Evaluation of Medial Retrotympanum by High-Resolution Computed Tomography and Endoscopy – A Descriptive Study

Abstract

The aim of this study was to determine and compare the capability of the High-Resolution Computed Tomography (HRCT) and endoscope in detecting hidden areas of medial retrotympanum in cases with chronic otitis media (COM). Per-operative endoscopic evaluation of structures in medial retrotympanum was done in 74 patients suffering from COM and was compared with pre-operative HRCT of the temporal bone in 50 patients. HRCT revealed type-A Sinus Tympani (ST) in 61.2%, type-B in 34.7%, and type-C in 4.1% of the patients; endoscope could reveal type-A in 71.6%, type-B in 25.7% and type-C in 1.4% of the patients. Ponticulus was discovered in 84.1%, subiculum in 100%, finiculus in 97.3% and sinus subtympanicus (SST) in 100% of the cases using endoscope. HRCT could detect ponticulus in 38.6% and subiculum in 4.0% of the cases. HRCT could not very clearly detect finiculus and SST. HRCT could not differentiate the content of the ST as compared with endoscopy. HRCT is a useful diagnostic tool in predicting the presence and type of ST, but its accuracy in detecting SST, finer bony ridges of ponticulus, subiculum, and finiculus and the contents is poorer than endoscope.

Introduction

Surgical intervention in cases of chronic otitis media (COM) aims at complete eradication of the disease and restoring the near normal physiology of the middle ear that includes hearing and ventilation, thus providing a disease free life during follow up [1].

The middle ear holds multiple hidden areas such as the retrotympanum (RT), epitympanum, supratubal recess, protympanum and the hypotympanum. RT is the posterior most outpouching in the middle ear. The pyramidal eminence in RT forms the fulcrum to which the chordal ridge and ponticulus gets attached, dividing the RT into multiple recesses and sinuses. The vertical segment of facial nerve divides the RT into medial and lateral compartments of which the medial retrotympanum (MRT) hold potential spaces for residual disease after surgery [2–5].

Eradication of disease from these potential spaces require extensive drilling using microscopes. In the attempt to acquire adequate visualization and disease clearance, the normal physiology of the middle ear gets altered permanently [2]. However, the advent of endoscopes has not only improved our expertise while operating in the middle ear but also has helped to achieve complete removal of the disease and providing a good clinical outcome [6].

High resolution computed tomography (HRCT) is being widely used pre-operatively to assess the disease extent, ossicular erosion, plan the surgical approach and detect associated complications. It also helps us to see beyond the hypertrophied or oedematous mucosa and granulations in diseased middle ear into the bony anatomy [7]. However, the radiology of the retrotympanum has received far less attention.

Majority of the earlier studies have described the details of the structures revealed with the help of endoscope or HRCT, but have not tried to compare the capability of these tools in identifying these structures in different stages of management in diseased ears [8–13]. An otologic surgeon needs to know precisely the capability of both of these diagnostic tools in divulging the anatomical details of RT in diseased ears, for reporting the extent of the disease, and to predict the degree of exploration required during the surgery. The present study thus aims at comparing the capability of endoscopy and HRCT in describing the anatomical details of MRT in diseased ears.

Materials and Methods

The study was conducted in the department of Otorhinolaryngology and Head-Neck Surgery, at a tertiary care hospital in northern India over a period of 18 months from January 2018 to June 2019. This study was approved by the institutional review board (Reference No. AIIMS/IEC/18/99). Informed consent was obtained from all participants regarding the materials used in the study.

Study Population

Total of 74 patients of chronic otitis media were enrolled in the study after excluding known cases of neoplastic or granulomatous condition and patients with history of temporal bone fracture or middle ear surgery. Among them, 32 had squamous type and 42 had mucosal type of chronic otitis media. The enrolled patients underwent relevant otological surgery (tympanoplasty/canal wall down or intact canal wall mastoidectomy) out of which a pre-operative HRCT of the temporal bone was done for all cases with squamous variety of COM and for 18 cases of mucosal variety. This study has primarily focused on the MRT since it was challenging to visualize the lateral retrotympanum without a posterior tympanotomy. Two spaces of the MRT, sinus tympani (ST) and sinus subtympanicus (SST), as well as three bony projections of the MRT, the ponticulus, subiculum, and finiculus were studied in detail with endoscopy (n = 74) [Figs. 1, 2 and 3] and then compared with the pre-operative HRCT in cases where both the modalities were available (n = 50) [Figs. 4 and 5].

Fig. 1.

Diagram of retrotympanum showing the types of Sinus Tympani (ST) (A) Type A is a limited space with no medial or posterior extension, (B) Type B extends medialy, (C) Type C has both medial and posterior extension, with respect to the vertical portion of facial nerve; ST = Sinus Tympani; Po = Ponticulus; Su Subiculum; SS = Sinus-subtympanicus; FR = Facial recess; LTS = Lateral tympanic sinus; PS = posterior sinus; Ce = Chordal eminence; Cr = Chordal ridge; Se = Styloid eminence; Pe = Pyramidal eminence; Rw = Round window; FN = Facial nerve; Pr = Promontory; St = Stapedial tendon; S = Stapes

Fig. 2.

Axial section of High Resolution Computed Tomographic Scan of right temporal bone showing types of sinus tympani (A) Type A, (B), Type B, and (C) Type C; Pr = Promontory; ST = Sinus tympani; FN = Facial Nerve; Pe = Pyramidal eminence; PSCC = Posterior semicircular canal; Co = Cochlea

Fig. 3.

Endoscopic view of the medial retrotympanum (A) Visualization of ST by introducing the 0^o endoscope from the opposite side, (B) Medial retrotympanum as seen with 30° endoscope, (C) Medial retrotympanum as seen with 0° endoscope; ST = Sinus tympani; SST = Sinus subtympanicus; FN = Facial nerve; Pe = Pyramidal eminence Po = Ponticulus; Su = Subiculum; Fi = Finiculus; In = Incus; Hom = Handle of malleus; Rw = Round window; St = Stapedial tendon; Sp = Stapes; Pr = Promontory

Fig. 4.

Axial section of High Resolution Computed Tomographic Scan of right temporal bone showing (A) Ponticulus (= Po, yellow arrow), (B) Subiculum (Su, white arrow) and Sinus Subtympanicus (= ST); Po = Ponticulus; Pr = Promontory; ST = Sinus tympani; FN = Facial Nerve; Pe = Pyramidal eminence; PSCC = Posterior semicircular canal; Co = Cochlea

Fig. 5.

Diagram of retrotympanum showing the types of Ponticulus, Subiculum and Finiculus (A) Ridge type (B) Bridge type (red arrow) (C) Incomplete/Absent; ST = Sinus Tympani; Po = Ponticulus; Su Subiculum; SS = Sinus-subtympanicus; Fi = Finiculus; FR = Facial recess; LTS = Lateral tympanic sinus; PS = posterior sinus; Ce = Chordal eminence; Cr = Chordal ridge; Se = Styloid eminence; Pe = pyramidal eminence; Rw = Round window; FN = Facial nerve; Pr = Promontory; S = Stapes

HRCT

Pre-operative HRCT was obtained using 128 Slice dual source Siemens somatom definition flash computed tomography scanner. Helical acquisition in the axial projections were obtained with sequential slice of 1.0 mm section thickness with an interval of 0.7 mm between the slices; scanning was performed from the arcuate eminence to the jugular fossa. Radio-morphological evaluation of RT in digital images was done using RadiAnt PACS DICOM viewer version − 5.0.2.

Endoscopy

Intraoperative endoscopic evaluation of the MRT was done using 4 mm outer diameter 0^o, 30^o and 45⁰ (Karl Storz, Tuttlingen, Germany) before and after complete removal of the diseased mucosa and decongestion with 1:200,000 adrenaline-soaked gelfoam. A 3-chip high-resolution monitor and camera (Karl Storz, Tuttlingen, Germany) were used for all of the procedures.

Sinus Tympani

The ST, which lies medial to the vertical segment of the facial nerve between the ponticulus and subiculum, was assessed at the level of the posterior semi-circular canal in the axial sections in HRCT. Two important landmarks used were the pyramidal eminence and the facial nerve. The type of ST was classified as type A,B,C with type A being a shallow space with no medial or posterior extension and the depth of the ST corresponding to the medial limit of the facial nerve, type B being a deep space with the medial extent of the space reaching posterior to the vertical segment of the facial nerve and type C being a deeper space with medial and posterior extension with respect to the vertical segment of the facial nerve [Figs. 1 and 2]. The diameter of the orifice of ST was measured at its medial opening. Additionally the soft tissue density in the ST and the depth were noted.

Endoscopically, the type of ST was then assessed after removal of the diseased mucosa and cholesteatoma and before reconstructing the middle ear. To visualize the medial boundary of the ST, the endoscope was introduced from the opposite side with respect to the affected ear. The type of ST was analyzed using the similar classification as used for HRCT [Fig. 3] and the contents of the ST was noted before removal.

Ponticulus, Subiculum & Finiculus

Ponticulus is the superior limit of ST separating it from the posterior sinus arising from the pyramidal eminence to the promontory. Subiculum which arises from the posterior pillar of the round window to the styloid complex marks the inferior limit of ST dividing the RT into superior and inferior MRT. Finiculus forms a clear borderline between the RT and hypotympanum. It arises from the inferior portion of the styloid prominence and connects the anterior and inferior lip of the round window niche. Using HRCT, these three structures were identified using the axial or coronal sections and attempt was made to further categorise them as bridge, ridge, incomplete, or absent [Fig. 4]. On endoscopy, these structures were further classified as mentioned above. [Figure 5].

Sinus Subtympanicus

Between the subiculum and the finiculus lies the SST. On endoscopy, It can be shallow or deep based on its extension beyond the styloid eminence. The space was identified on HRCT with subiculum as the landmark.

Statistical Analysis

The data was analysed using IBM SPSS software (Statistical Product and Service Solutions). Sensitivity, specificity, positive or negative predictive values (PPV and NPV), and hence diagnostic accuracy (DA) of HRCT was determined considering endoscopy as the gold standard for all the structures other than sinus tympani, for which HRCT was considered as the gold standard [9–11]. All categorical data were expressed as count and proportions. Cohen’s Kappa test was used to check to test agreement between the diagnostic tests. Values of k ≤ 0 were taken as indicating no agreement and 0.01–0.20 as none to slight, 0.21–0.40 as fair, 0.41– 0.60 as moderate, 0.61–0.80 as substantial, and 0.81–1.00 as almost perfect agreement [16]. Fisher’s exact test was used to determine if there are non-random associations between two categorical variables. A p-value of < 0.05 was considered statistically significant.

Results

Sinus Tympani

Endoscopic assessment of the ST after complete removal of the diseased mucosa or secretions and decongestion showed type A in 71.6% of cases, type B in 25.7% and type C in 1.4% and the type could not be assessed in 1.4% cases (n = 74) due to extensive hypertrophied mucosa and excessive bleeding. Pre-operative HRCT in axial sections revealed type-A ST in 61.2%, type-B in 34.7%, and type-C in 4.1% of the patients (n = 50). Presence of hypertrophied mucosa and bleeding made endoscopic classification of ST difficult and hence HRCT was considered a better tool to assess the type of ST. Cohen’s Kappa test reveals only a fair agreement (k = 0.588, p = 0.01) between both tools in detecting the types of ST [Table 1]. The sensitivity, specificity, NPV and PPV of endoscopy in detecting the type of ST were calculated [Table 2].

Table 1.

Agreeability of HRCT of temporal bone and endoscopy in detecting ST types

HRCT	Endoscopy				Cohen’s kappa
	Type A	Type B	Type C	Total	κ value	P-value
Type A	27 (55.1%)*	3 (6.1%)	0 (0.0%)	30 (61.2%)	0.588	0.01
Type B	8 (16.3%)	9 (18.4%)*	0 (0.0%)	17 (34.7%)
Type C	0 (0.0%)	1 (2.0%)	1 (2.0%)*	2 (4.1%)
Total	35 (71.4%)	13 (26.5%)	1 (2.0%)	49 (100.0%)

Total (n = 49);Types A,B&C represent sinus tympani types. *Cases where the two methods agreed. HRCT = High resolution computed tomography; ST = Sinus tympani

Table 2.

Diagnostic capability of endoscopy for detecting type of ST with HRCT of temporal bone as ‘gold standard’

ST types	Sensitivity	Specificity	PPV	NPV	Diagnostic Accuracy
Type A	87	58	77	73	76
Type B	53	88	69	78	76
Type C	50	100	100	98	49

Total n = 50. Data represent percentages. HRCT = High resolution computed tomography; PPV = positive predictive value; NPV = negative predictive value; ST = Sinus tympani

The diameter of the orifice of ST ranged between 0.76 and 3.19 mm and depth ranged between 1.01 and 5.67 mm on HRCT which was not measurable using the endoscope. Unlike endoscope which revealed disease in 29.6%, secretions in 40.7% and hypertrophied mucosa in 29.6%; HRCT showed only soft tissue density and was not able to differentiate further. This was statistically significant (p < 0.001) using Fisher Exact test.[Table 3].

Table 3.

Comparison of the contents of ST as seen with endoscopy and in HRCT temporal bone

Endoscopy	HRCT			Fisher’s Exact Test
	STD	Clear	Total	X^2	P Value
Disease	8 (29.6%)	0 (0.0%)	8 (16.0%)	50.000	< 0.001
Secretions	11 (40.7%)	0 (0.0%)	11 (22.0%)
Mucosa	8 (29.6%)	0 (0.0%)	8 (16.0%)
Clear	0 (0.0%)	23 (100.0%)	23 (46.0%)
Total	27 (100.0%)	23 (100.0%)	50 (100.0%)

Total n = 50. HRCT = High resolution computed tomography; ST = Sinus tympani; STD = Soft tissue density

Ponticulus

Ponticulus was present in 85.1% of which 55.4% was ridge type, 18.9% was bridge type while it was incomplete in 1.4% and absent in 14.9% of the cases using endoscope. There was difficulty in visualization of ponticulus in 9.5% of cases because of the mucosal hypertrophy and bleeding (n = 74). HRCT could however detect ponticulus only in 38.6% of the cases (n = 50). Cohen’s kappa gives only a fair agreement between HRCT and endoscopy (k = 0.213, p = 0.176) [Table 4]. Diagnostic accuracy of ponticulus was only 54.6% with a low sensitivity of 45.5% [Table 5] .

Table 4.

Agreeability of HRCT of temporal bone and endoscopy in detecting the ponticulus

HRCT	Endoscopy			Cohen’s kappa
	Present	Absent	Total	κ value	P-value
Present	17 (38.6%)*	0 (0.0%)	17 (38.6%)	0.213	0.176
Absent	20 (45.5%)	7 (15.9%)*	27 (61.4%)
Total	37 (84.1%)	7 (15.9%)	44 (100.0%)

Total n = 44. Absent and present refer to detection of the ponticulus. *Cases where the two methods agreed; HRCT = High resolution computed tomography

Table 5.

Diagnostic capability of HRCT of temporal bone for detecting structures in retrotympanum with endoscopy as ‘gold standard’

Structure	Sensitivity	Specificity	PPV	NPV	Diagnostic Accuracy
Ponticulus	46	100	100	26	55
Subiculum	4	-	100	0	4
Finiculus	-	-	-	-	0
SST	4	-	100	0	4

Total n = 50. Data represent percentages. HRCT = High resolution computed tomography; PPV = positive predictive value; NPV = negative predictive value; SST = Sinus subtympanicus

Subiculum

HRCT could detect subiculum only in 4.0% of the cases (n = 50). While, endoscope could detect subiculum in all the patients (n = 74); in 21.6% of the cases, it was bridge type, 78.4% of the cases it was ridge type. Cohen’s Kappa test reveals no agreement (k = 0, p = 1.0) between both tools in detecting the subiculum [Table 6]. HRCT has only 4% sensitivity in detecting subiculum [Table 5].

Table 6.

Agreeability of HRCT of temporal bone and endoscopy in detecting the subiculum

HRCT	Endoscopy			Cohen’s kappa
	Present	Absent	Total	κ value	P-value
Present	2 (4.0%)*	0 (0.0%)	2 (4.0%)	0.000	1.000
Absent	48 (96.0%)	0 (0.0%)*	48 (96.0%)
Total	50 (100.0%)	0 (0.0%)	50 (100.0%)

Total n = 50. Absent and present refer to detection of the subiculum. *Cases where the two methods agreed; HRCT = High resolution computed tomography

Finiculus & Sinus-subtympanicus

Finiculus on endoscopy was identified in 97.3% of which 79.7% showed ridge type, 13.5% showed bridge type and incomplete 4.1%. Similarly sinus sub-tympanicus was deep in 16.2% and shallow in 83.8%, depending upon its extension with respect to the styloid eminence [Table 5]. However, both these structures were poorly detected on HRCT.

Ponticulus, subiculum and finiculus could not be further classified in HRCT as seen with endoscopy.

Discussion

The RT is a complex space in the middle ear that is divided into multiple spaces by bony ledges arising from various prominences, such as the pyramidal eminence, chordal eminence, and styloid eminence [3, 4]. Though endoscopes in otology are over two decades old, its use in otologic surgeries are still being explored. There has been several studies done on the anatomy of the temporal bone especially the middle ear with the use of microscope. The complexity of the middle ear anatomy still demands better techniques to explore the difficult-to visualize areas with the operating microscope and endoscope has not yet been used to understand the same at large except for a few studies [6]. Endoscopic exploration of the ST was started by Thomassin et al. using angled endoscopes in 1993 followed by Badr-El-Dine et al. and Baki et al. in 2002 to 2003 [6, 14, 15]. However it was in 2009 that Marchioni et al. published a series of work which not only explored the retrotympanum to greater depths but also studied the anatomical variations and its significance in residual disease [9].

Endoscopes have greatly improved visibility of hidden areas of the middle ear, and their use in cholesteatoma surgery can significantly reduce the amount of drilling required. As a result, its use in ear surgery has surged in the last decade. However, mucosal oedema, hypertrophy, granulations, adhesions, and cholesteatoma impair its ability to distinguish fine outpouchings like ST and bony prominences like subiculum in a diseased ear. However, HRCT of the temporal bone can help in detailed preoperative understanding of these areas, overcoming this limitation. However, HRCT is not always reliable and may occasionally over- or under-diagnose the significant findings. [8–12]. The surgeon must be well-versed in which region is best visible preoperatively with HRCT of the temporal bone, or whether it would be wiser to check for it during surgery with the help of an endoscope.

The efforts of various authors in describing the structures of RT using HRCT or endoscope have been compared [Table 7]. Parlier-Cuau C et al. published the first detailed study of structures in RT as seen in HRCT in 1998. They scanned 66 temporal bones with HRCT in 49 patients and 24 cadaveric temporal bones. The findings were double-checked using anatomical sections from three cadaveric bones. They found that HRCT could detect ST in 95% of cases, ponticulus in 63%, and subiculum in 57% of cases [8].

Table 7.

Areas of retrotympanum reported by previous studies

Study (year)	Cases (n)	Tool	Structures in retrotympanum	% retrotympanum areas reported
Parlier-Cuau et al. [8] (1998)	66 normal + 24 cadaveric bones	HRCT + anatomical sections	Sinus tympani	95
			Ponticulus	63
			Subiculum	57
Marchioni et al. [9] (2009)	40 diseased	HRCT + endoscopy	Sinus tympani	95
			Ponticulus	95
Marchioni et al. [10] (2010)	25 diseased	Endoscopy	Subiculum	84
			Finiculus	56
			Sinus subtympanicus	56
Marchioni et al. [11] (2015)	298 diseased	HRCT + endoscopy	Sinus tympani	Types A = 33.1, B = 62.5, C = 4.4. (Sensitivity = 91%, specificity = 65%, PPV = 68%, NPV = 90.)
Marchioni et al. [12] (2015)	65 diseased	Endoscopy	Finiculus	89.2
			Sinus subtympanicus	70.7
El-Anwar et al. [13] (2019)	56 normal	HRCT	Sinus tympani	Types A = 28, B = 71, C = 1
Present study (2021)	50 diseased	HRCT + endoscopy	Sinus tympani, ponticulus, subiculum, finiculus, sinus subtympanicus	See Tables 1, 4 and 6

PPV = positive predictive value; NPV = negative predictive value; HRCT = high-resolution computed tomography of temporal bone

In the study conducted by Marchioni et al. in the year 2015, the capability of HRCT in predicting disease in ST was evaluated. They found that HRCT has high sensitivity and NPV for showing the involvement of ST by cholesteatoma. They also suggested that for a shallower ST, an exclusive endoscopic exploration is better; while in the case of a deeper type C ST, a retrofacial approach is preferable [9–12, 15]. El-Anwar et al. recently published a study in which they studied HRCTs of normal temporal bones and found that they can help predict the condition of ST and supratubal recess in patients [13]. These studies provide a basis for understanding the endoscopic and radiological anatomy of RT, but they do not compare the diagnostic capabilities of these two tools.

` The aim of this study is to compare the diagnostic capabilities of endoscopes and temporal bone HRCT in detecting structures in MRT in diseased ears. It was observed that HRCT is a better tool for detecting the type of ST than the endoscope, which shows DA as low as 49% in the case of type C and as high as 76% in the case of type A. The reason for this is the obscurity caused by the disease in the middle ear. It was noted that the sensitivity & diagnostic accuracy of endoscopy in detecting the type of ST decreased as the depth of the ST increased. However, the NPV and specificity increased with increase in the depth of the ST using endoscope. This clearly shows that the endoscope is not a good tool as compared to HRCT in finding out the type of ST in diseased ears. Rather than relying on per-operative findings, a surgeon should look for the type of ST in temporal bone HRCT to be prepared to explore this area during surgery. Also, the orifice of ST could be as narrow as 0.7 mm giving rise to a bottle neck like space leading to higher chances of residual disease as compared to a deeper ST with a wider orifice. However the contents of ST had to be assessed endoscopically as HRCT could not differentiate the contents clearly.

In contrast to ST, structures like ponticulus, subiculum, finiculus and sinus subtympanicus are best detected with endoscope, even in the diseased ears. Although HRCT had a 100% specificity in identifying ponticulus, it did not detect the type of ponticulus. It can thus be explained that, multiplanar HRCT with sliced sections at 0.7 mm intervals can detect wider spaces like ST but miss out on thin bony ridges at times. This also means that in the ears where HRCT detects these structures, surgeons will almost certainly detect them with the endoscope during surgery. This may also aid the surgeon in determining, predicting, and reporting the extent of bony destruction in the extensive disease during the preoperative period.

Conclusion

HRCT is a useful diagnostic tool in predicting the presence and type of ST, but its reliability in detecting SST, finer bony ridges of ponticulus, subiculum and finiculus is poorer than endoscope. High PPV of HRCT suggests that if it detects these structures, surgeon can surely detect them with the endoscope during surgery. Endoscope, on the other hand, may miss ST in the diseased ears if not already detected or suspected in HRCT. It can be said, therefore, that the two tools are complementary to each other, but neither can act as a diagnostic substitute for the other.

Declarations

Competing interests

The author has no competing interests to declare that are relevant to the content of this article.

Conflict of interests

The author declares no conflict of interests

Ethics approval and consent to participate

This study was approved by the institutional review board of All India Institute of Medical Sciences, Rishikesh (Reference No. AIIMS/IEC/18/99). The study was performed in accordance with the ethical standards as laid down in the Declaration of Helsinki and its later amendments.

Informed consent

Informed consent was obtained from all individual participants or their legal guardians (if minor) included in the study.