Morphometric Comparison of Tympanic Membrane in Unilateral Mucosal Chronic Otitis Media

Abstract

Previous studies of the middle ear in Chronic Otitis Media have focussed on radiological assessment of temporal bone; endoscopic studies were focussed on perforation size and location. The malleus handle (manubrium) is a constant ossicular feature visible on otoendoscopy. It’s relative position may indicate the previously documented differences in ossicular chain and middle ear cleft development relative to the tympanic ring and inner ear capsule between affected ear and contralateral ear. Design: descriptive. Setting: hospital based. Subjects: 84 patients with unilateral mucosal chronic otitis media. Methods: each patient underwent Otoendoscopy and the findings were recorded through a camera connected to a laptop. Using image j software these images were analysed. Participants were of mean age of 35 years with left preponderance, and no gender preponderance was found. In our study, majority of the perforations were small sized perforations. The malleus handle foreshortening was about 92.6% compared to unaffected ear which was statistically insignificant. There was a significant inter group difference in the distance between the tip of the manubrium and the inferior annulus—which was decreased in affected ears and in the distance from the lateral process to the anterior tympanic ring—which was increased in affected ears. There was no difference in the malleo scutal angles. These differences were not dependent on the site or size of the perforation. The differences we found between ears were unexpected. The significance of the differences in the shape of the tympanic membrane between affected and unaffected ears in unilateral mucosal COM need further study. We did not find a statistically significant difference in the malleo-scutal angles between ears.

Introduction

Chronic Otitis Media (mucosal disease) is a common diagnosis among patients attending ENT OPD. This disease causes significant burden on the community with the reported prevalence of 18.7% [1]. Many reasons have been attributed to non-healing perforation such as persistent mucosal disease and permanent perforation syndrome [2].

Previous studies have documented significant differences in the ears affected compared to opposite side including changes in mastoid pneumatization and ossicular erosion [3]. Congenital middle ear anomalies are also another cause of conductive hearing loss [4]. Wullstein opined that the slope of the scutum indicates the extent of pneumatization of the temporal bone [5]. Developmental anomalies of the middle ear cleft may include position anomalies of the ossicular chain and defects in pneumatization of the temporal bone [6].

Previous studies have focused on otoscopy, audiological and radiological assessment of temporal bone [7]. Newer tools and techniques permit more detailed evaluation of otoendoscopic image [8]. The malleus handle (manubrium) is a constant ossicular feature visible on otoendoscopy. Relative position of handle of malleus may be an indicator of the differences in ossicular chain and middle ear cleft development relative to the tympanic ring and inner ear capsule between affected ear and contralateral ear.

Otoendoscopy is a commonly performed relatively inexpensive method which permits examination of the tympanic membrane and record images. Problems with using otoendoscopy images for morphometric analysis include optical distortions like fisheye effect, distortions due to perspective as the tympanic membrane is at an angle to the endoscope, and due to distortions of the monitor and other apparatus. Despite its apparent disadvantages, otoendoscopy is a convenient method of examination for the clinician seeking to evaluate a patient with CSOM. Previous studies with otoendoscopy were focussed on perforation size and location [9]. Mohammed Radef Dawood et al.—in their paper acknowledged that newer tools and techniques permit more detailed evaluation of otoendoscopic image [7]. Alternative to make accurate measurements of the middle ear would be HRCT of the temporal bone which would introduce radiation risks and cost. Even if we accept these limitations, the radiological techniques currently commonly available to the clinician may not permit surface measurements of the tympanic membrane. Taking these points into consideration, we decided to proceed with morphometric analysis of the otoendoscopic images. We augmented the visual evaluation of otoendoscopic images with computer aided measurements to assess the differences between affected and contralateral ears in unilateral mucosal Chronic Otitis Media (COM). We did not find a similar approach used previously within the available literature.

Objectives

1. To enumerate the differences between affected and contralateral ear in cases of unilateral mucosal COM in terms of malleus handle position.

2. To enumerate the differences between affected and contralateral ear in cases of unilateral mucosal COM in terms of angle between scutum and handle of malleus.

Materials and Methods

Source of Data

Adult patients attending ENT OPD with clinical diagnosis of unilateral mucosal COM.

Study design: Hospital based observational study.

Study period: One and half years between 1st January 2021 and 31st December 2021 but due to covid 2nd and 3rd waves the study duration was extended till 1st July 2022.

Place of study: An otorhinolaryngology OPD at a tertiary care medical college hospital.

Sample size: 84.

Inclusion Criteria

1. Unilateral mucosal chronic otitis media (COM).

2. Age more than 18 years.

Exclusion Criteria

1. External ear anomalies like microtia or anotia.

2. Consent not given for otoendoscopy.

3. Cases of unilateral acute otitis media and otitis externa.

4. Patients taking antiplatelet or anticoagulant medications.

5. Impacted cerumen which cannot be cleared up in one sitting.

Methodology

Patients underwent detailed clinical examination and discharge was cleaned. All the patients who have fulfilled inclusion criteria were subjected to otoendoscopy with Karl Storz Zero-degree endoscope. The images were recorded and analyzed using image j software [open source software available at https://imagej.net/ij/index.html] where the measurements were done. A Jobson–Horne probe with a ring of known inner diameter was placed close to the tympanic membrane as a reference to which other measurements were calculated.

We defined the points of measurements on the tympanic membrane as illustrated by the following photograph (Fig. 1a).

Fig. 1.

Measurements

Horizontal measurements (lines perpendicular to the manubrium).

LP–AW a horizontal line from lateral process to the anterior edge of the tympanic membrane.

LP–PW a horizontal line from lateral process to the Posterior edge of the tympanic membrane.

T–AA a horizontal line from tip of manubrium to the anterior part of the annulus.

T–PA a horizontal line from tip of manubrium to the posterior part of the annulus.

Vertical measurements (Lines along the manubrium).

S–LP a vertical line from the scutum to the lateral process of the manubrium.

LP–T a vertical line from the lateral process to the tip of the manubrium.

T–I a vertical line from the tip of the manubrium to the inferior part of the annulus.

We defined the malleo-scutal angle illustrated (Fig. 1b)—as an approximator to the slope of the scutum—which would otherwise be a subjective measurement. One line is drawn parallel to the inferior edge of the scutum(S) and another along the manubrium (M) and the angle formed by their intersection was measured.

The size of the perforation was measured as a proportion of the area of the perforation to the area of the entire tympanic membrane (as a percentage). They were then classified as small (< 25%), Medium (25–50%), Large (> 50%).

For this analysis the perforations were classified as anterior perforations if the perforation was anterior to a line through the manubrium, inferior perforations if they were inferior to a line perpendicular to the manubrium tip at the umbo and posterior perforations if they were posterior to the line through the manubrium.

Z-score was used to analyze significance of difference of means, regression analysis was performed to check correlation of quantitative measurements. One way ANOVA test was performed to check the dependence of the LP–AW and T–I differences with the location of the perforation. The analysis was performed on Microsoft Excel for Mac 16.69.1 and R for Mac version 4.2.2.

To ensure quality of the measurements, the second author selected nine (every 10th participant starting from the first) participants otoendoscopic images, printed them out and measured the defined parameters on the images. We compared the mean of the difference of manual versus software derived measurements.

Results

Mean age of the participants was 35 years and there was no gender preponderance (M:F = 1:1).

Majority of the perforations were small (66.67%), few were medium (28.57%), and few were large (4.76%).

In this study we had perforations located in anteriorly, inferiorly and all quadrants. We did not have any affected ear with exclusively posterior perforation.

We could not find any statistically significant difference when comparing between the ears the distance from the manubrium tip to the anterior annulus (Z = 1.324, P = 0.186), manubrium tip to posterior annulus (Z = 1.343, P = 0.179), Lateral process of malleus to posterior annulus (Z = 0.548, P = 0.583).

Vertical measurements along the malleus handle were also compared. We could not find a statistically significant difference in the distances from Scutum to Lateral Process (Z = 1.366, P = 0.1719); and Lateral Process to tip of handle of malleus (Z = 0.230, P = 0.818).

We also compared the angles between the scutum and the line along the malleus handle (malleo-scutal angle) and the mean angle was smaller in affected ears indicating a higher slope of the scutum; but we did not find any statistical difference between affected and unaffected ears (Z = − 2.355, P = 0.0185) (Table 1).

Table 1.

Malleo scutal angles

	Affected ear	Normal ear
Range	35.214	36.559
Mean	36.8758333	37.8592738
Standard deviation	7.28134459	8.61862124

We found statistically significant differences at 5% level of significance in the distance from the lateral process to the anterior annulus (Z = 2.345, P = 0.0136) and the tip to the inferior annulus in a line along the handle of malleus (Z = − 4903, P < 0.001).

The results are displayed in Fig. 2.

Fig. 2.

Summary of results

There was no significant correlation between the size of the perforation and difference in LP–A (correlation coefficient = 0.03424) or difference in T–I measurements (correlation coefficient = 0.05795).

One way ANOVA test showed no dependence of LP–A differences with site of perforation (dof = 3, F = 0.587, P = 0.625) and no dependence of T–I differences with site of perforation (dof = 3, F = 1.636, P = 0.188).

With our quality control technique as described in the methods, we found an average variation of 9.23% between software derived and manual measurements.

Discussion

In our study mean age of the participants was younger than the age reported in some previous studies; some studies have reported preponderance in young adult age group—Nahata et al., found that COM was more commonly seen in younger individuals (15–24 years) [2].

Keeping the handle of malleus as the traditionally defined and clearly visible landmark in the tympanic membrane, we defined the measurements as discussed in the methodology section. Thus, we have avoided ambiguity in the level and methodology of measurement which will help to standardize further studies. We used ImageJ for measurements on the tympanic membrane—which is an open source software and using simple ratios between the diameter of a probe place in proximity to the tympanic membrane, we could calculate the other measurements. This software has been used previously in studies on the tympanic membrane in chronic otitis media [10]. We cross verified our measurements by manually measuring selected images and found very little variation between software derived and manual measurements on the otoendoscopic image.

The results showed a small but statistically significant difference in the LP to Anterior wall measurement but there was no statistically significant difference when considering LP to posterior wall and total horizontal measurements at LP. If these measurements reflect changes occurring in the corresponding parts of underlying middle ear spaces, we can speculate that this means that affected ears have a difference in the morphology of the anterior mesotympanum or protympanum and a change in the anterior drainage pathway. It is unlikely that this result was influenced by the site or size of the perforation as regression analysis and analysis of variance did not find any correlation or dependence.

The measurements at TIP to AA and TIP to PA are surface markings overlying the anterior mesotympanum. These measurements—tip to anterior annulus and tip to posterior annulus did not show any significant difference; as these values represent the horizontal distance of the mesotympanum at the level of the umbo, we may be able to conclude that there was no difference in mesotympanic size at the level of the umbo.

We expected that affected ears would have a shortening of the handle of malleus (LP to tip length) which is explained as “foreshortening.” We did not find any studies which gave information about extent of foreshortening. In our study, the foreshortening was statistically insignificant. Perhaps the foreshortening couldn’t be seen due to the endoscope picking up the entire length of the manubrium, or due to unrecognized middle ear disease causing retraction of the handle of malleus on the unaffected side. As majority of the diseased ears in the study had a small perforation that also could be a reason why we didn’t pick up foreshortening.

The tip to inferior annulus length was decreased in affected ears—which was statistically significant. Maybe these measurements were affected by optical distortions caused by the endoscope or by actual decrease in the inferior mesotympanic height—we could not differentiate with this study; maybe future studies can confirm this finding. However, perhaps we can say that the apparent decrease in inferior mesotympanic height is a more reliable finding on otoendoscopy rather than shortening of the malleus length itself. A real decrease in the tip to inferior wall distance may also lead us to speculations about predisposing factors for development of COM such as—decreased hypotympanic volume, increased rigidity of the TM or decrease in the compliance of the tympanic membrane in such ears. Further studies will be needed to confirm or deny these possibilities. As there was no correlation of this difference with the size of the perforation and no dependence on site of perforation by statistical analysis, these issues were ruled out.

Our results showed no statistically different result in the heights of the pars flaccida between affected and unaffected ears. If this measurement reflects the pneumatization of the upper mesotympanum, then we can postulate that there was no difference in this part of the middle ear cleft between affected and non-affected ears.

Several hypotheses have been formulated to explain differences in mastoid pneumatization between affected and non-affected ears. Wullstein et al. speculated that the slope of the scutum would indicate the pneumatization of the epitympanum [5]. Therefore, we compared the malleus—scutum angles between affected and non-affected ears as detailed in the methodology. However, we did not find a significant difference in the malleus-scutum angle between affected and non-affected ears. It is possible that these changes do not occur if COM develops after childhood—when mastoid and temporal bone pneumatization has been completed. Since the pneumatization completes in the childhood, if patient gets com in the adult age, it will not affect the pathway of pneumatization or will not change the course of pneumatization which implies the scutum angle also will not change. Therefore, it is possible that there is no significant difference in epitympanic pneumatization between affected and non-affected ears in our study since majority of the participants might have had an onset of disease after completion of middle ear cleft development. Further studies will be needed to confirm our speculations.

In this study we have demonstrated a capability to perform morphometric measurements of tympanic membrane otoendoscopic images. We will need to refine and further study the images to formulate hypotheses and arrive at new and useful information. The application of machine learning algorithms and augmented reality techniques to otoendoscopic images may make it possible for the clinician to identify anatomical variations such as a shortened LP to AW measurement which might indicate middle ear cleft differences. This would lead to easier decision making and development of preventive interventions in such affected individuals. Better treatment plans may be formulated for COM with a better understanding of its pathophysiology and predisposing factors.

Conclusion

In this study we have introduced and defined a methodology for morphometric analysis of the tympanic membrane based on available literature and experience. This study had a novelty in the method of analysis of the tympanic membrane in individuals with unilateral mucosal COM. We had expected a difference in the malleus-scutum angle between affected and unaffected ears as hypothesized by Wullstein et al., but we did not find such a difference in our participants, we found a significant reduction of anterior distance between the lateral process of the malleus and the annulus. We also found a significant increase in the inferior vertical length between the tip of the malleus and annulus. These differences in the shape of the tympanic membrane between affected and unaffected can enable speculation of new hypotheses regarding the pathophysiology of mucosal COM.

Author Contributions

SA: data collection and curation, writing—original draft. RRS: conceptualization, writing—original draft, writing—review and editing, visualization, supervision.

Funding

The authors declare that no funding received.

Declarations

Conflict of interest

None of the authors have any conflicting interests with this paper.

Consent to Participate

Written informed consent obtained from every participant.

Ethical Approval

Institutional Ethics Committee of Rajarajeswari Medical College approved the study with the title ‘Comparison of Relative Position of Handle of Malleus Between Affected and Contralateral Ear in Unilateral Mucosal Chronic Otitis Media on Otoendoscopy’ on date: 16/01/2021 and reapproved for publication on 09/12/2022. The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.