1. Introduction
Cholesteatomas are considered epidermal inclusion cysts that originate within the temporal bone, consisting of squamous epithelial cells and related debris. These growths tend to enlarge progressively, leading to tissue destruction, frequently accompanied by inflammation and granulation tissue formation [
1].
Cholesteatomas are typically categorized into three groups: congenital, primary acquired, and secondary acquired. Hypotheses suggest that keratinized squamous epithelial tissue in the middle ear cleft gives rise to congenital cholesteatomas. The primary location of RP often occurs in the pars flaccida, a region commonly associated with the most frequently diagnosed syndrome among symptomatic patients. Eustachian tube dysfunction (ETD) and gradual thinning of the tympanic membrane frequently coincide with RP, alongside recurrent occurrences of otitis media with effusion [
2]. Secondary acquired cholesteatomas, on the other hand, result from tympanic membrane perforations with epithelial migration [
3]. Several factors, including tobacco use, previous infections, and genetic predispositions, may also contribute to the development of cholesteatomas [
4,
5].
Cholesteatomas possess the capacity to induce recurrent infections and osseous erosion, which can affect diverse structures within the temporal bone. Consequently, this can lead to auditory impairment, facial nerve compromise, vestibular dysfunction, and potentially intracranial sequelae [
1].
Various treatment options for retraction pockets are available, ranging from conservative approaches to surgical interventions [
6]. For this condition, the insertion of ventilation tubes, excision of RP with tympanic membrane reconstruction, and tympanoplasty with canal wall-up or canal wall-down procedures can be performed depending upon the case. The lack of consensus stems from the challenge of predicting which instances of RP will progress into cholesteatoma and which will remain stable and symptom-free [
2].
Canal wall-down procedures are preferred for complete cholesteatoma removal, while canal wall-up procedures preserve the middle ear’s anatomy and offer better reconstruction options [
7].
One of the significant challenges associated with cholesteatomas is their tendency to recur. Consequently, close follow-up of surgically treated patients is necessary [
8]. Early diagnosis of cholesteatoma recidivism is crucial in preventing complications [
9]. Traditionally, a second surgery was scheduled 6 months to 1 year after the initial surgery to check for recurrence. However, there was uncertainty about when and how to perform this second procedure. Cholesteatoma recurrence is often linked to pre-existing chronic conditions or when some of the cholesteatoma remains unintentionally or intentionally due to its proximity to important structures [
10]. Initially, High-resolution Computed Tomography (CT) scans were the preferred way to diagnose and assess the extent of cholesteatoma. However, they had limitations in distinguishing between fibrotic tissue, cholesteatoma, or inflammation. Later, Magnetic Resonance diffusion-weighted imaging with echo-planar sequences (EPI-DWI MRI) was employed for lesions larger than 5 mm. Nevertheless, its effectiveness was limited when dealing with smaller lesions, leading to the generation of several artifacts. The development of Magnetic Resonance diffusion-weighted imaging with non-echo-planar sequences (non-EPI-DWI MRI) marked a significant breakthrough in imaging for detecting small cholesteatomas. This technique may help diagnose cholesteatoma recidivism after the first surgery, potentially avoiding unnecessary revision surgeries and associated risks.
Objectives
The objective of this study is to measure the rate of cholesteatoma recidivism after a specific surgical technique, Attic Exposure-Antrum Exclusion (AE-AE), and the role of non-EPI-DWI MRI sequences for the detection of cholesteatoma recidivism. Additionally, it seeks to investigate different parameters that may contribute to its recurrence.
3. Results
During radiologic follow-up, all 63 (100%) patients underwent non-EPI-DWI MRI to detect the presence of cholesteatoma recidivism. The study population consisted of 51% females and 49% males, with a mean age of 41 (standard deviation: 21) at the time of surgery. Among these patients, 57% had cholesteatoma on the left side, while the remaining 43% were on the right side. The analysis of the parameters in this patient cohort, using non-EPI-DWI MRI as the imaging follow-up technique, yielded the following results (
Table 1).
According to the European Academy of Otology and Neuro-Otology (EAONO) and Japanese Otological Society (JOS) classification for cholesteatoma, we found 47.6% of patients in Stage I (cholesteatoma localized in the attic), 47.6% in Stage II (cholesteatoma involving two or more sites), and 4.8% in Stage III (cholesteatoma with extracranial complications such as facial palsy, labyrinthine fistula, or canal wall destruction). No patient was found with Stage IV (cholesteatoma with intracranial complications).
About the surgical technique used, 67% of patients underwent a simple attic exposure and antrum exclusion (AE-AE) technique, 22% underwent an extended attic exposure and antrum exclusion (AE-AE) technique, and 11% underwent a radical mastoidectomy.
For aditus closure, cartilage was used in 90% of cases; preferably, conchal cartilage was utilized, thus avoiding obtaining it from the tragus, which could compromise its supporting function in a potential need for hearing aid use. Cortical bone was used in in 5%, and a combination of cartilage and cortical bone in 5% of cases.
Regarding reconstruction of the ossicular chain, in 60% of cases, reconstruction was unnecessary because the cholesteatoma was located laterally and did not affect the integrity and mobility of the ossicular chain. In 22% of cases, the stapes suprastructure was preserved, and a partial ossicular chain prosthesis (PORP) was placed. However, in 18% of cases, the stapes suprastructure was absent, and a total ossicular chain prosthesis (TORP) was needed.
Regarding the pure-tone average (PTA), the pre-surgical PTA was 42.2 dB (standard deviation: 15.7) and the post-surgical PTA at 12 months after surgery was 40.9 dB (standard deviation: 15.9). The difference between the 12-month post-surgical PTA and the pre-surgical PTA was −1.3 dB (standard deviation: 13.0). The
p-value was 0.457, as shown in
Table 2.
Statistical analyses using logistic regression showed no statistically significant association between cholesteatoma recidivism and any of the parameters examined in
Table 3.
However, logistic regression analysis revealed a significant association between pre-surgical PTA and cholesteatoma recidivism on non-EPI-DWI MRI scans (p-value: 0.003).
So, in
Table 3, where logistic regression was used, the coefficient values indicate the relative contribution of each independent variable to predict cholesteatoma recidivism. A significant coefficient value suggests that the variable is associated with the probability of recidivism, regardless of other variables in the model. In this context, the value of 0.185 for Pre-Surgical PTA indicates that, after controlling for other variables in the model, an increase in PTA is associated with a change in the probability of recidivism.
On the other hand, when we evaluate the relationship between imaging findings and preoperative audiometry using logistic regression. Here, the p-value of 0.003 indicates a statistically significant association between PTA and imaging findings. This suggests that PTA has an impact on imaging outcomes, which could have significant clinical implications, such as the ability to predict cholesteatoma recidivism.
In summary, even if it may seem contradictory on the PTA results, both analyses may be providing valuable information but are focused on different aspects of the study.
Table 3 examines variables predicting cholesteatoma recidivism, while the significant
p-value of 0.003 explores the relationship between PTA and imaging findings.
The results obtained from the non-EPI-DWI MRI scans revealed that 87% (55 patients) did not show any signs of cholesteatoma recidivism. However, in 13% (8 patients), the imaging results suggested the possibility of cholesteatoma recidivism. Of these 8 patients, one individual underwent a second non-EPI-DWI MRI scan, which yielded negative results, suggesting a false positive in the initial scan. The remaining 7 patients were scheduled for a second look surgery. Unfortunately, 2 cases were lost to follow-up, but the remaining 5 patients underwent the surgical intervention. Among these operated patients, 4 cases were found to have a cholesteatoma recidivism stage II due to multiple locations, which was successfully removed. However, in one of the operated cases, it turned out to be another false positive as no cholesteatoma was found during the intervention.
Additionally, three of these patients required a third revision surgery. All three patients had already undergone the surgical revision. In one case, the positive diffusion MRI result corresponded to a cholesterol granuloma; in the second case, cholesteatoma was found during both revision surgeries, and in the third case, inflammatory tissue was detected. The results are summarized in
Figure 1.
Given the observed results, the values of non-EPI-DWI MRI for detecting cholesteatoma recidivism in terms of sensitivity, specificity, positive predictive value, and negative predictive value were 100% (CI: 100.0–100.0%), 96.4%(CI: 91.5–100.0%), 60% (CI: 47.4–72.6%), and 100% (CI: 100.0–100.0%), respectively. As for the surgical technique of AE-AE, it was successful, with a recidivism rate of 5.2% (CI: 0.0–10.87%).
4. Discussion
Surgical resection stands as the primary treatment approach for cholesteatoma, with a spectrum of surgical techniques available, each tailored to specific considerations. These techniques include tympanotomy, atticotomy, cortical mastoidectomy, and canal wall up or down mastoidectomy, as well as more extensive interventions such as modified–radical and radical mastoidectomy. The overarching aim remains the complete eradication of the cholesteatoma while striving to reduce the potential for recidivism, conserve auditory function, and improve ear hygiene. The choice of surgical approach is contingent upon the cholesteatoma’s extent and the surgeon’s experience. In contrast, atticotomy is preferred when the cholesteatoma is confined to the lateral side of the malleus and incus within the “attic”. Success rates for complete cholesteatoma removal with atticotomy typically fall between 70% and 90%. Cortical mastoidectomy comes into play when the cholesteatoma extends towards the ossicles and penetrates the mastoid region via the antrum. This approach typically yields success rates ranging from 70% to 95%. The selection between canal wall-up and canal wall-down procedures depends on the attainability of complete cholesteatoma elimination. Opting for canal wall-down procedures has been demonstrated to offer enhancements in postoperative physical examination outcomes and improved surgical visualization. On the other hand, selecting canal wall-up procedures allows for the preservation of the natural middle ear anatomy, thereby facilitating improved reconstruction capabilities. This results in a middle ear that closely mimics its physiological state, necessitates less postoperative care, improves the compatibility with hearing aids, and the absence of water-related restrictions [
10]. The reported success rates for complete cholesteatoma removal with canal wall-up mastoidectomy typically range from 70% to 90%. Canal wall-down mastoidectomy demonstrates success rates of 80% to 95%. More extensive procedures, such as modified-radical and radical mastoidectomy, yield success rates ranging from 85% to 95% [
14,
15,
16,
17,
18].
The surgical technique of AE-AE has shown to be an effective technique for the resection of cholesteatoma, with a recidivism rate of 5.2% in our study population. Another study tracked 42 patients for 6 months to 7 years after their initial cholesteatoma surgery using the same technique. They found a recidivism rate of 4.8% [
7]. The lower rate in this study might be because they used less sensitive imaging like CT scans, which may have missed smaller cholesteatomas that MRI could have detected. Additionally, our study included patients with more aggressive cholesteatomas that affected the ossicular chain, which is associated with a higher recidivism rate.
Regarding imaging and postoperative monitoring, high-resolution CT scans were widely used as the main imaging technique for diagnosing and characterizing cholesteatoma. However, its ability to accurately distinguish between fibrotic tissue, cholesteatoma, and inflammation is limited. The interpretation of temporal bone CT entails a labor-intensive process due to the intricate anatomical structures and microarchitecture involved, heavily relying on the reader’s expertise [
19]. Later, EPI-DWI MRI was introduced as a mean to identify cholesteatoma recidivism. Nevertheless, its efficacy in detecting smaller lesions, measuring less than 5 mm, was hindered by magnetic field inhomogeneity artifacts, posing challenges for accurate interpretation. The emergence of non-EPI-DWI MRI has significantly advanced imaging techniques specifically tailored for detecting small cholesteatomas. This method enhances both sensitivity and specificity, as it minimizes artifacts in the temporal bone region and provides superior spatial resolution. Traditional morphological MRI sequences, weighted in T1 and T2, though widely used, exhibit limited specificity in diagnosing inflammatory conditions of the middle ear. The characteristics of cholesteatoma, granulation tissue, or chronic otitis media can manifest differently in these sequences, resulting in a lack of consistent reliability when attempting to differentiate between these conditions [
9].
Diffusion MRI is a highly sensitive and specific technique. This enables the selection of patients for revision surgery, avoiding unnecessary interventions and associated risks. The sensitivity (100%), specificity (96.4%), positive predictive value (60%), and negative predictive value (100%) obtained in our study with this technique are similar to those usually reported.
It is essential to highlight that non-EPI-DWI MRI can occasionally yield false-positive results if another tissue with restricted diffusion is present. Ordinary postsurgical inflammatory alterations do not impede diffusion, but in rare instances, different tissues may develop such behavior. Common factors contributing to false-positive findings are cholesterol granuloma, purulent content, or an abscess in the middle ear. Conversely, cholesteatomas measuring less than 2 mm are frequently reported as the main culprits behind false-negative outcomes [
9].
Our study’s results are consistent with those obtained in another study that evaluated non-EPI DWI MRI’s efficacy. The follow-up study included 35 patients who had undergone initial surgery for cholesteatoma resection, and non-EPI-DWI MRI was used as the diagnostic method. This technique detected 9 out of 10 cholesteatoma recurrences, with the only missed lesion being a 2 mm cholesteatoma with imaging affected by motion artifact in a child. These findings highlight the superior performance of non-EPI-DWI MRI compared to other imaging techniques [
20,
21]. High-resolution CT scans have shown limited reliability in detecting cholesteatoma recurrence, with a sensitivity of 43% and specificity of 48% in similar cases [
8]. Another study involving 45 patients found that EPI-DWI MRI conducted a few months after initial cholesteatoma surgery had a sensitivity of only 12.5% [
22]. These results underscore the effectiveness of non-EPI DWI MRI, which can accurately detect cholesteatoma recurrence with high sensitivity and specificity. However, also in terms of follow-up, recent studies such as that of Covelli et al. in 2022 [
23], advised the possibility of extending follow-up to 7 years to obtain a better perspective and characterize tumor monitoring, if possible. Nevertheless, the authors, in accordance with our work, after analyzing their series of 64 patients, concluded that extending the follow-up to at least 5 years after primary surgery was also recommended to detect cholesteatoma recidivism beyond this time frame. Additionally, they also raised the possibility of performing an early imaging test, just one month after the procedure, to detect the presence of any relevant cholesteatomatous residue, thus suggesting a potential therapeutic failure and the possibility of carrying out an early and minimally invasive reintervention [
23].
However, it is worth noting that sensitivity in another recent study was, unlike ours, 59%. This study included 33 patients who underwent a non-EPI-DWI MRI and a second revision surgery after the first intervention [
24]. Nevertheless, several factors that may have influenced this discrepancy in results need to be evaluated. The number of false negative Non-EPI-DWI MRIs is not known since small recidivism in cholesteatoma can be asymptomatic for some years, this situation makes even harder some parameters to be calculated and it can call into question the role of Non-EPI-DWI MRI in some circumstances as a tool for detecting cholesteatoma recidivism. It is for this reason that a second MRI, i.e., 5 years after surgical intervention may be an alternative to a second look.
The initial aspect to consider is the duration between the imaging examination and the second surgical procedure. In most patients, there was a gap of two months between the two tests, but some patients had to wait from 188 to 201 days until the second surgical intervention was performed. The problem with long periods between both tests is that, given the high capacity of recidivism of this pathology, a new cholesteatoma may appear that did not exist at the time of the imaging test. Another factor is the waiting time between the first surgery and the non-EPI DWI MRI. In this study, approximately nine months elapsed; in ours, it was two years. This implies that the lesions found at nine months may be significantly smaller than those found at two years, making their diagnosis more challenging and increasing the number of false negatives.
Some studies consider the lack of surgical re-intervention as a limitation after not detecting any indication of recidivism in the non-EPI DWI MRI since they consider it the only way to confirm true negative results. Nonetheless, the patients in our study have been followed up through otoscopic examinations and radiological tests. Therefore, along with the high sensitivity of the imaging test, negative results can be classified as true negatives with a high level of certainty. In cases of recidivism MRI is not useful unless intracranial complication are suspected.
A limitation of our study is the heterogeneity among the cases included regarding the aggressiveness of the cholesteatoma. Based on the EAONO classification explained earlier, the study includes cases from the first, second, and third stages. This could lead to an overestimation of the aggressiveness of the pathology and its capacity of recidivism. One possible solution to this problem would be to add only those patients classified within the same EAONO stage as an inclusion criterion. Another possible alternative would be to increase the number of patients, allowing for the stratification of cases into different stages within the same study.
While our study did not identify any parameters demonstrating a statistically significant association with an increased risk of new cholesteatoma recidivism, our investigation did reveal a notable finding regarding the correlation between pre-surgical Pure Tone Average (PTA) and the likelihood of detecting cholesteatoma recidivism in non-EPI-DWI MRI scans (p-value: 0.003). This correlation indicates that as the PTA increases, there is a higher probability of detecting cholesteatoma in imaging tests. This finding underscores the critical importance of early diagnosis and treatment of this pathology before it infiltrates additional structures and exacerbates the progression of the disease. In summary, our study suggests that while certain parameters may not directly predict cholesteatoma recidivism, monitoring PTA levels can provide valuable insights into the likelihood of recidivism, aiding in timely intervention and management strategies.
To conclude future studies could optimize radiological techniques for earlier diagnosis of cholesteatoma recidivism. To address these gaps in knowledge and advance our understanding of this disease, it’s crucial to prioritize and support future research in this area.