)
Preoperative ABG (dB)b
Mean (SD)	19.96 (10.3)	20.03 (9.2)	19.88 (11.4)	0.9371
Median	21.25	20	21.25
IQR	10-28.75	12.5-27.5	8.375-28.75
Range	3.75-45	3.75-37.5	3.75-45
Postoperative ABG (dB)b
Mean (SD)	15.1 (7.2)	15.38 (7.1)	14.77 (7.4)	0.6498
Median	15	15	14.38
IQR	10-21.25	10-22.5	7.81-21.25
Range	0-35	2-27.5	0-35
ABG change (dB)b
Mean (SD)	NA	NA	NA
Median	2.5	2.5	1.25	0.2416
IQR	0-8.75	0-8.75	0-11.25
Range	(−5) to 45	(−5) to 45	(−5) to 21.25
Recidivism rate, n (%)
Recurrent	14 (11.38)	11 (17)	3 (5)	0.0447
Residual	12 (9.75)	10 (16)	2 (3.3)	0.0303
Rec+Resc	19 (15.5)	15 (23.8)	4 (6.8)	0.0116

^a NH:H=Nonhispanic:Hispanic.

^b ABG or air-bone gap=Average air-conduction hearing threshold minus average bone conduction hearing thresholds (for 0.5, 1, 2, and 3 kHz).

^c Number of patients with either recurrent and/or residual disease.

Figure 2.

A, Axial high-resolution computed tomography of a patient who underwent CWU with bone pate obliteration 12 months prior, showing failure of the obliteration material to create neo-ossification of the mastoid cavity. B, DWI-MRI obtained 22 months after the primary surgery of the same patient, showing a lesion with reduced diffusivity in DWI-MRI in the mastoid cavity. This lesion was confirmed to be a residual cholesteatoma by revision surgery.

Discussion

Two recent systematic reviews demonstrated reduced disease recidivism for tympanomastoidectomy with epitympanum and mastoid obliteration compared with tympanomastoidectomy without obliteration.^4,5 However, both reviews underscored the significant heterogeneity of the reviewed series, raising concerns about potential bias. Most reviewed series were retrospective observational studies and contained several confounding factors. For instance, in most series, the type and stage (ie, extent) of cholesteatoma were not specified, the indications for obliteration were not stated, the duration of follow-up varied widely, and the methods for detecting recidivistic disease were not uniform. Furthermore, many pooled series originated from geographical areas (ie, Europe) where obliteration techniques are prevalent. As such, it has been suggested that more prospective trials of matched groups are necessary to establish the true benefits of obliteration over non-obliteration techniques.⁵ While this aim is commendable, challenges to randomization in clinical practice exist. We attempted to minimize bias by uniformly collecting preoperative and postoperative data, and we focused exclusively on retraction cholesteatomas affecting the pars flaccida in this series, as obliteration techniques are more likely to be practiced and beneficial for this cholesteatoma type. Our data indicate that obliteration of the epitympanum and mastoid following CWU and CWD tympanomastoidectomy reduces the rate of cholesteatoma recidivism compared with cases without obliteration; obliteration cases exhibited statistically significantly lower rates of recurrent (5% vs. 17%) and residual (3.3% vs. 16%) cholesteatoma. The combined rate of recidivistic disease was 23.8% in cases without obliteration and 6.8% in cases with obliteration. The median postoperative observation time was 34 months for both groups, with a minimum of 12 months. Survival analysis revealed that recidivistic cholesteatoma rates at 5 years of follow-up were 9% in the obliteration group and 37% in the non-obliteration group (P=0.0367, log-rank test). Our findings align favorably with those from the systematic reviews and a recent series by Erfurt et al⁹ involving 143 adult patients undergoing CWU tympanomastoidectomy with and without obliteration, where the group treated with mastoid obliteration (73 ears) exhibited significantly lower rates of recurrent (4.1%) and residual (6.8%) cholesteatoma compared with the group without obliteration (70 ears, 25.7% and 20%).

None of the other demographic and clinical variables studied influenced recidivism. The obliteration and non-obliteration groups were matched in terms of age, sex, type of surgery, ear canal status (ie, CWU and CWD), degree of mastoid pneumatization, middle ear mucosal status, EAONO/JOS cholesteatoma stage, preoperative PTA, and length of follow-up. Most cases in both groups were classified as stage II (~80% of cases in the entire cohort), with fewer cases in stages I and III. Stages I and II represent local cholesteatoma extension into the middle ear and mastoid. According to the STAM system of middle ear sites, stage I involves only 1 STAM site, while stage II involves multiple sites. Stage III indicates extracranial extension, and stage IV denotes intracranial extension and complications. While this staging system is widely accepted, criticisms regarding its prognostic value have been raised.^10–12

One significant limitation of this staging system is the heterogeneity of disease involvement in stage II. The STAMCO classification system attempts to address this limitation by refining EAONO/JOS stage II cases based on the number of involved sites and ossicular chain status.¹³ The grading of cholesteatoma extent is further categorized: STAM 1 (stage I) for involvement in only 1 location, STAM 2 (stage II) for involvement in 2 locations, and STAM 3 (stage III) for disease extension into 3 locations or when one of the areas of difficult access (ie, S1 and S2) is involved. Three categories of ossicular chain status are also incorporated. Interestingly, in the recent series by Erfurt et al,⁹ the greatest benefit of obliteration concerning disease control was reported in a subgroup analysis of STAMCO stage III cases, showing recidivism rates of 54.5% in non-obliteration cases versus 7.7% in obliterated cases. Another proposed classification system, ChOLE, is a numerical system that also considers the number of involved sites and ossicular chain status while grading middle ear ventilation based on mastoid bone pneumatization observed in CT scans.¹⁴ We attempted to separate EAONO/JOS stage II cases based on ossicular involvement, mastoid pneumatization, and the number of involved sites. While we staged our cases using STAMCO and ChOLE systems, we were unable to demonstrate an effect of these classifications on recidivism. Although it is reasonable to argue that a more granular definition of cholesteatoma extent and ossicular chain status may add prognostic value, we suspect that our series lacked sufficient size to achieve the statistical power required by the increased number of disease categories in the STAMCO and ChOLE systems. We chose to report our data using the EAONO/JOS classification system due to its widespread consensus and implementation globally.

Obliteration techniques aim to prevent the reformation of a retraction cholesteatoma by separating the middle ear from the epitympanum and mastoid, thereby reducing recurrent disease. Another proposed mechanism for the associated lower recurrence rates compared with non-obliteration techniques is that mastoid obliteration diminishes variations in middle ear gas pressure and tympanic membrane retractions by decreasing the available mucosal surface for gas exchange.¹⁵ However, the mechanisms by which obliteration reduces residual disease remain a topic of debate. It has been speculated, based on animal studies, that obliteration interferes with the conditions necessary for residual keratinocytes to develop into a growing cholesteatoma pearl, thus creating an unfavorable environment for residual cholesteatoma growth.¹⁶ Furthermore, improved intraoperative visualization of disease, which may result in better cholesteatoma resection, occurs when extending the atticotomy or converting from CWU to CWD surgery, as surgeons may feel more comfortable doing so knowing that the external auditory canal will be effectively reconstructed with obliteration. Additional advantages of obliteration include the reduction of the size of the mastoid cavity in CWD, facilitating postoperative care, and preventing the unsightly retroauricular soft tissue depression that often follows a mastoidectomy procedure. The disadvantages of epitympanum and mastoid obliteration include the potential for infection and extrusion of the obliteration material, longer operative times, and the necessity for MRI scanning for postoperative surveillance due to the risk of “silent” residual cholesteatoma in the obliterated spaces. Moreover, epitympanic obliteration can only be performed when the ossicular chain has been removed or is absent, limiting this technique to more advanced cholesteatoma cases. Furthermore, epitympanum obliteration is likely not beneficial for cases of pars tensa retraction cholesteatoma and adhesive otitis media with disease confined to the tympanic cavity.

Various materials and surgical approaches are utilized, each with its advantages and limitations.^17,18 Obliteration using autologous bone pate collected during cortical mastoidectomy is a popular technique, though autologous cartilage, soft tissue, hydroxyapatite cement, and active bioglass^9,18 have also been utilized as fillers. Both bone pate and bioglass were used in this series; however, bioglass was utilized in only 8 cases. While a comparison between these filler materials would be valuable, the limited number of bioglass cases precludes drawing meaningful conclusions. Off-the-shelf materials such as active bioglass are particularly practical in situations where cortical bone is unavailable for harvesting, as in certain revision cases or when native bone is infected. In CWD, we now routinely employ pedicled flaps of temporoparietal fascia or periosteum with axial blood supply to cover the filler material.^19,20 These flaps bring vascularized tissue to facilitate rapid re-epithelization, infection control, and prevent the extrusion of filler materials. As experienced early in our series, prolonged otorrhea occurred in 2 cases where the bone pate used for obliteration was only covered with a free graft of temporalis fascia, resulting in granulations and partial loss of the mastoid cavity reconstruction.

Over 90% of cases in our series were primary surgeries, while revision surgeries corresponded to recidivistic cholesteatoma after unsuccessful primary CWU or CWD tympanomastoidectomy performed at our institution, where obliteration had not been implemented. Revision surgery was either a planned second-stage tympanomastoidectomy or indicated after recidivistic disease was identified through otoscopy or imaging. The choice of tympanomastoidectomy type is individualized and varies according to the surgeon’s preference. At our institution, we strive to preserve the external auditory canal whenever possible and recommend CWD techniques in cases where there is an external auditory canal defect that is too large to reconstruct or in instances of irresectable cholesteatoma (ie, cholesteatoma adherent to vital structures such as the facial nerve, membranous labyrinth, dura, or sigmoid sinus). The obvious disadvantages of removing the external auditory canal in CWD procedures include generally poorer postoperative hearing outcomes compared with CWU and the creation of an open mastoid cavity that is not self-cleaning, necessitating lifelong regular debridement, a concern that is particularly pronounced in young children.²¹ One argument in favor of CWD procedures is the lower recurrence of retraction cholesteatoma and the avoidance of second-look surgery. However, a recent systematic review indicated that the rates of recurrent or residual disease necessitating mastoid revision surgery may not differ between CWU and CWD procedures.²²

We performed a subgroup analysis of CWD and CWU cases. Among the 27 CWD cases, the overall rate of recidivistic disease was low at 4%, with no statistically significant difference between the obliterated and non-obliterated groups. The benefit of obliteration was most evident in the CWU group, with significantly lower recurrence and total recidivistic rates in obliteration cases. However, with the implementation of obliteration techniques for canal wall reconstructions, the lines between CWD and CWU techniques are less distinct, as the surgeon may feel more comfortable extending the exposure at the expense of the canal wall to facilitate disease removal while at the same time avoiding the creation of an open mastoid cavity. The primary aim of our study was to assess the effectiveness of obliteration as a technique across different surgical approaches. In this context, obliteration is viewed as an independent factor to either CWU or CWD, rather than focusing on the underlying differences between the procedures themselves. By grouping both, the analysis gains statistical power to detect general trends in how obliteration affects outcomes (eg, recidivism and hearing function) regardless of surgical technique. In clinical practice, the decision to perform CWU versus CWD is driven by disease extent, anatomy, patient factors, and surgeon preference. Obliteration has been increasingly used across both types to improve outcomes (eg, reduce cavity problems in CWD, or promote healing in CWU with extensive disease). Therefore, studying them together may reflect current real-world usage of obliteration and provide more generalizable insights.

Recurrence of disease is typically identified through the observation of reformation of a retraction pocket via oto-microscopy. Detection of residual disease after CWU tympanomastoidectomy has traditionally been performed through planned revision surgery a few months following the primary surgery (ie, second-look surgery). DWI-MRI has emerged as a substitute for second-look surgery and gained popularity alongside the advent of obliteration techniques due to the inherent surgical challenges of exploring the obliterated spaces of the epitympanum and mastoid. In this series, 2 cases of residual disease were identified after CWU with obliteration. In 1 case, residual disease was found during revision surgery for a recurrent epitympanic retraction cholesteatoma, while separate residual disease was observed in the mesotympanum. In the second case, DWI-MRI indicated a lesion with restricted diffusion 22 months after the primary surgery (Fig. 2). A pre-MRI CT scan had shown failure of ossification and loss of the obliteration material. We routinely fill the entire volume of the epitympanum and mastoid during CWU with compacted obliteration material, and typically, new bone or bioglass creates a solid image on postoperative CT scans (Fig. 3). Radiographic evidence of failed neo-ossification or the presence of large lacunae within the obliterated spaces should raise suspicion for granulation tissue or residual keratinous rests, warranting differential diagnosis via DWI-MRI. Cholesteatomas appear nonenhancing or show only rim-enhancement in contrast-enhanced T1-weighted spin-echo imaging, allowing differentiation from granulation tissue. Cholesteatomas are hyperintense on DWI, whereas granulation tissue, fibrous tissue, cholesterol granulomas, and serous fluid exhibit low signal intensity on DWI. The observed high signal intensity is attributed to the restricted molecular diffusion of cholesteatoma.²³

Figure 3.

Temporal bone computed tomography of CWU tympanomastoidectomy with obliteration with (A) bone pate and with (B) bioglass.

Due to susceptibility artifacts and the low spatial resolution associated with thick sections of echo-planar DWI, false-negative scans can occur in lesions smaller than 6 mm.^7,23 The rate of false-negative DWI-MRI decreases with increasing cholesteatoma size; therefore, it is recommended to delay imaging for 18 months postsurgery to allow any epithelial rest in the mastoid to grow to a detectable size. It has also been proposed that a second DWI-MRI be obtained 5 years after obliteration surgery to identify any residual disease that may have been missed in the initial postoperative DWI-MRI.⁹ Non–echo-planar DWI and newer DWI modalities have demonstrated greater sensitivity and positive predictive value than echo-planar DWI, and when negative for cholesteatoma, these modalities may help avoid a second-look surgery.²⁴

Hearing outcomes

Most cases in this series were classified as stage II disease, where the incus and malleus were typically involved or removed to facilitate disease extirpation. Ossicular reconstruction with partial ossicular replacements (PORPs) and total ossicular replacements (TORPs) was performed either secondarily during the second-look surgery for cases without obliteration or primarily in cases with obliteration. The median ABG change was only 1.875 dB, and some patients experienced worse hearing despite disease control, highlighting the challenges of middle ear reconstruction in cholesteatoma surgery. Revision surgery was performed in 11 (18%) of the obliteration cases, primarily for reconstruction of the middle ear in cases with a large ABG, not as a planned second-look, except for the 4 cases with recidivistic disease. The decision to pursue revision surgery for middle ear reconstruction versus habilitation with a hearing aid or bone conduction auditory implant is typically a shared decision between the surgeon and the patient and varies individually.

Limitations

The limitations of this study are inherent to retrospective reviews, including selection bias, confounding, and attrition bias. Although we rigorously and uniformly recorded cholesteatoma type, location, and extent both preoperatively and intraoperatively, the decision to perform obliteration and the timing of postoperative visits, MRI, and/or second-look surgeries by the surgeon may have influenced both the surgical approach and the identification of recidivistic disease. Regarding attrition, 39 cases of pars flaccida cholesteatoma were excluded from analyses due to loss to follow-up or lack of postoperative DWI-MRI, potentially resulting in an undercount of recidivistic disease. Twenty of these 39 cases were treated without obliteration, indicating that undercounting could have affected either group, which may not have altered the reported inter-group differences. Furthermore, although none of the 19 obliterated cases excluded for lacking a postoperative MRI showed recurrent disease on oto-microscopy, their follow-up duration was <12 months. A significant limitation of this series is that detecting residual disease in obliterated cases relied on DWI-MRI scans, whereas cases without obliteration typically underwent second-look surgeries for exploration of the cholesteatoma-involved areas. Most DWI-MRIs obtained in this series were echo-planar DWI, which has a lesser negative predictive value than non–echo-planar DWI and other newer modalities. We attempted to mitigate this by delaying the first MRI until at least 16 months postoperatively and by introducing non–echo-planar DWI scanning during the latter part of the study period. Nevertheless, cholesteatoma recidivism after 5 years has been reported,²⁵ and “silent” disease in the mastoid cavity could potentially occur even after a “negative” second-look and/or DWI-MRI performed within the first 18 months postsurgery, highlighting the need for prolonged active surveillance. Another limitation is that, given the relatively low number of CWD cases, the subgroup analysis failed to provide more granular comparisons and insights into the interaction between surgical approach and obliteration efficacy. In addition, the purported benefit of obliteration as reported here should not be extrapolated to cases with significant disease extension into the mesotympanum.

Conclusions

Epitympanum and mastoid obliteration during CWU and CWD tympanomastoidectomy for cholesteatomas originating from retractions of the pars flaccida is both safe and effective in reducing rates of recurrent and residual disease when compared with CWU and CWD procedures without obliteration. Observed complications associated with the obliteration technique were minor and potentially preventable with refinements in the surgical technique aimed at reducing postoperative infection and filler extrusion. Hearing outcomes did not show significant improvement with obliteration compared with non-obliteration, and hearing gains were modest, underscoring the necessity for long-term surveillance to identify “late” recidivistic disease and the ongoing need for habilitation of hearing function.