Surgical treatment of middle ear cholesteatoma using an oral approach in 2 dogs

Abstract

Middle ear cholesteatoma is caused by the formation of epidermoid cysts that result in distention and enlargement of the tympanic bulla with subsequent destruction of surrounding tissues. We report treatment of middle ear cholesteatoma in 2 dogs, via an oral surgical approach. Abnormal tympanic bulla contents and the wall compressing the pharynx were successfully removed in both cases. Computed tomography imaging, surgical findings, and histopathology results were consistent with middle ear cholesteatoma in both cases. The outcomes in both cases suggest that an oral surgical approach may be an alternative treatment for middle ear cholesteatoma in dogs.

Key clinical message:

Despite the limited number of cases described herein, our report indicates that the direct oral approach for canine cholesteatoma may be and alternative approach.

Middle ear cholesteatoma has been reported in humans, rodents, dogs, and 1 cat (1–4). This disease is caused by epidermoid cyst formation, followed by distention and/or enlargement of the tympanic bulla, with destruction of surrounding tissues occurring in the end stage (5). It is either congenital or acquired. The acquired form is secondary to chronic otitis media and/or externa or as a postoperative complication following total ear canal ablation (TECA), lateral bulla osteotomy (LBO), or ventral bulla osteotomy (VBO) (6–8).

If the lesion expands ventrally (3–11), it can compress or occlude the pharynx and result in stertor, stridor, dyspnea, and/or dysphagia (7,10). Abnormal bulla bone formation, expansion, resorption, and sclerosis impair the temporomandibular articulation and induce changes in the paracondylar process (12). Patients can then develop trismus, which may secondarily cause dysphagia, anorexia, and pain at the temporomandibular joints or when opening the mouth (12,13). Secretion of degradative enzymes and chemokines leads to bone resorption (1,2,5), whereas neuropathy occurs due to extension of the disease towards the brain parenchyma or nerves near the tympanic bulla; such as, the trigeminal, facial, vestibulocochlear and glossopharyngeal nerves (5,14).

Diagnosis is based on history and relevant clinical findings, computed tomography (CT) or magnetic resonance imaging (MRI), and histopathological findings (8). However, recognizing clinical symptoms and identifying an oral lesion may aid in detecting bullae extending into the oral cavity (4). Early diagnosis and surgical treatment results in better outcomes and prognosis and moreover, increases the likelihood of curative treatment (4,7–9).

The goal of treatment is to remove all keratinous debris and stratified squamous epithelium, either surgically or non-surgically, and to control infection (7,9,11,15,16). Non-surgical conservative treatment often involves a transcanal endoscopic procedure and medical treatment (11). Surgical approaches are selected based on disease localization, extent, and clinical symptoms. Options include TECA-LBO, VBO, a caudal auricular approach, an oral approach, and/or adjunct segmental mandibulectomy (5,10,13). However, recurrence is common in dogs with middle ear cholesteatoma, with rates of 41.6 to 50% for surgical approaches and 30.8% for transcanal endoscopic procedures (7,8,11). Other studies have reported that complete removal of debris is difficult both in brachycephalic breeds and in severe cases, resulting in higher recurrence rates (6,11).

In the present report, we describe an alternative technique for treating middle ear cholesteatoma in dogs using an oral approach in 2 cases. To our knowledge, no studies have reported long-term outcomes following this approach.

Case descriptions

Case 1

A 10-year-old intact female Lhasa Apso was referred to our institution for dyspnea, stertorous breathing, dysphagia, insomnia, and exercise intolerance. On physical examination, stertor and cyanosis were observed with mild stress or excitement. Left-sided head tilt and ipsilateral facial paralysis were observed on neurological examination. Upper airway obstructive disease and a left-sided neuropathy were the differential diagnoses. As the dog was presented in respiratory crisis, a detailed neurological examination was not performed. Manipulation of the mouth was noted to cause severe cyanosis. Cervical and thoracic radiographs and fluoroscopic findings did not identify a cause for the presenting problem. Complete blood (cell) count (CBC) and serum biochemical analysis findings were within reference limits, except for mildly elevated alkaline phosphatase [303 U/L; reference range (RR): 47 to 254 U/L]. Midazolam (Dormicum; Maruishi Pharmaceutical, Osaka, Japan), 0.3 mg/kg body weight (BW), IV and propofol (Propofol; Pfizer Japan, Tokyo, Japan), 6 mg/kg BW, IV were administered for induction and isoflurane (Isoflurane Inhalation Solution, Pfizer, Tokyo, Japan) was used to maintain a general plane of anesthesia for the oral examination and the CT (Toshiba Medical Systems, model: Asteion TSX-021B, Tochigi, Japan). Helical images were acquired as a volume with the following settings: voxels, 0.2 to 0.35 mm; reconstruction interval: 0.5 or 1.0 mm; rotation speed: 0.75 s; matrix: 120 kVP; tube current, 150 mA; and pitch, 3.5. The volume data were reconstructed in high and low frequency algorithms, and displayed in bone and soft tissue windows, respectively. Isovolumetric transverse, sagittal, and dorsal plane images with a 1-mm slice thickness were reviewed. Oral examination revealed what appeared to be a smooth, pink, mucosa-covered, firm mass compressing and displacing the soft palate from the left dorsal region. There was no inflammation or erosion on the surface of the lesion (Figure 1 a). Due to mild trismus, a thick tongue, and the large left-sided pharyngeal mass, it was extremely difficult to visualize the epiglottis during endotracheal intubation.

Figure 1.

Lesions protruding into the soft palate (circle) in Case 1 (a) and Case 2 (b). The soft palate was dissected longitudinally (b).

P — Palate; T — Tongue.

Computed tomographic findings in Case 1

The left bulla was severely expanded ventrally into a lobular shape (2.3 × 2.3 × 2.5 cm) with non-uniform width of the bulla wall. The medial and lateral walls were thick and sclerotic with lobular bony proliferation, whereas the ventral aspect had multifocal thinning with foci of complete lysis. Intra-luminally, the bulla contained heterogeneous soft tissue contents. The ventral expansion of the bulla resulted in medial-lateral narrowing of the nasal and oropharynx and displacement of the airway and endotracheal tube to the right and ventral. Expansion of the epitympanic recess and widening of the jugular foramen without displacement of intracranial structures was seen. The right tympanic bulla was normal in size and thickness with smooth margins, although was nearly filled with soft tissue-attenuating contents other than a small laterally positioned focus of gas. Both horizontal external ear canals had peripheral, well-defined, curvilinear mineralization of the wall. After advanced imaging, a left middle ear cholesteatoma was highly suspected (Figure 2 a, Figure 3 a–c). The distance to the expanded bulla was 1.4 mm from the soft palate, 17.5 mm from the lateral aspect, and 15.7 mm from the ventral aspect of the surface of the skin.

Figure 2.

Pre- and postoperative computed tomography (CT) images in 2 canine cases. (a) Transverse images for Case 1: preoperative (left) and postoperative (right). Most of the abnormal bulla (white arrows) was resected using an oral approach, except on the ventromedial and lateral sides (black arrows). Contents in hypotympanum and mesotympanum were left intact (b) Transverse images for Case 2: preoperative (left) and postoperative (right). The abnormal bulla (white arrows) was completely resected.

Figure 3.

Preoperative and postoperative three-dimensional (3D) CT of Case 1. a, d — Left lateral view; b, e — Frontal view; c, f — Ventral view. Arrows show the left-sided middle ear cholesteatoma.

A — Anterior; P — Posterior; D — Dorsal; V — Ventral; L — Left; R — Right.

Cytology findings in Case 1

Fine-needle aspiration and cytology were performed after imaging. A 23-G needle was inserted through the soft palate, and white-colored material was collected. Keratinized epithelium, without signs of malignancy, was detectable on microscopy.

Diagnosis and treatment plan for Case 1

Given the strong suspicion of a middle ear cholesteatoma based on CT, surgical treatment was planned for Case 1.

Surgical procedure and recovery for Case 1

Midazolam (0.3 mg/kg BW) was subcutaneously administered as a preanesthetic medication. Propofol (6 mg/kg BW, IV) was infused to induce general anesthesia, and the dog was intubated. Isoflurane inhalation was used to maintain a surgical plane of anesthesia. Cefmetazole (Cefmetazon; Daiichi-Sankyo, Tokyo, Japan), 22 mg/kg BW, IV was administered every 2 h during surgery. Fentanyl (Fentanyl injection; constant rate infusion; Daiichi-Sankyo, Tokyo, Japan), 5 to 15 μg/kg/h, IV was administered as an analgesic.

A temporary tracheostomy was performed before surgery, which was used for intubation and inhalant anesthesia administration, and the dog was positioned in ventral recumbency to allow an unobstructed oral approach. Bandaging tape was used to suspend the dog’s head by passing the tape under the maxillary canine teeth and attaching it to the metallic frame attached to the operating table, using pillows to stabilize the head position.

A 2.5 to 3 cm ventral midline mucosal incision in the soft palate was created using a No. 15 scalpel blade and stay-sutures were applied to retract the incised palatal edges laterally. Bipolar hemostasis was used when approaching the enlarged tympanic bulla. The surface of the bulla appeared immediately, and a bone rongeur was used to create an initial opening into the bulla which was expanded ventrally. Once the ventral part of the bulla had been opened and removed with a rongeur, the bulla contents (including debris and the epithelial lining) were removed using a rongeur and curette. Following removal of the bulla contents and hemostasis, the soft palate was closed with 3-0 and 4-0 polydioxanone (PDS II; Ethicon, Tokyo, Japan) in 2 layers (nasal mucosa and oral mucosa) using a simple continuous pattern. The temporary tracheostomy was closed routinely. The surgical time for the oral surgical approach and bulla debridement was 65 min. The overall surgical time, from temporary tracheostomy to its closure, was 150 min. Large amounts of caseous debris and epithelial tissue were obtained from the enlarged bulla, and some hair was observed within the debris. Samples were submitted for microbial culture, susceptibility testing, and histopathologic evaluation.

Postoperative CT was performed after closure (Figure 2 a; Figure 3 d–f). The severely ventrally expanded hypotympanum was partially reduced by removing the ventral, medial, rostral, and caudal bulla walls, and several remaining bony margins were seen, particularly dorsally. The ventral aspect of the region of the removed bulla contained gas with a thick, soft tissue rim. The contents of the hypotympanum and mesotympanum persisted and were soft tissue attenuating. A single round bony proliferation was seen medially, which persistently displaced the nasopharynx to the right. The oral cavity caudally was normally positioned along the midline, without ventral displacement.

The dog recovered uneventfully. One dose of methylprednisolone sodium succinate (Solu-Medrol; Pfizer, Tokyo, Japan), 5 mg/kg BW, IV was administered postoperatively to reduce the risk of pharyngeal inflammation. This medication is given routinely following staphylectomy at our institution and there was no notable swelling observed at the surgical site at the conclusion of surgery, nor were there clinical signs of upper airway obstruction signs following extubation. A mild cough, attributed to the temporary tracheostomy, was observed during hospitalization, which resolved by the time of discharge. Although mild trismus still persisted, stertor, dyspnea, dysphagia, and insomnia completely resolved postoperatively. The presenting clinical signs were most likely due to the large mass effect extending to the pharynx and oral cavity. The dog was kept in an intensive care unit (ICU) for 2 d after surgery, and in a general ward for the following 3 days. No major postoperative complications were observed. Gabapentin (Gabapentin tablets; Fuji Pharma, Toyama, Japan), 10 mg/kg BW, PO, tramadol hydrochloride (Tramal OD Tablets; Pfizer, Tokyo, Japan), 5 mg/kg BW, PO, and cefmetazole (22 mg/kg BW, IV) were administered twice a day until discharge. The owner declined CT monitoring and the dog was lost to follow-up after discharge from the hospital.

Histopathology results indicated that multi-layered keratin and keratinization-stratified squamous epithelium was present within an epidermoid cyst, corresponding to a cholesteatoma. It was aseptic based on culture results.

Case 2

A 10-year-old spayed female miniature dachshund presented with loud stertorous breathing during neck flexion. Owners also mentioned that they had observed a reluctance to fully open the mouth for a couple of months prior to presentation. On detailed physical examination, a smooth, pink, mucosa-covered, firm mass was detected along the left side of the soft palate, extending to the oral cavity. The dog did not have observable external ear canal disease, irregularity of the oral mucosa, bleeding, or halitosis. No neurological signs were observed. Differential diagnoses included an oral or nasopharyngeal mass, which includes a middle ear cholesteatoma. Findings from a CBC and serum biochemical analysis were within reference limits, except for mildly decreased blood urea nitrogen (7.2 mg/dL; RR: 7.3 to 29.2 mg/dL) and creatinine (0.34 mg/dL; RR: 0.40 to 1.45 mg/dL) levels. Thoracic radiographs revealed no abnormalities. The dog was prepared for head CT. Propofol was administered for anesthesia induction (8 mg/kg, IV), and inhaled isoflurane was used to maintain general anesthesia.

Computed tomographic findings in Case 2

The left bulla was severely expanded ventrally and medially retaining a circular to oval shape (2.0 × 2.0 × 2.6 cm) with irregular and non-uniform margins of the bulla wall. There was a small focus of complete lysis seen laterally and mild irregularly margined bone formation seen medially. Luminally, the bulla contains heterogeneous soft tissue contents. The ventral and medial expansion of the bulla resulted in medial-lateral narrowing of the oropharynx and ventral displacement of the oral cavity and endotracheal tube. Hyoid bone deviation was also observed. Expansion of the epitympanic recess and widening of the jugular foramen without displacement of intracranial structures was seen. The right tympanic bulla was normal in size and thickness with smooth margins, although it contained soft tissue-attenuating contents. Both horizontal external ear canals contained gas. After advanced imaging, a left middle ear cholesteatoma was highly suspected (Figure 2 b, Figure 4 a–c). The distance to the expanded bulla was 0.78 mm from the soft palate, 23 mm from the lateral aspect, and 21.9 mm from the ventral aspect of the surface of the skin.

Figure 4.

Preoperative, postoperative, and 14 mo postoperative 3D CT of Case 2. a, d, g — Left lateral view; b, e, h — Frontal view; c, f, i — Ventral view. Arrows show the left-sided middle ear cholesteatoma. No recurrence is observed on CT images.

Diagnosis and treatment plan for Case 2

Cholesteatoma was highly suspected based on CT findings, and surgical treatment was planned.

Surgical procedure and recovery for Case 2

Oral gabapentin (10 mg/kg BW) was administered 3 h before induction for analgesia. Midazolam (0.3 mg/kg BW) was subcutaneously administered as a preanesthetic medication. Propofol (8 mg/kg BW, IV) was infused to induce general anesthesia, and the dog was intubated. Isoflurane inhalation was used to maintain a surgical plane of anesthesia. Cefmetazole (22 mg/kg BW, IV) was administered every 2 h during surgery. Fentanyl (5 to 15 μg/kg/h, IV, CRI) was administered as an analgesic.

A temporary tracheostomy was performed before the oral approach. The dog’s position, preparation, and approach were the same as those in Case 1 (Figure 1 b).

Small arterial bleeding occurred, which was suspected to derive from a branch of the maxillary artery, and hemostasis was achieved using titanium surgical clips (Hemoclips; Mizuho, Tokyo, Japan). After removing the enlarged ventral part of the bulla, debris, and epithelial lining, the tympanic bulla was inspected using a 5.4 mm 0-degree rigid scope (Olympus, Tokyo, Japan) positioned through the incision site; no abnormal contents remained (Figure 5). Closure was performed in the same manner as in Case 1. The duration of the oral surgical procedure was 50 min. The overall surgical time, from temporary tracheostomy to closure, was 138 min. Samples were submitted for microbial culture, susceptibility testing, and histopathologic evaluation.

Figure 5.

Scope image after treating a middle ear cholesteatoma from the oral cavity in Case 2. No persistence of keratin debris or epithelial tissue is left inside. White dashes show the edge of the soft palate incision. Debrided tympanic cavity is shown in between the white dashes.

D — Dorsal; L — Left; R — Right; T — Tongue; * — Epiglottis.

Postoperative CT was performed following closure (Figure 2 b; Figure 4 d–f). The severely rounded and expanded bulla was subtotally excised by removing the ventral, medial, rostral, and caudal bulla walls with minimal remaining bony margins seen. The surgical site contained gas with a thick, soft tissue rim. The nasopharynx and oral cavity were normally positioned along the midline, without displacement. The dog recovered uneventfully. Methylprednisolone sodium succinate (5 mg/kg BW, IV) was administered to reduce pharyngeal inflammation. Except for a mild cough, which resolved by the time of discharge, there were no major postoperative complications. The dog was kept in an ICU for 2 d after surgery, and in a general ward for the following 2 d. Gabapentin (10 mg/kg BW, oral), tramadol hydrochloride (5 mg/kg BW, oral), and cefmetazole (22 mg/kg BW, IV) were administered twice daily until discharge. Clinical signs of stertor and reluctance to fully open the mouth were no longer noted. No recurrence was observed on CT at 3 and 14 mo (Figure 4 g–i) after surgery. No clinical signs of recurrence have been observed to date (64 mo after surgery).

Histopathology revealed an epidermoid cyst with multilayered keratin and keratinization-stratified squamous epithelium consistent with cholesteatoma. Culture was negative.

Discussion

Both medical and surgical treatment of middle ear cholesteatoma have been previously reported (3–11,13). The goal of treatment is to remove all keratinous debris and stratified squamous epithelium while controlling secondary infection (5,7,8). Reducing the size of the abnormal bulla is critical in advanced cases because it can occlude the pharynx and inhibit jaw movement, as was seen in our cases. Two major surgical approaches to the bulla are the ventral and the lateral. The ventral approach is commonly selected when ear canal ablation is not required. The lateral approach is chosen in subjects with severely diseased external ear canals, in which TECA is concomitantly required. Although, some surgeons perform the lateral approach regardless of the presence of external ear canal disease.

In our cases, the external ear canal was not involved as it can be in other canine cases and most human cases (5,17,18). During oral examination, the lesions bulged from the soft palate which prompted us to use an oral approach, even with mild trismus in Case 1. Furthermore, the distance to the lesions was the shortest using an oral approach according to the CT images. An oral approach was planned to resect and remove the ventrally distended abnormal bulla and its contents. It was reasoned that performing a VBO would have necessitated greater tissue dissection and been more difficult due to the greater distances to the lesions.

The only structure dissected using this oral approach was the soft palate. The short surgical times, especially from incision to removal of the bulla, further highlight the straightforwardness of the procedure. Because of altered anatomy, the approach in the narrow working space was more complicated than a typical staphylectomy. The oral approach provided a clear view of the ventral aspect of the expanded bulla. Although it is reported in other literature (14), there were no notable differences between the brachycephalic breed (Case 1) and the dolichocephalic breed (Case 2) even with the benefit of hindsight. The effectiveness of this surgical approach is observed in postoperative CT images, which showed approximately 70 to 80% removal (author’s subjective opinion) or complete removal of the expansile bulla walls and large amount of keratinous debris/epithelial tissue. In addition, postoperative improvement of clinical signs, namely stertor, dyspnea, dysphagia, and insomnia, also document the effectiveness of the approach in these 2 dogs.

In their cadaveric study, that was performed after our procedures, Manou et al (14) reported that the oral approach to the normal bulla is feasible in dogs. The key difference in our study and the cadaveric study is the altered anatomy by the middle ear cholesteatoma. Due to this difference, incision of the soft palate was required in our cases, whereas in their study, the soft palate was retracted laterally (14). An oral approach to the normal tympanic bulla was reported to be safe from the anatomic aspects; however, we encountered hemorrhage in Case 2. Because the anatomy was altered, we suspected that the branch of the maxillary artery may have been deviated medially. Caution is needed to avoid serious hemorrhage when treating middle ear cholesteatomas through any surgical approach.

Recurrence is frequently observed in subjects with cholesteatoma (41.6 to 50%), with a mean time to recurrence of 7.5 mo (7,8). This high incidence may be due to the persistence of keratin debris or the perimatrix (3,8). In Case 1, the dog presented as an emergency, and the surgical treatment aimed to remove the lesion and its contents to alleviate symptoms of respiratory distress; surgery achieved this goal. Some of the soft tissue bulla contents were left in place, which may result in recurrence. The dog was lost to follow-up and the final outcome remains unknown. Although brachycephalic dogs are reported to be the most difficult facial conformation to approach the normal bulla ventrally (14), the remaining hypotympanum and middle ear content in Case 1 is in part due to the disease severity and emergency nature of the surgery. As the authors were comfortable with the size reduction and resolution of clinical signs, repeat surgery was not considered necessary. In Case 2, there were no clinical signs of recurrence at the time of writing (64 mo after surgery). Previous studies have reported that severe cases were associated with higher rates of recurrence than moderate cases (7,8,11), as the chance of leaving debris, keratin, or epithelium is greater (11). It is difficult to achieve resolution of the disease, regardless of the technique used, with no apparent difference in clinical outcomes between LBO and VBO (7). Use of video endoscopy may help ensure clearance of the epithelial lining or keratin debris, as was done in Case 2. With video endoscopy, further resection could have been possible in Case 1; however, this was not available on emergency basis. In humans, video-assisted surgery has been reported to reduce the incidence of recurrence in patients with middle ear cholesteatoma (19–21). This may help remove more of the abnormal bulla wall, keratin debris, and epithelium for canine middle ear cholesteatoma. Due to the high rate of recurrence, management and follow-up after surgery are recommended in dogs with this disease.

In conclusion, our report indicates that the direct oral approach for canine cholesteatoma is an effective procedure that may be an alternative to TECA-LBO or VBO. An oral approach involves minimal tissue dissection and is associated with short surgical times; however, it does require familiarity with the regional anatomy and technical expertise from the surgeon. Visualization is good for the ventral aspect of the lesion with this approach, and can be improved at the dorsal and lateral aspects with the use of video endoscopy to avoid persisting abnormal bulla wall, keratin debris, or the perimatrix. Furthermore, temporary tracheostomy, which was performed in our dogs, may not be necessary in all dogs, particularly in less severe cases and in cases with a coiled endotracheal tube, which can reduce surgical and anesthetic times. Future studies with a larger number of dogs with middle ear cholesteatoma and varying surgical approaches are necessary to determine best management techniques for this disease.