MRI Findings of a Middle Ear Cholesteatoma in a Dog

Introduction

Cholesteatomas are lesions formed from keratinizing stratified squamous epithelium mislocated in the middle ear. The secondary hyperkeratosis and shedding of keratin debris usually results in an expansile, cystic mass surrounded by an inflammatory reaction.¹ This condition is frequently associated with inflammation of the middle ear and has been reported in 11% of canine ears affected by otitis media.² Cholesteatoma appears to be an uncommon condition in dogs; however, the incidence may be higher than reported. Indeed, this condition may either not be recognized or misdiagnosed as an inflammatory polyp or cyst.

Previous studies have described the radiographic and computed tomographic (CT) features of cholesteatomas.^2–4 CT appears to be the method of choice for detection of cholesteatomas. In veterinary medicine, CT is considered to be superior to MRI in detecting lesions of the inner and middle ear affecting bony structures; however, MRI provides better resolution of soft-tissue lesions.^5,6

MRI was used in two previous studies to diagnose canine cholesteatomas, but no detailed description of the imaging features was presented in those cases.^4,7 This report describes the detailed radiographic and MRI appearance of a middle ear cholesteatoma in a dog.

Case Report

An 8 yr old castrated male flat-coated retriever presented for assessment of chronic intermittent ear disease. Pain on opening the jaw had been present for 2 wk, including several episodes of acute pain. The owner reported circling and a head tilt, neither of which were observed on physical examination. Physical examination revealed pain on opening the jaw, although there was no reduction in the range of motion. There was enophthalmos and ptosis of the left eye, as well as intermittent head shaking. The remainder of the physical examination was unremarkable. Neurologic examination revealed no additional abnormalities. Ophthalmic examination revealed conjunctivitis of the left eye and bilaterally reduced lacrimation (the Schirmer tear test was 10 mm and 15 mm in the left and right eyes, respectively). Otoscopic examination showed bilateral rupture of the tympanic membranes with a hyperplastic area at the entrance to the middle ear on the left side. Serum biochemistry and hematology were unremarkable. Ear cytology and culture revealed Malassezia and Staphylococcus pseudintermedius bilaterally.

On the ventrodorsal radiograph of the skull, calcification was evident around the right external ear canal. The left ear canal was obliterated by soft-tissue opacity with proliferative new bone extending into the surrounding tissues, superimposed on the caudal aspect of the zygomatic arch and temporomandibular joint. There was a small area of new bone on the medial aspect of the caudal zygomatic bone that was level with the lateral aspect of the mandibular condyloid process. On the rostrocaudal open-mouth projection, both the right and left tympanic bullae were poorly delineated and difficult to assess further.

MRI of the head was subsequently performed with a 1.0T magnet^a. Sagittal T2-weighted (T2W; repetition time [TR] = 5,315 msec, echo time [TE] = 105 msec, slice thickness = 3 mm), transverse T2W (TR = 6,090 msec, TE = 100 msec, slice thickness = 3.5 mm), dorsal T1-weighted (T1W; TR = 425 msec, TE = 12.5 msec, slice thickness = 3 mm), transverse and sagittal T1W (TR = 450 msec, TE = 12 msec, slice thickness = 3.5mm), and fluid-attenuated inversion recovery (TR = 5,900 msec, TE = 100 msec, inversion time = 2,000 msec, flip angle = 90°, slice thickness = 3.5 mm) sequences were acquired. Additional transverse and dorsal T1W sequences were acquired following IV administration of 0.1 mmol/kg of gadopentetate dimeglumine^b with the settings described above.

The MRI showed a severely expanded left tympanic bulla with a markedly thickened and irregularly shaped wall consistent with the radiographic findings. The distorted bulla extended in a caudocranial direction from the level of the foramen magnum to the left temporomandibular joint and lateromedially from the level of the vertical ear canal to the temporomandibular joint. The temporomandibular joint appeared normal. The marked distortion of the wall prevented the authors from distinguishing between the bony part of the bulla tympanica and the inner lining. The wall of the left tympanic bulla appeared of mixed intensity on T2W images and hypointense to normal bone on T1W images, suggesting a lack of medullary fat and probable sclerosis of the bone. Similarly, the adjacent area of the petrous temporal bone was hypointense on both T1W and T2W images, indicating probable reactionary sclerosis. Some partial enhancement of the inner lining of the bulla was observed on T1W images following contrast administration, suggesting the presence of inflammation.

The content of the left bulla was predominantly isointense to brain tissue on T1W and of mixed intensity on T2W and fluid attenuated inversion recovery sequences (Figure 1). On T2W images, the medial one-third of the bulla content was hyperintense to brain tissue, consistent with a fluid signal. Contents did not enhance following contrast administration (Figure 2). The left cochlear and vestibular canals were incorporated into the mass of this material and were not visible as discrete structures. The left horizontal ear canal was distorted because it was being pushed dorsally and was partially filled with material arising from the lumen of the bulla. The left retropharyngeal lymph nodes were mildly enlarged compared to those on the right.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

On the right side, the bulla was small with a thickened wall. The inner wall and the external ear canal lining enhanced following contrast administration. The horizontal canal was narrowed and partially filled with T1W and T2W hyperintense material at the level of the temporomandibular joint, probably representing proteinaceous fluid.

Based on the radiographic and MRI findings, bilateral chronic otitis externa and media and left-sided otitis interna with massive distortion of the left bulla were suspected. The left bulla was suspected to contain inspissated material. Neoplasia could not be excluded, but was less likely due to the lack of postcontrast enhancement. Finally, there was a mild enophthalmos of the left eye, probably due to disruption of the sympathetic nerve supply via the left bulla.

Total ear canal ablation and a lateral bulla osteotomy were performed on the left ear. The facial nerve was not directly visualized during this procedure. Exploration of the left bulla revealed a mass of inspissated material filling the chamber. This material was removed, and samples were submitted for histopathology, bacterial culture, and sensitivity. The remaining cavity was deep (>4 cm). The normal anatomy was obliterated, leaving an undulating and irregular contour. All of the secretory lining or capsule that could be visualized was removed.

Sections were taken from the surgically resected left ear canal and from a part of the mass adjacent to the left tympanic bulla. Mild lymphoplasmacytic inflammation and associated epithelial hyperplasia of the lining of the external ear canal was identified. The tissue mass comprised a more severe chronic inflammatory reaction with lymphoplasmacytic inflammatory cells and pigment-laden macrophages infiltrating the stroma between dilated ceruminous glands. A sample of the wall of the left tympanic bulla was also obtained (Figure 3). Strips of pigmented stratified squamous epithelium with accumulation of overlying keratin were noted. The microscopic appearance of this lesion was consistent with aural cholesteatoma associated with underlying otitis externa and media. Bacterial culture was positive for S. pseudintermedius.

View Figure

• Download as Powerpoint
• Download Figure

The dog underwent a total ear canal ablation and lateral bulla osteotomy on the right ear 2 wk later. Histopathology confirmed the provisional diagnosis of chronic otitis media and externa.

Discussion

Cholesteatomas are usually classified as either congenital or acquired. However, their exact pathophysiology continues to be debated widely, and recent studies favor a possible common origin and overlap in the pathophysiology.¹ In most clinical situations, such as the case presented here, cholesteatomas are associated with chronic otitis media and are therefore considered to be acquired. However, a congenital lesion is difficult to completely exclude as secondary infection may mean that a cholesteatoma is not recognized until late in the course of the disease process^2,4

The clinical signs presented in this case are similar to those described in previous reports.^2,4 Head tilt, head shaking, scratching, pain on opening the jaw, ataxia, nystagmus, and circling are the most common signs reported with cholesteatomas. S. pseudintermedius cultured from the middle ear was frequently observed with cholesteatomas in a previous study. This pathogen is typically associated with otitis media.^4,8

Previous reports of middle ear cholesteatomas in dogs have noted stenosis of the ipsilateral ear canal associated with calcification, increased density of the middle ear, and either a disruption or change in shape of the wall of the tympanic bulla. In some cases, the temporomandibular joint and the retroglenoid process showed radiologic evidence of bone remodeling.² Expansion of the bulla, especially when coupled with the presence of bone lysis in regions adjacent to the bulla, often leads the clinician to suspect aggressive neoplasia. However, clinicians should be aware that middle ear neoplasia is rare in dogs and usually involves extension of disease from the external ear causing permeative lysis.⁹ Cholesteatomas, on the other hand, induce bone destruction as a consequence of tympanic bulla expansion due to the accumulation of keratin debris.¹ In the case reported herein, the radiographic changes were nonspecific. Obliteration of the external ear canal and calcification of the surrounding tissues were noted, but were consistent with otitis externa. The walls of the tympanic bullae were indistinct, and this, coupled with the proliferative new bone extending into the soft tissues on the left, led to the suspicion of a more aggressive process than a simple otitis media. It was therefore considered necessary to use advanced imaging to define the changes further, which would permit informed selection of appropriate treatment options.

In humans, the diagnosis of a primary cholesteatoma is based mainly on clinical suspicion, and imaging provides useful additional information.¹⁰ Nevertheless, the detection of either a primary or residual cholesteatoma with the use of imaging techniques remains challenging. The primary examination tool for the evaluation of a suspected primary acquired cholesteatoma is CT because CT provides good information on cholesteatoma delineation and extension, delineation of the tympanic segment of the facial nerve, bony erosion, and anatomic features such as the delineation of the tegmen and bony labyrinth.^11,12 In veterinary medicine, CT is used routinely as the method of choice to investigate canine middle ear disease.^6,13,14 Moreover, CT enables differentiation between inflammatory changes, cholesteatoma, and neoplasia in most patients.⁶ Previous studies have reported the CT findings associated with cholesteatoma in dogs.^3,4 One reported heterogeneous contrast enhancement within the middle ear in six dogs and diffuse contrast enhancement in one dog.⁴ No significant contrast enhancement of the bulla content was found in another study.³ The lack of contrast enhancement was suspected to be due to the avascular nature of the accumulated keratin debris and the presence of enhancement to increased vascularity within the perimatrix.^4,15 A peripheral ring enhancement pattern was detected in 36% of patients.³ No histologic features were found to explain the ring enhancement, and it has been suggested that ring enhancement could be an indicator of potential recurrence because 75% of the patients with recurrence had ring enhancement.³ Ring enhancement was not observed in another report.⁴

It should be noted that CT is not widely accessible to veterinarians outside specialist referral practices and institutions in the UK. In contrast, mobile MRI facilities are more readily available. MRI was used in two previous studies to diagnose either primary or recurrent cholesteatomas in dogs, but no detailed description of the imaging features was reported.^4,7 The cholesteatomas were simply described as expansile masses arising from the middle ear that were isointense to the brain on T1W images and nonenhancing following contrast administration. Some intracranial involvement was reported in each study, but was not seen in this case. The pre- and postcontrast MRI findings observed in the case described herein were similar to those reported in humans in that absent to moderate postcontrast peripheral enhancement was present.¹⁶

In surgically and nonsurgically treated human ears, it has been demonstrated that pre- and postcontrast conventional MRI allows differentiation between inflammatory mucosa, granulation tissue, and a cholesteatoma when CT could not consistently enable the differentiation.¹⁷ Inflammatory and granulation tissues enhance strongly postcontrast injection, whereas cholesteatomas do not. Differentiation between granulation tissue and soft-tissue masses is crucial and has implications in the choice of treatment. In this case, the partially enhancing rim was suspected to represent a portion of inflammatory lining of the left bulla associated with surrounding inflammatory tissues. The surgery was recommended due to the severe expansion of the lesion and severe distortion of the bulla.

This report describes an advanced and large cholesteatoma that was easily detected on conventional sequences. Small cholesteatomas may be hard to detect, and reports of human cases have shown that standard MRI sequences are unable to detect residual cholesteatoma prior to a second surgical intervention (termed second-stage surgery in humans).¹⁸ The detection of small lesions is essential in human patients as most cholesteatomas recur within the first 2 postoperative years, with 60% occurring during the first year after surgery.¹⁹

A study in dogs showed that cholesteatomas also frequently recur postoperatively. In that study, 10 of 20 dogs presented with recurrence of clinical signs after surgery. A recurrent cholesteatoma was confirmed in five cases with either CT or MRI, but no information concerning the size of the lesions was available. The variables that had a significant effect on recurrence after surgery were an inability to open the mouth, neurologic signs, bulla lysis, and lysis of the squamous or petrosal portion of the temporal bone. The authors of that study also suggested that early CT imaging of dogs with persistent signs of ear disease might allow intervention earlier in the course of the disease when surgery is more likely to be curative.⁴

Spin echo diffusion-weighted MRI (DWI), which has primarily been used for the diagnosis of ischemic brain infarction, has recently been described as an additional diagnostic tool in detecting cholesteatomas.^10,12,20 Cholesteatomas appear to have consistently high signal intensity on DWI, and a combination of restricted water diffusion and a T2 shine-through effect is thought to be responsible.¹² However, some authors suggest that DWI should still be used in conjunction with standard MRI sequences because not all cholesteatomas will be detected by DWI alone.¹² Moreover, a study assessed the use of DWI associated with delayed postcontrast T1W images and advocated the concurrent use of DWI. DWI was considered more specific than delayed postcontrast T1W images (73% and 55%, respectively), but delayed postcontrast T1W images were more sensitive (60% and 90%, respectively).²⁰ Unfortunately, DWI sequences or delayed postcontrast images were not available in this case, which prevented any comparison with human reports.

Conclusion

In veterinary medicine, the MRI appearance of cholesteatomas has not previously been described in detail. MRI is likely to be of use in selected cases for further characterization of soft tissue changes and identification of possible complications of cholesteatomas. However, further studies are needed to characterize the range of MRI findings in cases of cholesteatomas to fully assess its usefulness.