Oronasal Fistula Repair Utilizing a Temporalis Muscle Flap in a Dog with Severe Trismus

Introduction

Squamous cell carcinoma (SCC) is the second most common neoplasm affecting the oral cavity in dogs and is characterized by its locally invasive behavior with a predilection for bone involvement.¹ Radical ablative surgical excision is currently the standard of care for management of these neoplasms but often leaves the surgeon with large tissue defects requiring complex reconstruction. The complication rate following maxillectomy in dogs is high with incisional dehiscence being the most common cause of surgical failure.^2,3 Excessive tension during closure, compromise of tissue vasculature, and infection further complicate revision surgeries for failed maxillectomies.^2–5 The temporalis myofascial rotation flap is commonly used in human oncologic and reconstructive maxillofacial surgery.^6–9 Use of this flap for reconstruction of a complicated orbitonasal fistula has been described in one dog.¹⁰ The purpose of this case report is to describe the application of this surgical treatment in the reconstruction of a chronic oronasal fistula in a dog undergoing treatment for maxillary SCC.

Case Report

A 9 yr old spayed female cocker spaniel presented with a 2 mo history of progressive right-sided facial swelling, halitosis, and ocular discharge. Abnormalities on physical examination at the time of presentation included a moderate lymphadenomegally of the right mandibular lymph node, blephrospasm and epiphora of the right eye, and a large mass involving the right maxilla. Hematologic and serum biochemical analyses were performed and revealed a mild thrombocytosis (461×10³/μL; reference range, 150–430×10³/μL) and a mild elevation of alkaline phosphatase (144 U/L; reference range, 16–111 U/L). A urinalysis obtained by cystocentesis was normal. Radiographic examination of the thorax was negative for metastases, and ultrasonography of the abdomen was unremarkable. Fine-needle aspiration (FNA) and cytology of the right mandibular lymph node revealed the presence of a moderately cellular lymphoid population consisting primarily of small well-differentiated lymphocytes with frequent intermediate and large lymphocytes and plasma cells. These findings were consistent with reactive lymphoid hyperplasia.

Anesthesia was induced with ketamine^a (5 mg/kg IV) and midazolam^b (0.25 mg/kg IV) and maintained with isoflurane^c. A helical computed tomography (CT) scan^d of the thorax with adjoining 2 mm slices was performed. Adontia of the right maxillary canine tooth, third and fourth premolar teeth, and first and second molar teeth was confirmed. Irregular lysis of the associated periodontal bone was identified with extension onto the rostral one-third of the zygomatic bone and ventral floor of the orbit. A mildly contrast-enhancing soft-tissue mass measuring 3.5 cm × 2.7 cm was present within the labial aspect of the maxillary gingiva, which extended across the dental arcade to near midline within the mucoperiosteum of the hard palate. After completion of the CT, an incisional biopsy was obtained from the maxillary mass. Histopathology of the mass revealed a highly cellular, unencapsulated, and poorly demarcated multilobular mass that was effacing and replacing the dermis with invasion into surrounding salivary gland tissue. The mass was composed of cords, lobules, and islands of cuboidal to polygonal plump cells with variably distinct borders and intercellular bridges containing a large amount of granular basophilic cytoplasmic material. A moderate amount of anisokaryosis and anisocytosis was observed, and mitoses averaged one to two per high-power field. These findings were consistent with SCC of the right maxilla.

Two weeks after the CT scan, the dog was readmitted for surgical excision of the maxillary mass. Anesthesia was induced and maintained as described previously. A routine right-sided hemimaxillectomy was performed using a combined dorsolateral and intraoral approach.¹¹ Inferior orbitectomy and exenteration of the globe were also performed based on the extent of the mass identified on the CT scan and to avoid the morbidity associated with high postoperative radiation doses delivered to the eye. Visual inspection of the excised block of tissue revealed tumor extension very close to the caudal and medial margins. Concerns for the inability to adequately reconstruct the defect precluded removal of additional tissue. The defect was closed using a random pattern buccal mucosal advancement flap. The mucosa of the flap was undermined to advance the flap into apposition with the hard palate mucoperiosteum. The flap was sutured with 3-0 polydioxanone (PDS)^e in three layers. The middle layer was anchored to the hard palate using small holes created with a Kirschner wire^f. The deep layer of the flap was closed with simple interrupted sutures from the submucosa of the hard palate to the submucosa of the flap, and the superficial layer was closed with direct mucosal apposition of the flap to the mucoperiosteum. The dog recovered uneventfully from anesthesia and was discharged from the hospital 48 hr postoperatively with deracoxib^g (2 mg/kg per os [PO] q 24 hr for 7 days) and tramadol^h (2 mg/kg PO q 6–8 hr for 5 days).

Histopathology of the submitted mass confirmed an invasive SCC of the right maxilla. Microscopic margin assessment revealed tumor cells within millimeters of the cut edge of the specimen, which correlated with the gross observations made at the time of surgery. A small, unnamed lymph node that was submitted within the tissue block was found to be infiltrated by the tumor.

Six days after surgery, the dog presented for evaluation of a bilateral mucopurulent nasal discharge and bouts of intermittent sneezing after eating. Oral examination revealed a 0.5 cm × 0.5 cm region of dehiscence at the caudal aspect of the maxillary reconstruction where the buccal mucosal flap had been sutured to the mucoperiosteum of the hard palate. The remainder of the dog's examination was unremarkable. Amoxicillin trihydrate/clavulanate potassiumⁱ (14 mg/kg PO q 12 hr for 14 days) was prescribed. The dog's diet was modified to include feeding only large morsels of a semisolid canned food to minimize aspiration of food into the nasal cavity through the fistula. Two weeks after surgery, the dog was readmitted for closure of the acquired oronasal fistula and for another CT scan of the head and neck to be used for radiation therapy planning. Anesthesia was induced and maintained as described previously. To produce a tension-free closure of the fistula, a random pattern bipedicle sliding buccal mucosal advancement flap was created and sutured in place with interrupted 3-0 PDS^e. The dog recovered uneventfully from anesthesia and was discharged from the hospital the same day with instructions to continue all previously prescribed medications.

One week after surgical closure of the acquired oronasal fistula, a 3.5 wk course of fractionated radiation therapy (RT) was initiated. At the time of presentation for initiation of RT, the fistula had recurred but was considered small enough not to warrant further delay in adjuvant RT. Radiation was given in daily fractions on a Monday–Friday schedule. The treatment plan was to administer 16 × 3 gray (Gy) fractions for a total dose of 48 Gy. Anesthesia was induced with IV propofol^j (6 mg/kg) and maintained with isofluorane^c. The radiation dose was delivered through equally weighted bilaterally opposed treatment portals from a 4 MV CLINAC^k. Beam shaping to spare normal tissues was done with prefabricated lead blocks positioned on a leucite tray. The dog was prescribed an anti-inflammatory dose of prednisone^l (0.5 mg/kg PO q 24 hr for 5 days) and amoxicillin trihydrate/clavulanate potassiumⁱ (14 mg/kg q 12 hr for 14 days) midway through the course of radiation. RT proceeded uneventfully, and the patient finished the prescribed course as planned. Two weeks after completion of RT, the dog presented for a routine recheck evaluation. Marked epidermal crusting and alopecia were noted within the previous field of radiation over the right maxilla. Oral examination revealed mild to moderate erythema of the gingival mucosa (that was within the radiation field), and persistence of the oronasal fistula was documented. Because the oral mucosal epithelium was still recovering from acute radiation side effects, repair of the oronasal fistula was not pursued. Clinical signs associated with recurrence of the fistula were seemingly mild and were limited to moderate mucoid discharge from the nares. Recommendations for adjunctive chemotherapy were also discussed during this visit based on the knowledge that the tumor was infiltrating a locoregional lymph node at the time of the original surgery. An alternating regimen of carboplatin^m (220 mg/m² IV q 3 wk for three treatments) and doxorubricinⁿ (30 mg/m² IV q 3 wk for three treatments) was administered to treat metastatic disease over the ensuing 4 mo without complication.

Three months after the completion of RT, the dog was readmitted for evaluation of the oronasal fistula. Clinical signs associated with the fistula consisted of a persistent moderate mucoid nasal discharge. The owners had become adept at performing intraoral lavage of the fistula to prevent accumulation of food material. General anesthesia was induced and maintained using a combination of IV propofol^j (a 6 mg/kg bolus followed by boluses of 1–6 mg/kg totaling 18 mg/kg) and remifentanyl^o (a 5 μg/kg bolus followed by a continuous rate infusion of 0.4 μg/kg/min). The fistula measured approximately 5 mm × 5 mm, and a 3.2 cm silicone nasal septal obturator^p was cut to fit within the fistula so the intraoral portion of the silicone nasal septal obturator extended beyond the border of the fistula by 25 mm.^12,13 Total procedural time for placement of the silicone obturator was 14 min, and at completion of the procedure, the device lay flush with the mucoperiosteum of the palate producing functional closure of the oronasal fistula. The dog recovered uneventfully from anesthesia and was discharged from the hospital the following day.

Three and one-half months after placement of the nasal septal obturator, the dog was evaluated for a 7 day history of progressive sneezing and respiratory stertor. It was reported that nasal discharge had not been observed since the obturator was used to cover the oronasal fistula. Examination of the oral cavity was complicated by the development of a moderate to severe trismus that was presumed to be secondary to the late effects of RT producing fibrosis of the right temporomandibular joint. Despite the trismus, a limited exam of the oral cavity revealed that the nasal septal obturator had become displaced. Under general anesthesia, the obturator was replaced using the same anesthetic protocol and technique described previously.

One year after surgical treatment of the primary tumor, the dog presented for a scheduled comprehensive restaging evaluation. The dog was reported to be doing very at home; however, progressive severe halitosis had been noted since the second nasal septal obturator was placed (4 mo previously). Hematologic and serum biochemical analyses were performed, and an enlarged right mandibular lymph node was aspirated. Thoracic radiographs were taken to assess for pulmonary metastasis. No evidence of metastatic disease was identified, and the results of the blood work were normal. Progression of the previously documented trismus was confirmed, which prevented a comprehensive oral examination. The owners did not report clinical dysfunction (e.g., dysphagia, pain) associated with the trismus. To facilitate oral examination, a brief general anesthesia was induced using IV propofol^j (6 mg/kg). Even under general anesthesia, the jaw could open no more than 3–4 cm using manual distraction. The nasal septal obturator was again found to be dislodged, and the diameter of the fistula was noted to be approximately two-fold larger than at the last evaluation. Attempted surgical revision of the fistula was recommended, but the owners were reluctant due to the previous failed attempts and the fact that the dog's quality of life was only marginally affected by the fistula at that time.

After an additional 6 mo, the dog presented for a second planned restaging evaluation. The dog was deemed to be free of local and metastatic disease. The owners reported that the dog had worsening halitosis and an intermittent nasal discharge. The owners opted for a third surgical attempt at resolving the fistula. Under general anesthesia, a CT scan of the head and neck was performed to assist with surgical planning of the fistula repair and to confirm the absence of any local disease recurrence. The oronasal fistula was identified at the level of the fourth premolar and first molar, with rostrocaudal and mediolateral dimensions measuring 16 mm and 14 mm, respectively (Figure 1). A multilobulated ring-enhancing mass lesion was identified at the medial aspect of the caudal mandible, which extended to the level of the right medial retropharyngeal lymph node. The lymph node was mildly enlarged and demonstrated homogenous contrast enhancement. Ultrasound-guided FNA and cytology of the lymph node demonstrated changes consistent with reactive lymphadenopathy. FNA, cytology, and bacterial culture of the caudal mandibular mass lesion were consistent with a sterile sialocele that likely resulted from radiation-induced damage to the mandibular and sublingual salivary glands.

View Figure

• Download as Powerpoint
• Download Figure

The following day, the dog was anesthetized with propofol^j (6 mg/kg to effect) and positioned in ventrolateral recumbency for surgical reconstruction of the oronasal fistula via a myofascial transposition flap utilizing the temporalis muscle. Due to the presence of the previously documented severe trismus, the surgeons elected to approach the fistula through the nasal cavity and the scar of the previous dorsal skin incision. An approximately 14 cm incision was made from the dorsal aspect of the base of the right ear, starting halfway between the base of the ear and external sagittal crest at the level of the external occipital protuberance, advancing to the dorsal aspect of the commissure of the right lip. The subcutaneous tissue and frontalis muscle were separated until exposure of the entire temporalis muscle and its associated fascia were exposed. The superficial temporal artery, vein, and associated deep temporal branches (serving as the primary blood supply to the temporalis muscle) were identified at the caudal aspect of the zygomatic arch and carefully preserved (Figure 2). As described in previous reports, circumferential and subperiosteal dissection of the temporalis muscle were performed in preparation for transposition of the muscle.^10,14 Ostectomy of a portion of the zygomatic arch to within 5 mm of the retroarticular process facilitated mobilization of the temporalis flap about its base. The remaining portion of the right orbital wall and nasal cavity within the region of the previous maxillectomy were exposed. During this dissection, fibrous bands felt to be causing the trismus were transected, allowing for a greater range of motion of the mandible. The oronasal fistula was then easily identified, and circumferential debridement of the fistula was performed (Figure 3). The temporalis muscle flap was then rotated into the defect and secured with the fascial side of the flap placed intraorally using preplaced simple interrupted sutures of 3-0 PDS^e (Figure 4). The proximal portion of the temporalis muscle flap was split longitudinally, and the dorsal aspect of the coronoid process of the mandible was positioned through the flap. The portions of the flap within the orbit and rostral temporal and parietal regions were secured to the periosteum of the underlying bone using simple interrupted sutures of 3-0 PDS^e (Figure 5). The frontalis muscle and subcutaneous tissue were closed using 3-0 PDS^e in a simple continuous pattern, and skin staples were placed. The dog was then repositioned into dorsal recumbency, and complete intraoral coverage of the fistula was confirmed. Anesthetic recovery was uneventful. At the request of the owners, the dog remained in the hospital for 5 days after surgery. On the second postoperative day, there was an acute episode of mild to moderate intraoral hemorrhage with unilateral (right-sided) epistaxis. A large (approximately 5 cm × 5 cm) well-organized blood clot was identified within the mouth, covering the temporalis muscle flap site, which was immediately removed. The flap appeared intact, viable, and no evidence of the previous hemorrhage existed. The origin of the bleeding was not confirmed but was thought to be from within the nasal cavity. A large clot subsequently formed, which passed uneventfully into the mouth. The bleeding resolved spontaneously, and on the fifth postoperative day, the dog was discharged with amoxicillin trihydrate/clavulanate potassiumⁱ (14 mg/kg PO q 12 hr) and tramadol^h (2 mg/kg PO q 6–8 hr).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Two weeks after the temporalis muscle flap procedure, the dog returned to the hospital to have the skin staples removed (Figure 6). The flap appeared to be intact and partially epithelialized. At the 3 mo postoperative examination, oral examination of the surgical site revealed complete integration and epithelialization of the flap, and the owners reported that signs related to the oronasal fistula had completely resolved.

View Figure

• Download as Powerpoint
• Download Figure

Discussion

SCC is the second most common oral malignancy in dogs, is locally invasive in nature, and has a low prevalence of regional and distant metastasis.¹ At the time of diagnosis, radiographically detectable bone invasion is documented in up to 77% of dogs, necessitating radical ablative surgery to gain local control of the disease.¹ Surgical excision of large tumors within the oral cavity may not be curative because obtaining wide surgical margins is precluded by normal anatomic restrictions. Adjunctive RT is often combined with ablative surgery for oral SCC. With this combined approach, median survival times of up to 34 mo have been reported.^1,15

Although effective for down staging and controlling localized disease, complications associated with surgical therapies for maxillary neoplasms remain high. Frequently reported complications include infection, flap necrosis, and suture line dehiscence leading to oronasal fistula formation.^2,3 In two clinical reports reviewing a series of dogs undergoing maxillectomy, the dehiscence rate with surgery alone ranged from 18% (3/17) to (9/39) 23%.^2,3 Because surgery for oral SCC is often combined with neoadjuvant or adjuvant RT, complication rates are likely higher. A comprehensive study evaluating the effects of RT on reconstructive skin or mucosal (oral) flaps identified complications in 71% (20/26) of dogs.¹⁶ The purpose of the aforementioned study was to determine if flap complication rates could be correlated to the timing of adjunctive RT. Contrary to the findings reported in some human studies, preoperative RT and postoperative RT had similar complication rates in dogs; however, dogs that received preoperative RT were prone to developing more severe flap-related complications.^16–19 Importantly, although the reported complication rate in one study was high, successful healing was ultimately achieved in 85% of cases.¹⁶

The dog described in this report underwent two failed attempts at closure of the oronasal defect. Initially, a less-extensive technique was used employing a bipedicle advancement flap from the surrounding mucosal tissue. Justification for this technique was based on the small size of the defect and the fact that the mucosal viability had not yet been altered by the effects of radiation. In this instance, the definitive cause of local flap failure was unknown; however, excessive tension on the flap is generally cited as the most common cause of failure during reconstruction of the palate.^4,5 Errors in technique such as iatrogenic damage to the flap's blood supply, creation of a flap that is too thin, or insufficient debridement of the recipient bed are other common causes of flap failure.^4,5 Side effects from RT cannot be implicated in the failure of the flap in the dog described herein as dehiscence had occurred prior to initiation RT. In general, a delay of 7 days between surgical wounding and the initiation of RT has been suggested to minimize the incidence of healing complications.¹⁶ Unfortunately, even after 7 days, healing-related complications are possible.

As an interim step in the management of the oronasal fistula in this dog, a silicone prosthetic nasal septal obturator was placed within the palatal defect using previously described techniques.^12,13 The procedure was beneficial in that successful palliation of the clinical signs was achieved when the oronasal fistula wound bed would have been least receptive to tissue manipulation because of acute RT side effects. The surgical technique for placing the obturator was simple and efficient, and an immediate water-tight palatal seal was created. One reported complication of this treatment is premature dislodgement of the obturator, as was seen in this dog; however, placement durations of >2 yr have been achieved in some animals.¹²

A multitude of novel surgical techniques for repair of oronasal fistulas have been described to help circumvent the typical healing problems associated with simple mucosal and mucoperiosteal flap reconstruction. Potential options for this dog included microvascular free tissue transfer (e.g., rectus abdominis muscle flap) and axial pattern flaps. The axial pattern buccal flap based on the angularis oris artery and vein was reportedly particularly useful for palatal reconstruction when local advancement techniques either have failed or were not feasible.⁵ This island flap is freely movable and strong with a robust blood supply. Cadaveric evaluation demonstrated that the flap can be applied to defects as far cranial as the distal gingival margin of the ipsilateral canine tooth as well as defects extending across the palate to the contralateral dental arcade.⁵ Conformational variations in individual clinical patients, and an in depth understanding of the anatomic distribution of the angularis vasculature should be established prior to attempting this procedure.

An extended pedicle flap based on the superficial cervical artery was another axial pattern flap considered for use in the dog described in this report.²⁰ With this technique, the harvested graft tissue is depilated using a pneumatic dermatome and passed into the palatal defect through a carefully placed tunnel within the caudolateral region of the oropharynx. Although cadaveric studies suggested that flap advancement to the level of the canine tooth can be achieved, application in clinical patient results in extension no further rostral than the third premolar and spanning 50–70% of the width of the palate.²⁰ Another disadvantage of this procedure is that harvesting the flap must occur in at least two stages with a minimum convalescent period of 7 days prior to definitive transfer of the tissue.

Unfortunately, the development of progressive and severe trismus complicated the ability to use an intraoral approach for flap generation and dissection. Therefore, an axial pattern flap could not be used. Microvascular free tissue transfer was also deemed excessively risky because the size and orientation of the radiation field was such that the ideal recipient vessels would have received sufficient doses of radiation that could have altered their structural integrity.

A myofascial flap based on the ipsilateral temporalis muscle was used for definitive repair of the oronasal fistula in this dog. In human reconstructive and oncologic ablative surgery, the temporalis muscle flap is one of the most commonly used regional flaps for successful reconstruction of moderate to large intraoral defects. The flap is reported to be versatile, with high survivability and low patient morbidity.^6–9 In veterinary medicine, clinical use of the temporalis muscle for facial and intraoral reconstructive surgery is uncommon. Only one report appears to have been published.¹⁰ In that report, an orbitonasal fistula was repaired with a temporalis muscle flap after five failed closure attempts using other reconstructive techniques. A comprehensive anatomic review of the canine temporalis muscle and technical considerations for generation of the temporalis flap were also discussed in that report.¹⁰

In humans, the temporalis muscle flap, once harvested, can typically span a length of 12–16 cm and tolerates an arc of rotation up to 135°.⁶ The flap can be used for defects associated with the ipsilateral cheek and tongue and can extend as far as the anterior floor of the mouth, including the contralateral palate.⁶ A study using cats as an animal model for temporalis muscle flap healing demonstrated that the flap could be used for successful reconstruction of a defect as large as that created from a hemimaxillectomy.¹⁴ Based on anatomic and breed specific differences in the functional maxillary length in dogs, it may be inappropriate to assume that a similar length of closure could be achieved in dogs until more thorough prospective or cadaveric studies are completed.

After debridement, the size of the repaired oronasal fistula in the dog in this report was approximately 2 cm × 2 cm and was centered within the caudal half of the hard palate. A tension-free closure was easily achieved with near complete mucosal re-epithelialization by the end of the second postoperative week. Rotation of the flap was facilitated by removal of the zygomatic arch and longitudinal transection of the flap so that the flap could be situated over the coronoid process of the mandible. Although the risk of flap necrosis is increased, coronoidectomy and removal of the zygomatic arch are commonly performed in humans receiving a temporalis muscle flap to facilitate rotation into the oral cavity.⁷ Although aesthetics was not of paramount importance for the owners of the dog in this report, the robust nature of the temporalis muscle allowed for a cosmetic reconstruction of the palatal defect as well as the defect created from the previous orbitectomy/ocular exenteration. In cases where enucleation is not performed or when the palatal defect is small, a partial temporalis muscle flap may be adequate for reconstruction. In dogs, the ramifications of partial flap creation on flap viability are poorly understood; however, care should be taken until further research on flap survivability is conducted.

The reported complication rate of the temporalis muscle flap procedure in humans is 30.8% (8/26), with all of these complications considered minor. Partial flap necrosis is one of the more commonly reported minor complications, with an incidence ranging from 8% to 13%.^6,9 In the previously mentioned study using cats as an animal model for healing of the temporalis muscle flap, 24/30 cats had uneventful healing without infection, wound dehiscence, or fistula formation. The remaining six cats were excluded from the study for reasons unrelated to healing complications. Interestingly, three of the cats developed postoperative proptosis of the ipsilateral globe that was thought to be caused by impingement of the coronoid process on the transposed flap in the infratemporal area leading to pressure on the globe from the ventral direction.¹⁴ Aside from a transient bout of oral hemorrhage and epistaxis, no complications were identified with the temporalis muscle flap placement in the dog in this clinical report.

Conclusion

This case report describes the reconstruction of a complicated oronasal fistula in a dog. The temporalis muscle flap is easy to perform, has a low rate of complications, and is potentially an underutilized option for either the reconstruction of oral cavity defects or management of complications associated with radical and ablative maxillofacial surgery.