Functional Outcomes After Transoral Plus Lateral Pharyngotomy Approach for Advanced Oral and Oropharyngeal Tumors

Abstract Objective The aim of this study was to evaluate our institutional experience with the combined transoral plus lateral pharyngotomy (TO+LP) approach in a subset of patients with advanced or recurrent oral and oropharyngeal malignancy. Study Design A retrospective study of procedures utilizing TO+LP for cancer resection between January 2007 and July 2019. Setting Tertiary academic medical center. Methods Thirty‐one patients underwent a TO+LP approach for the resection of oral and oropharyngeal tumors. Functional and oncologic outcomes were analyzed. Results Eighteen (58.1%) patients were treated with TO+LP for recurrent disease. Twenty‐nine required free tissue transfer and 2 (6.5%) had positive margins. The median time to decannulation was 22 days (range 6‐100 days). Thirteen (41.9%) patients still required enteral feeding at their most recent follow‐up. Patients without a history of prior radiation were decannulated sooner (p = .009) and were less likely to require enteral feeding at the first postoperative follow‐up (p = .034) than those who had prior head and neck radiotherapy. Conclusion A TO+LP approach can be used to achieve good functional and oncologic results for selected patients with advanced or recurrent oral and oropharyngeal cancer when minimally invasive options such as transoral robotic surgery, transoral laser microsurgery, or radiotherapy are not possible.

T ransoral approaches are regularly employed for oral cavity malignancies and for early tumors of the oropharynx through the use of transoral robotic surgery (TORS) or transoral laser microsurgery (TLM) techniques. Larger or recurrent tumors involving both the oral cavity and oropharynx, as well as those of the oropharynx which are not amenable to TORS, TLM, or nonoperative options, are typically managed with the lip split and mandibulotomy approach. While this approach affords excellent visualization for the resection of large oropharyngeal tumors, it comes with an increased risk of morbidity, with reported mandibulotomy-related complication rates of up to 11%. 1,2 The transcervical approach to the oropharynx, known as the lateral pharyngotomy, 3,4 has been well described for the treatment of early-stage tumors. [5][6][7] With the advent of TORS, TLM, and improvements in primary radiotherapy therapy, it has been utilized less frequently in recent years. 8 However, in select patients, a combined transoral and lateral pharyngotomy (TO+LP) approach remain useful. In our institution, we employ this approach for many larger or recurrent tumors involving both the oral cavity and pharynx that would otherwise be managed with a mandibulotomy with lip split when a TORS, TLM, or nonoperative approach is not possible.
Most existing reports of functional and oncologic outcomes after lateral pharyngotomy are dominated by patients with smaller and previously untreated tumors. [9][10][11][12] Furthermore, most utilized primary closure [5][6][7][8][9] or local flaps 11 to reapproximate the pharynx, with very few reported free flaps for reconstruction. 6 Due to the limited published data assessing outcomes among patients who undergo lateral pharyngotomy and free flap reconstruction for advanced or recurrent oropharyngeal malignancies, the aim of this study was to evaluate our institutional experience with the combined TO+LP approach in this subset of patients.

Materials and Methods
We obtained Institutional Review Board approval and queried the University of Wisconsin Head and Neck Cancer Database for all patients treated for head and neck cancer who underwent regional or free tissue transfer and lateral pharyngotomy between January 1, 2007 and July 1, 2019. Patients who underwent laryngectomy, maxillectomy, parotidectomy, or skull base resections were excluded. From the remaining list, patients with oral cavity or oropharyngeal malignancy who were treated with a combined TO+LP approach were included.
The technique for the TO+LP begins with the removal of the submandibular gland and level one lymph nodes with ligation of the facial artery and common facial vein branches. Next, cranial nerve 12 and the superior laryngeal nerve are freed from surrounding tissues, the lingual artery is ligated, the digastric muscle is divided, and the greater horn of the hyoid on the side of the resection is removed. With this, the pharynx is separated from the vasculature and the cranial nerve branches. With a malleable ribbon retractor placed lateral to the pharyngeal constrictors, the anterior and superior aspects of the resection can be accomplished transorally and mobilized through the pharyngotomy. The remaining resection is then accomplished transcervically. This often requires mobilization of the resection around the hypoglossal nerve to preserve this structure ( Figure 1). Reconstruction, typically with free tissue transfer, proceeds in the reverse order with the inferior and posterior aspects of the flap inset accomplished through the lateral pharyngotomy and the superior and anterior elements accomplished transorally ( Figure 2). Full inset is accomplished prior to revascularization of free tissue transfers.
Patient records were reviewed for demographic data, staging, treatment, postoperative course, recurrence, and survival information. The initial tumor stage was determined using the Tumor, Node, and Metastasis/ American Joint Committee on Cancer 8th edition. Complications directly relating to surgery were recorded if they occurred within 1 year after surgery. Postoperative swallow function was recorded based on first (1-2 week) postoperative follow-up notes from the surgeon or speech-language pathologist, as available. The most recent swallow function was recorded based on the last documented visit with an oncologist or speech-language pathologist in which the swallow function was discussed. The duration of tracheostomy dependence was calculated as the time elapsed from tracheostomy placement to decannulation. The duration of enteral feeding was calculated as the time elapsed between surgery and removal of all mechanisms of enteral feeding (nasogastric tube or gastrostomy tube [g-tube]).
Independent, 2-sample t tests and χ 2 tests of independence were used to compare tumor size, margin status, functional outcomes, complication rates, and time to most recent swallow assessment between patients who had undergone radiation therapy at some time prior to TO+LP, versus those with no prior radiotherapy. If data did not meet the statistical assumptions for parametric testing, a Wilcoxon-Mann-Whitney rank sum test was performed. Kaplan-Meier survival curves were generated to analyze disease-free survival for patients who had surgical resection via TO+LP approach for their initial disease presentation and for those treated for a recurrence. Statistical analysis was performed using SPSS, version 26 (SPSS, Inc and IBM Company) and a p value of less than .05 was considered statistically significant.

Demographics
Between 2007 and 2019, 31 patients underwent surgical resection of an oral or oropharyngeal tumor using a TO +LP approach. Patient ages ranged from 42 to 73 years (mean 60.1 years). The mean Charlson comorbidity index was 4.4, and 24 (77.4%) patients were current or former smokers with an average of 29.3 pack years. Other demographic details are shown in Table 1.

Disease Characteristics
An overview of patient disease characteristics and history is shown in Table 2. The tongue base was the most involved disease site (21/31, 67.7%). Twenty patients (64.5%) had stage IV disease at the initial diagnosis. Eighteen (58.1%) patients had surgery with a TO+LP approach for recurrent disease, with a mean time of 2.26 years between initial treatment and the TO+LP for recurrence. Seventeen (54.8%) patients received head and neck radiotherapy prior to undergoing the TO+LP procedure, either as part of the initial treatment for their cancer, which then recurred or for a separate primary tumor of the head and neck. The most common histopathologic tumor type was squamous cell carcinoma (25/31, 80.6%) ( Table 2).

Treatment and Operative Course
The mean maximum dimension of the tumors resected via TO+LP was 3.4 cm. Two (6.5%) patients had positive margins noted on the final pathology ( Table 3). Twentynine (93.5%) patients required free flap reconstruction of the TO+LP defect. Radial forearm and anterolateral thigh free flaps were the most used reconstructive methods at 58.1% (18/31) and 38.7% (12/31), respectively, with 1 patient requiring both types due to the size of the defect. One patient (3.1%) required regional tissue transfer for reconstruction. One patient underwent primary closure without the need for additional reconstruction. There were no instances of flap failure. Twelve patients, 1 of whom had received prior radiotherapy, went on to have adjuvant radiotherapy after surgical resection via TO+LP. Of those who received surgical resection followed by adjuvant radiotherapy, 7 had oral cavity

Postoperative Course
Data relating to the postoperative courses are shown in Table 4. Median postoperative admission length was 9 days (range 2-18), with 5 patients requiring readmission. Readmission reasons included 2 instances of neck surgical site infection without fistula requiring IV antibiotics, and 1 instance each of recurrent disease, chest pain, and aspiration pneumonia. The overall 1-year postoperative complication rate was 32%, with 10 patients having at least 1 complication. The most common postoperative complication was aspiration pneumonia. Only 1 (3.2%) patient developed a salivary fistula.

Functional Outcomes
Thirty (96.8%) patients had a tracheostomy placed during surgery and 1 patient did not require a tracheostomy. The median time to decannulation was 22 days (range 6-100). One patient was never decannulated ( Table 4). Twenty-five (80.7%) patients were discharged with a gtube. Five (16.1%) patients did not receive an immediate postoperative g-tube but were discharged with enteral feeding via nasogastric tube. Thirteen (41.9%) patients never had their g-tube removed. For those patients who temporarily required enteral feeding, the median time to discontinuation of enteral feeding was 71 days (range 11-451). A summary of patient swallow function (type of feeding tolerated) at the first (1-2 week) postoperative follow-up visit and at the most recent follow-up visit is shown in Table 4. The mean time to the most recent assessment of swallow function was 27.0 months (median 18.0 months, range 3.0-87.8).

Recurrence and Survival
Fifteen (48.4%) patients had no recurrence after resection via TO+LP; 2 additional patients experienced isolated distant metastases, for a total locoregional control rate of 54.8% ( Figure 3). The median duration of disease-free follow-up was 28.2 months (range 3.3-70.7). The median time to recurrence was 4.8 months (range 0.5-75.8).
Disease-free survival was analyzed for subjects who had surgical resection via TO+LP for their initial disease presentation and for those treated for recurrence (Figure 4). At the median follow-up time of 28.2 months, disease-free survival was 89% among patients treated for initial disease and 18% among patients treated for a recurrence.

History of Prior Radiation Therapy
Patients with prior radiotherapy were, on average, decannulated 22 days later than those without a history of radiation to the head and neck (p = .009). There was a significant association between a history of prior radiation and swallow function at first postoperative follow-up, with 75% (12/16) of patients with prior radiotherapy requiring enteral feeding at first followup, compared with 36% (5/14) of patients without prior radiotherapy (Χ 2 2 = 6.8, p = .034). While there was a wide range in time to the most recent assessment of swallow function across all patients, there was no significant difference in follow-up time detected between those with a history of radiation therapy and those without a history of radiation therapy (p = .313). The association between history of radiation and swallow function at the last documented follow-up approached significance when all categories of feeding type tolerated were included (Χ 2 3 = 7.0, p = .073). While there was not a significant difference in the time to removal of enteral feeding between groups, those with prior radiation had a significantly lower chance of ever having enteral feeding removed, with 69% (11/16) of patients with prior radiotherapy still requiring enteral feeding at last documented follow-up, compared with 21% (3/14) of patients without a history of radiation (Χ 2 1 = 5.4, p = .020). There were no significant differences in any other functional outcomes or the rate of postoperative complications between groups.

Discussion
Most patients treated with TO+LP had larger tumors, a history of radiation, or recurrent disease. Using the TO +LP approach for resection, an excellent negative margin rate was achieved in this population. Furthermore, postoperative complications were limited, and functional outcomes were comparable to those reported in prior studies of early-stage patients.

Margin Status After TO+LP
In the current study, negative margins were achieved using the TO+LP approach in most patients, many of whom had been previously irradiated. Surgical planes following high-dose radiotherapy can be obscured by soft tissue fibrosis and telangiectasias. Therefore, complete resection of the recurrent tumor frequently requires radical excision with wide clinical margins. 13 Our negative margin rate of 94% (88% among previously irradiated patients) compares favorably with rates of 61% to 97% reported in the literature. [5][6][7]9,11,12,14,15 This speaks to the quality of exposure achievable with the TO+LP even in more technically challenging situations.

Oncologic Outcomes After TO+LP
While the rate of disease-free survival at 28 months after LP for our entire cohort was 47%, it was 89% among patients treated for an initial diagnosis of head and neck cancer. This compares favorably with disease-free survival rates of 61% to 86% following lateral pharyngotomy among other previously untreated cohorts. 7,9,11 Diseasefree survival rates for patients treated for recurrence were 29.4%, which is comparable to rates of 23% to 43% reported in the current literature for surgical salvage of recurrent oral cavity or oropharyngeal tumors. [16][17][18][19][20][21] Postoperative Complications After TO+LP Postoperative complication rates in the current study compared favorably with other studies, despite the heavily pretreated patient population. In a study involving the use of the lateral pharyngotomy approach in a cohort of previously untreated T1 and T2 tumors of the pyriform sinus, Holsinger et al report 1 (3%) postoperative complication of death related to salivary fistula. 5 Most other studies report postoperative complication rates after lateral pharyngotomy between 26% and 39% 6,9,22 which is consistent with our rate of 32%. It is notable that our rate of complications remains consistent with these other studies despite the increased number of advanced tumors and prior radiotherapy in our patient cohort. Within our cohort, the rate of operative complications, such as infection, fistula, or chyle leak, was 25.8%, with only a 19.4% rate of complications specifically related to pharyngeal resection (aspiration pneumonia). Patients treated at our institution were discharged after a median of 9 days (range 2-18), compared to a range of 12 to 19 days reported in other studies. 5,9,12,22  Treatment of recurrent head and neck cancer is challenging and can be associated with substantial morbidity. Postoperative complication rates of 17% to 67% in this setting have been reported, 20,21,[23][24][25][26] with postoperative admissions up to 12 days. 19 Comparatively, the complication rate in our cohort of patients with recurrent head and neck cancer and prior radiotherapy undergoing TO+LP was 35%, and the median postoperative admission length was only 8 days.

Functional Outcomes After Lateral Pharyngotomy
Duration of postoperative tracheostomy and enteral feeding are important indicators of functional outcome after pharyngeal tumor resection. Prior studies of the lateral pharyngotomy for resection of untreated, mostly T1 to T2 tumors report almost exclusively temporary tracheostomy and enteral feeding requirements. Median time to decannulation between 3 and 8 days 5,9,12,22 have been reported for patients whose tumor defect was able to be closed primarily. Lim et al describe a cohort of mostly T1 to T2 hypopharyngeal tumors that were treated with primary resection via lateral pharyngotomy where 70% of patients required reconstruction with a radial forearm free flap. 6 Median duration of tracheostomy dependence in this group of patients was 28 days (range 11-113), with 1 patient permanently tracheostomy dependent. This compares to our patient cohort, in which 29 out of 31 patients required free flap reconstruction, where the median time to decannulation was 22 days (range 6-100).
Other studies report average postoperative enteral feeding durations of 9 to 12 days, 5,9,22 with very few patients requiring g-tube. Bertolin et al report that 1 patient out of 64 treated primarily for early to intermediate-stage oropharyngeal carcinoma required g-tube placement due to significant swallowing dysfunction. In a Laccourreye et al paper, where only 7% of subjects had prior radiotherapy, 6.5% of patients required g-tube placement for up to 12 months. 22 While 96.8% of patients in the current study required enteral feeding at discharge, and 54.8% of patients still required enteral feeding at their first postoperative visit, it is important to highlight that these rates are representative of a heavily pretreated population. Only 5 out of 14 patients without prior radiotherapy required enteral feeding at the first postoperative visit. Furthermore, 12 of these 14 patients received adjuvant radiotherapy following TO+LP, which likely contributed to the decision to maintain enteral feeding.
Patients with a history of prior radiotherapy had longer median times to decannulation and a higher likelihood of requiring enteral feeding at the first postoperative visit and at the last follow-up. In our study, 41% of patients with prior radiotherapy were nondependent on enteral feeding at the time of the last follow-up. This compares favorably to other studies, which have reported long-term g-tube dependence rates of 30% to 65% after salvage surgery for oral cavity and oropharyngeal tumors. 19,24,27,28 While the TO+LP approach is applicable for many patients who have combined oral/oropharyngeal tumors or combined oropharyngeal/hypopharyngeal tumors, there are situations where a lip split with mandibulotomy or lingual release is preferred. For tumors involving the masticator space or for those requiring mandibular resection, we find the view afforded with the TO+LP approach is limiting and can compromise confident resection. Also, for tumors requiring a subtotal or total glossectomy, a full lingual release is employed. Ideal tumors for the TO+LP approach are those that involve the middle and posterior thirds of the tongue, or the lateral oropharynx without extension to the retromolar area and masticator space. In today's era, the TO+LP approach will be most applicable for patients who have failed TORS, TLM, or nonoperative management of human papillomavirus-associated oropharyngeal tumors and who require salvage resection with smaller volumefree tissue transfer.

Limitations
This study has several limitations. First, our cohort of patients was heterogeneous, with advanced or recurrent tumors requiring individualized approaches for management. We were, therefore, unable to include a direct matched comparison to other surgical approaches used at our institution, such as the mandibulotomy with lip split. Second, this was a retrospective study where outcomes are reported in a primarily descriptive manner, and we were limited by the information available in the medical record. In some cases, relevant information, such as radiation dose, complete histopathologic data, and speech and swallow follow-up, could not be obtained in full for each patient. Additionally, markers such as duration of tracheostomy and enteral feeding are incomplete ways of assessing function. Quantitative measures, such as the PSSHN and MDADI exist, but could not be used due to the retrospective nature of this study.

Conclusion
This study demonstrates that the combined TO+LP approach to oral and oropharyngeal tumor resection can be used for select patients with advanced or recurrent cancer with good functional and oncologic results. This includes its application to patients with recurrent disease as well as those with a history of prior radiation where the prognosis is generally poor. In the setting of surgical salvage, the TO+LP approach can be used to obtain similar margin status and postoperative complication rates as reported in studies of early disease. While our locoregional control after TO+LP for salvage was inferior to many studies describing outcomes after lateral pharyngotomy for early-stage disease, disease-free survival was comparable to rates reported in studies of salvage surgery for recurrent head and neck cancer. These results support the use of a combined TO+LP approach to resection and reconstruction in this subset of patients with recurrent and advanced oral and oropharyngeal malignancy.