A retrospective analysis of surgery in prestyloid parapharyngeal tumors: Lateral approaches vs transoral robotic surgery.

OBJECTIVE: Surgery is the standard treatment for most tumors in the prestyloid parapharyngeal space (PPS) but it can be a challenging procedure because of the anatomical complexity of the area. Prestyloid surgery can be performed with various lateral approaches or with a medial approach using transoral robotic surgery (TORS)-either alone or in combination with a transcervical incision. We have retrospectively compared our center's results with lateral and medial approaches.
METHODS: Between 2015 and 2020, 28 patients with prestyloid PPS tumors underwent surgery at our center: 14 with lateral approaches, including transcervical, transcervical-parotid, and transcervical-mandibular, and 14 with medial approaches (12 with TORS and two with TORS plus a transcervical incision). We compared surgical time, postsurgical complications, length of hospital stay, need for feeding tube, and relapse-free survival in the two patient groups.
RESULTS: Pleomorphic adenoma was the most frequent tumor and 60.7% of the tumors were benign. Tumor volume and maximum length were similar in the two groups of patients. Intraoperative image guidance and ultrasound were used in 33% of TORS. TORS was associated with less surgical time, fewer complications, and shorter hospital stays. Recurrence rates were similar in the two groups.
CONCLUSION: The medial approach by TORS offers superior results in prestyloid tumors than the open lateral approach and can be refined by intraoperative guidance. Level of Evidence: 4.

INTRODUCTION

The parapharyngeal space (PPS) is located lateral to the upper pharynx and medial to the mandible, from the hyoid bone caudally to the skull base, connecting posteromedially with the retropharyngeal space and inferiorly with the submandibular space. It is an inverted pyramid‐shaped space divided into the prestyloid and retrostyloid components by the tensorvascular‐styloid fascia, running posteriorly from the styloid process to the tensor veli palatini muscle. PPS lesions account for only 0.5% to 1% of head and neck tumors, are generally located in the prestyloid, and are mostly benign. , Surgery is the mainstay of treatment but can be challenging due to the PPS anatomical complexity and dense neurovascular network. , , It is thus essential to select the surgical approach with the least risk of postsurgical complications and the greatest possibility of complete resection, especially in malignant tumors.

Surgical approaches for prestyloid tumors can be classified as lateral or transoral/medial. Multiple lateral approaches have been used alone or in combination, , including transcervical, transcervical‐parotid, and transcervical‐mandibular approaches. In contrast, until relatively recently, the medial approach was limited to small tumors near the constrictor muscle, due to the limited visualization of the PPS with this approach. With a medial approach—natural orifice transluminal endoscopic surgery—the tumor can be approached either endoscopically or via transoral robotic surgery (TORS). However, over the last few years, TORS has become more prevalent. TORS provides a high‐definition visualization of the PPS and allows access to narrow areas, enabling resection due to precise tissue manipulation and tremor filtering without line‐of‐sight limitation but lacking haptic feedback. Although TORS cannot usually benefit from navigational systems due to soft tissue modification in surgical positioning, it is now recognized as an important surgical option—either alone or in combination with a transcervical approach.

Here we report our retrospective analysis of two surgical approaches used in our center for the resection of prestyloid tumors. We compare results using TORS assisted by navigational systems with those using lateral approaches in a series of 28 patients.

MATERIALS AND METHODS

Study design and patient inclusion

This was a retrospective analysis of prospectively collected data on 28 patients with prestyloid tumors undergoing surgery at our hospital from February 2015 to February 2020. All patients underwent a clinical examination, magnetic resonance imaging (MRI) and computed tomography. A fine‐needle aspiration was taken transorally or transcervically.

The use of the database was approved by the hospital ethics committee (PI‐17‐267) and all individuals included in the study provided their signed informed consent, in accordance with the Declaration of Helsinki.

Surgery

Lateral approaches

Between 2015 and 2017, prestyloid surgery at our center was based on a lateral approach. The transcervical approach involves a 4‐cm transverse skin incision; once the digastric posterior belly and the stylohyoid are recognized, the stylomandibular ligament is cut to improve exposure. The dissection continues toward the head, and the inferior and posterior PPS can be exposed. The trancervical‐transparotid approach identifies the facial nerve and reaches the deep parotid lobe and the lateral prestyloid space. The transcervical‐transmandibular approach adds a mandibular osteotomy, making it possible to reach the upper PPS. Facial nerve intraoperative neuromonitoring using the AVALANCHE system (Dr. Langer Medical GmbH, Waldkirch, Germany) with four channels was used with these lateral approaches.

TORS

Since 2018, prestyloid surgery at our center has been based on a medial approach with TORS—either alone or in combination with a transcervical incision when necessary—using the da Vinci Xi (Intuitive Surgical, Inc., Sunnyvale, California) docked at the patient's left side. The oropharyngeal cavity is accessed with the Feyh‐Kastenbauer retractor modified by Wenstein‐O'Malley (Olympus Corp). Surgery is performed with a slightly tilted head‐up position. A single dose of dexamethasone 4 to 8 mg is administered to avoid postoperative airway edema. Prophylactic tracheostomy is not a usual procedure. Patients with postoperative edema or uncertain airway safety remain intubated for 24 hours; if the situation does not improve, a temporary tracheostomy is considered.

TORS starts with 0‐degree scope and the Endowrist TM instruments (Intuitive Surgical, Inc) to approach the middle and inferior PPS. To approach the superior PPS we use the 30‐degree scope. The initial incision is made in the pterygomandibular raphe; once the palatoglossus and constrictor muscles are dissected, the prestyloid PPS is exposed by blunt dissection. When there is lateral extension, the medial pterygoideus muscle is cut to expand the axial axis of the surgical corridor. When a lateral extension cannot be controlled with TORS alone, it is combined with a transcervical incision (Figures 1 and 2) to dissect the lateral edge of the tumor.

FIGURE 1

Decision‐making tree for surgical approaches in prestyloid PPS tumors. ICA, internal carotid artery; TO, transoral

FIGURE 2

MRI axial and coronal slices showing prestyloid PPS tumors. (A,B) Independent parotid gland tumors treated with TORS. (C,D) Prestyloid tumor with lateral extension due to primary parotid involvement treated with TORS in combination with a transcervical approach

We use a navigation system based on ultrasound guidance using a 20 MHz disposable Doppler probe (Mizuho Medical Co. Ltd.) to identify vascular structures, entered through the patient's mouth by the first assistant and then guided by the surgeon with the Maryland instrument.

Swallowing assessment

Prior to hospital discharge, functional assessment of swallowing is performed in the dysphagia unit of our center by fiberoptic endoscopic evaluation of swallowing (FEES) and/or video fluoroscopy. Patients are evaluated as to whether they should receive a normal diet, diet adaptation, or swallowing rehabilitation or if they should keep the feeding tube. The same evaluation is performed at 3 and 6 months after surgery.

Data collection and statistical analyses

We collected data on tumor extension, nasogastric tube and/or gastrostomy, length of hospitalization, complications, need for adjuvant treatment, and swallowing. Surgical margins were considered free if complete extracapsular dissection was achieved in benign tumors, or the margins were 1mm tumor‐free in malignant tumors. Descriptive statistics were used to summarize data. The Chi‐square test or the Fisher's exact test, as appropriate, was used to determine an association between categorical variables. The one‐way analysis of variance or the independent sample t‐test was used to determine differences between means. We tested for normality using the Shapiro‐Wilk test of normality and for homogeneity of variances using Levene's test. The Kruskal‐Wallis or the Mann‐Whitney U nonparametric tests were used when the normality assumption was not met. Survival curves were estimated using the Kaplan‐Meier method. All statistical analyses were performed with Stata, version 14.2 software (StataCorp LLC). Significance was set at P ≤.05.

RESULTS

Patient characteristics

The study included 28 surgically resected patients with prestyloid PPS tumors (Table 1). Between 2015 and 2017, 14 patients were treated with a lateral approach. Between 2018 and 2020, 14 were treated with a medial approach. In surgery with TORS, the use of ultrasound guidance helped to identify the main vascular structures in the vicinity of the parapharyngeal tumor. Thirteen tumors (46.4%) originated in the deep lobe of the parotid gland. Seventeen (60.7%) were benign. There were no significant differences in mean tumor size between patients treated with a lateral approach and those treated with a medial approach (21.6 ± 3.2 cm3 vs 17.15 ± 2.5 cm3, respectively).

TABLE 1

Patient and tumor characteristics and surgical outcomes in 28 patients with prestyloid tumors

	All patients	Lateral approach	Medial approach
	N = 28	N = 14	N = 14
	N (%)	N (%)	N (%)
Age, y
Median (range)	60.18 (34‐85)	62.36 (42‐85)	58 (34‐76)
Sex
Male	15 (53.57)	8 (57.1)	7 (50)
Female	13 (46.43)	6 (42.9)	7 (50)
Surgical approach	—
Transcervical	1 (3.57)	1 (7.14)	—
Transcervical‐transparotid	11 (39.28)	11 (78.57)	—
Transcervical‐transmandibular	2 (7.14)	2 (14.29)	—
TORS alone	12 (42.85)	—	12 (85.71)
TORS + transcervical	2 (7.14)	—	2 (14.29)
Pathology of tumor
Pleomorphic adenoma	8 (28.57)	6 (42.85)	2 (14.28)
Warthin's tumor	1 (3.57)	1 (7.14)	—
Carcinoma ex pleomorphic adenoma	2 (7.14)	1 (7.14)	1 (7.14)
Adenoid cystic carcinoma	1 (3.57)	1 (7.14)	—
Salivary duct carcinoma	1 (3.57)	1 (7.14)	—
Schwannoma	1 (3.57)	—	1 (7.14)
Hemangioma	2 (7.14)	—	2 (14.28)
Lipoma	1 (3.57)	—	1 (7.14)
Histiocytoma	1 (3.57)	—	1 (7.14)
Metastasis of squamous cell carcinoma	7 (25.0)	2 (14.28)	5 (35.7)
Myoepithelioma	1 (3.57)	1 (7.14)	—
Lymphoepithelial lesion	1 (3.57)	—	1 (7.14)
Liposarcoma	1 (3.57)	1 (7.14)	—
Type of tumor
Benign	17 (60.71)	9 (64.29)	8 (57.14)
Malignant	11 (39.29)	5 (35.71)	6 (42.86)
Mean tumor size, cm3	19.4	21.66	17.15
Mean max tumor length, cm	3.75	3.6	3.9
Need for feeding tube	8 (28.6)	4 (28.6)	4 (28.6)
Mean duration of surgery, min (range)	182.9 (40‐340)	221.4 (100‐340)	133.3 (40‐240)
Mean length of hospital stay, days	7.32	8.9	5.8
Rate of postsurgical complications	39.3%	64.3%	14.3%
Free margin rate	78.6%	78.6%	78.6%
2‐year relapse free survival rate	74.1%	69.3%	78.6%

Outcomes

The mean duration of surgery was longer for lateral than for medial approaches (P = .02; Table 1). Postsurgical complications were more frequent with lateral approaches (P = .01), with the highest rate (72%) associated with the transcervical‐transparotid approach, due to postoperative facial palsy. Three patients had transient facial palsy, while five had permanent facial palsy. These complications were not associated with more malignant tumors (Table 2). The only complications associated with TORS were one case of cervical hematoma and one of pharyngeal wound dehiscence with temporary velopharyngeal incompetence (Tables 1 and 2). The free margin rate was 78.6% for both approaches.

TABLE 2

Postsurgical complications for all patients and according to the surgical approach used

Complication	All patients	Lateral approach	Medial approach
	N = 28	N = 14	N = 14
	N (%)	N (%)	N (%)
None	17 (60.7)	5 (35.7)	12 (85.7)
Wound dehiscence	1 (3.6)	1 (7.1)	0
Transient facial palsy a	3 (10.7)	3 (21.4)	0
Permanent facial palsy b	5 (17.8)	5 (35.7)	0
Cervical hematoma	1 (3.6)	0	1 (7.1)
Pharyngeal wound dehiscence	1 (3.6)	0	1 (7.1)

Of the three patients treated with a lateral approach who had transient facial palsy, two had benign tumors and one had a malignant tumor.

Of the five patients treated with a lateral approach who had permanent facial palsy, four had benign tumors and one had a malignant tumor.

Four patients in each group required a feeding tube. The mean time in those treated with a lateral approach was 2.6 days, while the two undergoing transcervical‐transmandibular surgery required a tube for a mean of 13 days. The mean time in those treated with a medial approach was 1.9 days, while the patient undergoing TORS + transcervical required a tube for 5 days. The swallowing assessment in the patients who did not require a feeding tube showed a safe and effective deglutition at hospital discharge. The patients who required a feeding tube received swallowing rehabilitation and adapted diet until 3 months postsurgery. All the FEES at this time were normal and patients could receive a normal diet. The FEES evaluation at 6 months postsurgery showed completely normal and safe swallowing in all patients.

The mean length of hospital stay overall was similar with both approaches (P = .19; Table 1). Among patients treated with a medial approach, hospital stay was longer in those with postsurgical complications than in those without (12 vs 4.75 days; P = .01). With a lateral approach, length of hospital stay was similar for those with and without complications (9 vs 8.6 days; P = .93).

With a median follow‐up of 38 months (range, 10‐60), there were no significant differences in 2‐year relapse‐free survival rates between the two groups (Table 1).

DISCUSSION

The main objectives of surgical excision of prestyloid tumors are to relieve symptoms, prevent imminent morbidity, and avoid malignant transformation, especially in pleomorphic adenomas. However, there are often complications stemming from the anatomical complexity of the area. The ideal surgical approach should allow enough room to achieve a complete resection while still controlling the nearby neurovascular structures and managing possible complications.

The direct pathway to the prestyloid is a medial approach, which has been used since the 1950s. Until recently, however, this approach was controversial since limited visibility and a narrow working area often led to increased tumor spillage and recurrence and to the risk of neurovascular injury. TORS was first described in 2007 and has since evolved as a valid approach, with improved visual magnification and the precision of robotic instruments. Today, TORS is able to open a wide corridor in the middle PPS, thus minimizing the risk to neurovascular structures. However, the use of TORS is less than optimal in tumors located in the far lateral and posterior areas of the PPS, and in cases with lateral extension due to parotid gland primary involvement, it can be combined with a transcervical approach. , ,

While some decision‐making algorithms for choosing between different lateral approaches have been proposed, there are as yet no clear guidelines for deciding between lateral, medial, or combined approaches. We have developed a decision‐making algorithm for surgical approaches to prestyloid tumors based on the position of the internal carotid artery and parotid involvement on the lateral tumor extension (Figure 1). Based on this algorithm, we now use a medial approach in the vast majority of prestyloid tumors.

The skills and experience of the surgical team are clearly crucial when selecting the best approach to prestyloid tumors. In this study, the same surgical team used both the lateral and medial approaches, though at different time periods. We found no differences in tumor size, pathology, or patient age between the two groups of patients, and in line with previous reports, , tumors in both groups were primarily benign salivary gland tumors. The medial approach proved superior in terms of duration of surgery, length of hospital stay, and postsurgical complications. The lateral approach required a longer surgical time, even when considering the extra time needed for docking the da Vinci system in the medial approach, most likely because the medial approach provides the most direct corridor to the prestyloid PPS. , Length of hospital stay was shorter with a medial approach, perhaps due to the more direct route and fewer associated complications, , , although we cannot rule out a potential time bias since the two groups were treated in different years. Postsurgical complications were less frequent with the medial approach, possibly due to less structure dissection.

The most common lateral approach for prestyloid tumors is transcervical, , but this is not ideal if the tumor is in the upper PPS. The transcervical approach is often used in conjunction with a transparotid approach when the tumor is located in the deep lobe of the parotid, though this increases the risk of facial nerve involvement. Previous studies have reported a complication rate of 20% to 40% with these approaches. , A transmandibular approach, combined with the transcervical approach, was first proposed for patients with suspected malignant, very large, or vascular tumors and is now used in 3% to 40% of cases, , although it is associated with increased complications. , , This approach may require a covering tracheostomy, longer hospital stay, and delay in oral nutrition, and carries additional risks of mandibular dehiscence, temporomandibular joint dysfunction, and loss of dentition. These combined lateral approaches achieve sufficient working space and vascular control but are associated with high rates of postsurgical complications. , In our series, we observed complications in 64.3% in patients treated with a lateral approach, mainly due to facial nerve mobilization associated with the transcervical‐transparotid approach, which led to permanent facial palsy in 35% of cases. The transcervical‐transmandibular approach was associated with a longer hospital stay and greater need for a feeding tube. In contrast, TORS has been reported to have only 7% to 17% complications, with hematoma and pharyngeal dehiscence being the most common. In our series, the medial approach was associated with complications in only two cases: one pharyngeal dehiscence with temporary velopharyngeal incompetence and one cervical hematoma.

The disadvantages of TORS are the lack of haptic feedback and the limited space in the deep PPS without control of carotid vessels. However, image‐guided surgery, though not yet in widespread use, is improving these conditions and is a promising tool for reducing complications associated with TORS. , , In our practice, we use ultrasound intraoperative guidance. It is based on echography guided by a Doppler 20 MHz probe with 1 cm of tissue penetration to identify the external branches of the facial and lingual arteries and the internal carotid artery. This system helps to identify and protect the vascular structures surrounding the tumor.

Based on our good surgical results with medial approaches, we can recommend them for selected prestyloid PPS tumors. Those with the internal carotid artery in a lateral position can benefit from a medial approach with TORS. When there is a lateral extension due to a primary parotid tumor, a combined approach can be used to reach the external area. If this prolongation does not involve the parotid gland, TORS alone can be considered.

CONCLUSIONS

Surgery for prestyloid tumors is evolving in parallel with technology. At present, however, despite multiple reports, there are no specific guidelines for selecting TORS or lateral approaches. In the present single‐center study of two surgical approaches to these tumors, we have found better surgical results and lower morbidity with medial approaches using TORS alone or TORS plus a transcervical approach. TORS can be improved with intraoperative image guidance, and further investigation of these techniques is warranted.

CONFLICT OF INTEREST

The authors declare no potential conflict of interest.