Clinical Outcomes with Transoral and External Surgical Approaches for Resection of Parapharyngeal Space Tumors: Systematic Review and Meta-Analysis.

Key Message
Transoral approaches may be associated with fewer neurological complications and improved perioperative outcomes compared to external approaches (EAs) in the surgical treatment of PPST.

There were no significant differences in postoperative complications between the different types of transoral approaches.

Transoral approaches are best reserved for carefully selected patients in experienced centers.

Introduction
The parapharyngeal space (PPS) is an anatomically distinct region shaped like an inverted pyramid extending from the base of the skull to the level of the hyoid bone on both sides of the pharynx, bordered by the carotid sheath and the retropharyngeal space.
1
The region has been anatomically subdivided into the prestyloid (or true PPS) and the poststyloid parapharyngeal (or carotid space) spaces.
1

PPS tumors (PPSTs) are a rare subset of head and neck neoplasms that account for 0.5% to 1% of all head and neck lesions, with 80% to 90% being benign.2
-4 Nonetheless, they are clinically important from a diagnosis and management perspective due to their histological diversity and anatomy.1
-3,5
-7 This deep neck space is challenging to access due to its medial location within the head and neck and contains many vital structures, such as cranial nerves and major vessels.
1

Although surgical intervention is the standard treatment for most PPSTs,
3
different techniques may be used depending on the precise location of the PPST, its size, vascularity, histological characteristics, and the patient’s or surgeon’s preference. External surgical approaches (EAs) are most commonly used for tumors of the parapharynx, utilizing techniques such as transparotid, transmandibular, infratemporal fossa, or transcervical.
8
While transcervical and combined transcervical-parotid approaches are preferred for resecting most PPSTs, the transoral approach has recently been introduced as an alternative for selected tumors, offering a minimally invasive option.
9
The transoral approach involves accessing PPSTs through the oropharynx, thus minimizing external incisions and visible scarring.
9
Transoral approaches may include both endoscopic-assisted transoral approach (EATA)
10
and transoral robotic surgery (TORS).
11
EATA and TORS are increasingly becoming the preferred surgical approach and may provide comparable surgical outcomes.
9

While previous systematic reviews have evaluated transoral approaches, their scope focused on a single transoral modality. For instance, De Virgilio et al
12
evaluated the clinical characteristics and outcomes of PPST treated exclusively by TORS, while Chen et al
13
compared EATA to EAs, excluding robotic techniques. A more recent systematic review and meta-analysis by Faisal et al
14
focused primarily on neurological complications in benign PPST without a comprehensive comparison of surgical approaches.
The current systematic review and meta-analysis aim to address these gaps by comparing clinical outcomes, including postoperative complications (dichotomized as neurological and non-neurological), recurrence rates, operation time, hospitalization time, and intraoperative bleeding, across a broader spectrum of transoral techniques versus EAs. We also compared the clinical outcomes with various transoral approaches, such as conventional transoral, EATA, and TORS. Stratified analyses by tumor location and pathology were also performed.

Materials and Methods
This systematic review and meta-analysis was conducted following the PRISMA reporting guidelines. The protocol for this study was preregistered with PROSPERO (registration number CRD42024580626).

Search Strategy
Embase, MEDLINE, CINAHL, Cochrane, and Web of Science databases were searched from inception until August 2024 for relevant literature. With the support of an academic librarian, we devised a search strategy that combined relevant keywords, including surgery, parapharyngeal space tumors, and complications, with the Boolean expressions “AND” and “OR.” Furthermore, the database search was supplemented by a thorough review of the bibliographies of all included studies and relevant documents. The full search strategy developed to identify relevant articles from the above-mentioned databases is shown in Supplemental Table 1.

Study Screening and Selection
We utilized Covidence (Veritas Health Innovation, Melbourne, Australia), a systematic review management tool, to facilitate the study selection process. The titles and abstracts of the identified studies were independently reviewed by 2 reviewers (A.S.A. and M.C.G.). A full-text review of the shortlisted articles was used to determine their eligibility. Any disagreements or disputes during the selection process were resolved by discussing with the senior authors. Studies were eligible for inclusion if they adhered to the following PICO criteria: Patients (P): Adults diagnosed with any PPSTs; Intervention (I): Surgical approaches such as the transoral approach, transcervical approach, transcervical-parotid approach, transmandibular approach, infratemporal fossa approach, TORS, EATA, or a combination of any surgical approaches; Comparison (C): Direct and indirect comparisons between different surgical approaches; and Outcomes (O): Postoperative complications, recurrence rates, intraoperative bleeding, operation time, and hospitalization time. Research studies categorized as conference abstracts, letters to the editor, literature reviews, cadaveric studies, narrative reviews, case reports, case studies with fewer than 5 patients, and commentaries were excluded. Studies involving pediatric patients, those evaluating non-surgical treatment methods for PPST, or authored in a non-English language were also excluded.

Data Extraction
Two reviewers (A.S.A. and M.C.G.) independently extracted data required for analysis using a standardized Excel spreadsheet. Data extracted included the surname of the primary author, publication year, study location (Country), study design, pertinent characteristics of enrolled participants such as sample size, the type of PPST and mean age of enrolled participants, the follow-up duration, surgical approaches, postoperative complications, recurrence rates, operation time, intraoperative bleeding, and hospitalization time.

Study Outcomes
The primary outcome of this study was postoperative complications. Secondary outcomes were operation time, intraoperative bleeding, hospitalization time, and recurrence rates. The postoperative complications were further categorized into neurological (eg, cranial nerve deficits, Horner’s syndrome, first bite syndrome, etc.) and non-neurological complications (eg, wound infection, hemorrhage, seroma, etc.).

Study Quality Assessment
The methodological quality of the studies eligible for inclusion in the current systematic review and meta-analysis was assessed using the Newcastle–Ottawa Scale.
15
The Newcastle–Ottawa Scale enables studies to be independently assessed based on their selection, comparability, and outcomes. A maximum of 1 point was assigned for every criterion fully addressed. Otherwise, no point was assigned. The overall methodological quality of each study was categorized as poor (score 0-3), fair (score 4-6), or good (score 7-9).

Data Analysis
The STATA 18 (StataCorp, College Station, TX, USA) software was used to compare the rate of postoperative complications between transoral surgical approaches and EAs in a 1-group meta-analysis, for which the overall effect size was presented in terms of the event rate (ER). For each group, ER was defined as the number of events per total patients. Postoperative complications were grouped according to the tumor behavior (benign or malignant) and tumor location (prestyloid or poststyloid). Furthermore, a test for subgroup differences was used to compare the outcomes of EAs and transoral approaches, as well as between the different transoral approaches. Review Manager (RevMan version 5.4.1. The Cochrane Collaboration, version 2014) was used to compare operation time, intraoperative bleeding, and hospitalization time between transoral surgical approaches and EAs, with the overall effect size presented in terms of the mean difference (MD).
All pooled outcomes were provided with their corresponding 95% confidence intervals (CIs). Interstudy heterogeneity was assessed using I2 statistics, with values exceeding 50% considered high.16,17 The DerSimonian–Laird random-effects model was employed whenever significant heterogeneity was observed to provide a conservative estimate of the overall effect. Otherwise, the fixed-effects model was used for analysis.

Results

Search Results
A total of 505 records were identified from the electronic database search: 258 from Embase, 133 from Medline, 101 from Web of Science, and 13 from CINAHL. Of these, 174 duplicate records were excluded, and another 237 were excluded for having irrelevant titles and abstracts. The remaining 94 records were subjected to full-text review, from which we excluded 2 abstracts, 1 study with the wrong intervention, 8 published in different languages, 1 study involving the pediatric population, 6 articles without full texts, 8 case series with less than 5 individuals, and 20 that did not assess outcomes based on the individual surgical approaches. Ultimately, 48 records met the inclusion criteria.10,18
-64 The PRISMA flowchart for the study screening and selection process is presented in Figure 1.

Characteristics of Included Studies
The 48 included studies involved 1728 patients aged 29.7 to 66.3 years; 1418 had benign, and 155 had malignant PPST. The follow-up duration ranged between 6 and 180 months. All the studies were published between 1995 and 2024 (Table 1). The largest number of included studies was conducted in China (n = 10), followed by the United States (n = 8), Italy (n = 7), and India (n = 6). Two studies each were conducted in France, Korea, Mexico, and Greece, while 1 study each was conducted in Iran, Turkey, the Czech Republic, Germany, Israel, Taiwan, Japan, and Spain. One study did not specify the location where it was carried out.
24
In patients undergoing transoral surgical excision, the most common pathologies of PPST were pleomorphic adenoma and schwannoma; while for patients treated by EAs, the most common pathologies were pleomorphic adenoma, schwannoma, and paraganglioma (Supplemental Table 2).

Complications
Facial nerve deficits, palsy, paralysis, or injury (n = 12) were the most reported neurological complications in the included studies, followed by Horner’s syndrome (n = 9), cranial nerve deficits/palsy/paralysis/injury (n = 8), vocal cord paralysis/palsy (n = 6), marginal mandibular nerve palsy/paralysis/dysfunction (n = 6), first bite syndrome (n = 5), and Frey’s syndrome (n = 4). Common non-neurological complications included wound complications (including dehiscence and infection; n = 10), hematoma (n = 7), trismus (n = 6), and hemorrhage (n = 6). The complications reported in the included studies are detailed in Table 2.

Primary Outcomes: Transoral Approaches Versus EAs
Pooled results from 31 studies show that the incidence of postoperative complications was significantly lower in PPST patients treated with transoral approaches [ER: 0.11, 95% CI (0.08-0.15)] compared to those treated with EAs [0.40 (0.32-0.48); P < .0001; Table 3 and Supplemental Figure 1].
In the subgroup analysis, we observed a significantly higher incidence of neurological complications in patients treated with EAs compared to those treated with transoral approaches [0.30 (0.22-0.38) vs 0.05 (0.03-0.07); P < .0001; Table 3 and Supplemental Figure 2]. However, there was no significant difference in non-neurological complications between EAs and transoral approaches (Table 3 and Supplemental Figure 3). High inter-study heterogeneity was detected for neurological and non-neurological complications with EA (I2 = 89.21% and 67.79%) but not for transoral approaches (I2 = 20.71% and 31.11%).
Furthermore, after subcategorizing the postoperative complications according to tumor behavior, we found that transoral approaches were associated with significantly fewer postoperative complications than EAs when used to excise benign tumors [0.12 (0.22-0.35) vs 0.39 (0.29-0.48); P < .0001] and malignant [0.17 (−0.08-0.41) vs 0.63 (0.30-0.95); P = .03] tumors (Table 4 and Supplemental Figures 4 and 5). High inter-study heterogeneity was detected for postsurgical complications in both subgroup analyses with EA (I2 = 83.54% and 61.73%) but not transoral approaches (I2 = 46.42% and 0%).
When categorized by tumor location, we noted lower incidences of postoperative complications with transoral approaches, but only for tumors in the prestyloid compartment [0.11 (0.06-0.17) vs 0.39 (0.22-0.05); P = .001; Table 4 and Supplemental Figure 6]. High inter-study heterogeneity was detected for postsurgical complications with EA (I2 = 91.38%) but not with transoral approaches (I2 = 6.59%). No significant difference in postoperative complications was observed for poststyloid PPST excised using the transoral approaches and EAs, albeit with high inter-study heterogeneity (I2 = 68.78% and 56.49% with the 2 approaches, respectively; Table 4 and Supplemental Figure 7).
A subgroup analysis of recurrence rates among PPST patients showed no significant difference between transoral approaches and EAs [0.03 (0.02-0.05) vs 0.02 (0.01-0.03); P = .35; Table 5 and Supplemental Figure 8]. Low interstudy heterogeneity was observed in both EAs and transoral approaches groups (I2 = 16.41% and 0%, respectively).

Primary Outcomes: Comparison of Different Transoral Approaches
Our results showed no significant difference in postoperative complications between EATA, TORS, and the conventional transoral approach (Table 6 and Supplemental Figure 9). While low inter-study heterogeneity was detected for EATA (I2 = 0%) and conventional transoral (I2 = 37.01%) approaches, high heterogeneity was detected for TORS (I2 = 53.78%). Similarly, sub-group analysis by neurological or non-neurological complications did not show a statistically significant difference between EATA, TORS, and the conventional transoral approach (Table 6 and Supplemental Figures 10 and 11). Low heterogeneity was detected for all analyses except neurological complications with TORS (I2 = 61.17%).
In addition, our results demonstrated no significant difference in recurrence rates between EATA, TORS, and the conventional transoral approach [0.05 (0.01-0.08) vs 0.03 (0.00-0.06) vs 0.03 (0.00-0.06); P = .73; Table 5 and Supplemental Figure 12]. A low interstudy heterogeneity was also observed across all the subgroups (I2 = 0%)

Secondary Outcomes
The pooled results revealed that transoral approaches were associated with significantly lower intraoperative blood loss (MD: −104.30 mL; P < .00001) and a shorter length of hospitalization (MD: −1.70 days; P = .002) compared to EAs. However, no significant difference in operation time was observed between the transoral approaches and EAs groups (MD: −12.66 minutes; P = .28).

Quality of Included Studies
The summary of the quality appraisal of the included studies using the Newcastle Ottawa Scale is presented in Supplemental Table 3. Overall, all studies had fair methodological quality. Regarding the representativeness of the exposed cohort, we did not assign any stars to the studies, as they were all single-center studies. Also, none of the studies adjusted for confounders and thus did not receive a score under the comparability domain.

Discussion
The management of PPST is challenging due to the complex anatomy, proximity to neurovascular structures, and diverse pathology.65
-67 Nonetheless, surgical excision has been the preferred treatment for over 95% of patients with PPST.2
-4 The choice of approach often requires careful consideration to ensure complete tumor resection while minimizing the risk of morbidity. Therefore, the current meta-analysis investigated whether surgical outcomes differ between transoral approaches, including the conventional transoral approach, EATA, and TORS, and the conventional EAs.
We used broader inclusion criteria than those in earlier meta-analyses investigating postoperative outcomes with surgical approaches to manage PPST, allowing us to pool data from a larger number of studies (n = 48). For comparison, De Virgilio et al
12
included 22 studies to investigate the clinical outcomes of PPST treated by TORS, Chen et al
13
included 7 studies to compare the clinical outcomes with EATA and EAs, and Faisal et al
14
pooled data from 13 studies to identify the determinants of postoperative neurological complications of benign PPST with surgical approaches in general.
As a result, we categorized and compared the overall, neurological, and non-neurological outcomes of patients with PPST after surgical interventions using transoral and external approaches. This categorization provided a clearer definition of the distinct and complex anatomy in which critical neurovascular structures are located.
1
Given that neurological complications can have serious implications on long-term morbidity, functional outcomes, and quality of life,
5
delineating complications that affect neural structures, such as the cranial nerves, the sympathetic chain, and other major vessels, can help correlate clinical relevance with our findings. Our analysis shows that transoral approaches were associated with a lower incidence of postoperative complications than EAs. Specifically, we noted a lower incidence of neurological complications with transoral approaches. In line with our findings that transoral approaches may offer a safer alternative to EAs in preserving neurological function, earlier studies by Ducic et al
31
indicate that transoral approaches require less dissection of the neural structures.
We also conducted a subgroup analyses of postoperative complications based on tumor classification. When categorized by tumor behavior, patients with both benign and malignant tumors responded more favorably in terms of the rate of overall postoperative complications with transoral approaches than with EAs. This is an important finding, as many clinicians consider transoral approaches, such as TORS, unsuitable for excising malignant tumors,12,51 perhaps due to the greater risk of capsular violation with malignant compared to benign tumors. Chan et al
68
noted that 24% of the patients with pleomorphic adenomas in the PPS had unintended capsule violation or tumor fragmentation with TORS based on a case series of 44 patients. On the other hand, Boyce et al
22
reported 3 cases (17.6%) of capsular rupture, 1 involving basal cell adenoma and 2 involving pleomorphic adenoma, among 17 patients with PPST treated with TORS. However, in a more recent case series of 53 patients with PPST, Chu et al
29
used TORS to surgically manage 7 patients, of which 2 had malignant tumors, without any cases of capsule violation or tumor fragmentation. In comparison, capsule violation was reported in 1 of the 22 patients treated with the transcervical approach.
Moreover, malignant tumors tend to invade surrounding neurovascular structures, such as the carotid artery and cranial nerves, or spread to adjacent compartments; therefore, they often require extensive resection to attain clear margins, thus increasing the risk of complications regardless of surgical approach.
1
While our analysis did not specifically assess capsule violation, it does provide robust evidence for the lower risk of complication with transoral approaches for both benign and malignant PPST. We did not observe any differences in overall, neurological, or non-neurological postoperative complications between the different transoral approaches (ie, conventional, EATA, TORS).
Furthermore, we conducted a subgroup analysis by the anatomical site of the tumor, which indicated that transoral approaches were associated with lower rates of postoperative complications compared to EAs when used to excise prestyloid tumors. However, no difference in postoperative complications was noted between the 2 approaches for poststyloid tumors. The poststyloid compartment contains critical neurovascular structures, such as the carotid artery, internal jugular veins, cranial nerves IX-XII, and the sympathetic chain.
1
Therefore, tumors in this compartment necessitate resection closer to or around these critical structures, increasing the risk of complications.
Among our secondary outcomes, transoral approaches were associated with significantly lower intraoperative blood loss and shorter hospitalization duration than EAs, albeit with comparable intraoperative duration. These findings are generally consistent with earlier studies. A systematic review by Chen et al
13
reported lower blood loss (−89.02 mL vs −104.30 mL in the current study) and shorter duration of postoperative hospitalization (−2.44 days vs 1.70 in the current study) with EATA compared to EA. However, in contrast to our study, Chen et al
13
also reported significantly shorter operative time with EATA (−5.56 min) than EA. It is important to note that while Chen et al
13
compared EATA with EA, the current study compared all transoral approaches with EA. In another systematic review, Chan et al
68
reported a mean operative time of 157 minutes, a blood loss of 58.2 mL, and a mean length of hospitalization of 3.0 days for TORS. However, they did not compare these outcomes with EAs or any other transoral approaches.

Implications of Our Findings
The conventional transoral approach is often discredited by many surgeons for the treatment of PPST, given its association with high risk of bleeding, cranial nerve damage, tumor spillage, and reduced exposure based on evidence from over 2 decades ago.69
-71 However, more recent evidence suggests that the transoral approach offers advantages, including the absence of visible neck scars, shorter postoperative hospital stays, and shorter operation times compared to EAs,33,35 which is further supported by our meta-analysis, except for operation time, which was not significantly different in our analysis. This indicates improved clinical outcomes with the transoral approaches over time, potentially due to factors such as technological advancement and experience with this modality for treating PPST.72,73 The choice of transoral surgical approach likely requires careful consideration of the surgeon’s expertise and tumor characteristics while incorporating patient-centered decision-making.
While the current meta-analysis did not compare tumor rupture between transoral and EAs, the risk of tumor rupture is a biologically and clinically important event in PPST because it has been associated with increased risk of recurrence, particularly in pleomorphic adenoma.
43
Lim et al found that intraoperative spillage of pleomorphic adenomas occurred in 13.6% of patients treated with conventional EAs and 15.4% of patients treated with TORS, suggesting that there is no significant difference between transoral and EAs in terms of tumor spillage.
43
However, due to the lack of comparisons in other included studies, future research should strive to evaluate this important event for better comparisons.
From a practical standpoint, our findings suggest that EATA and TORS might be suitable for the excision of benign or malignant PPST in select patients, particularly those with PPST located in the prestyloid region, and in centers with expertise in advanced transoral surgery. These findings align with the recommendations of previous meta-analyses, which noted that EATA
13
and TORS
12
are safe, minimally invasive, and aesthetic surgical modalities for PPST. However, patient selection is a crucial determinant of success with EATA surgery, and this may not be the preferred modality for patients who have PPST in the postsyloid compartment if the tumor location presents limited exposure space for visualization of the great vessels or if there is an increased risk of tumor spillage or neurovascular damage.
13
Furthermore, TORS may have additional disadvantages as the first-line treatment option for PPST, such as the surgeon’s expertise, lack of haptic feedback in older systems, availability, and higher cost.2,22,68,74

Strengths and Limitations
The strength of the current systematic review and meta-analysis lies in its high statistical power, which is attributed to the inclusion of a large number of studies with a pooled participant size of 1728. For comparison, earlier studies by Riffat et al,
2
Kuet et al,
3
De Virgilio et al,
12
and Chen et al
13
included 17, 22, 22, and 7 studies with 1143, 1293, 113, and 318 patients, respectively. Another key strength of our study is that we categorized postoperative complications according to the tumor type and location. Such granular analysis yielded valuable insights for patient-specific management and the development of evidence-based guidelines for the surgical management of PPST. Future research should include surveys of current practices and surgeon attitudes toward management of PPST, which would provide complementary insights to our findings.
The current study has several significant limitations. First, there may be publication bias in existing literature reporting favorable outcomes with technically advanced or minimally invasive approaches. Second, there is potential case-selection bias as smaller or less complex tumors are more often managed using transoral approaches. Similarly, surgeon and center-level bias likely exists, since transoral resections are typically performed in tertiary, high-volume centers with greater expertise in robotic and endoscopic surgery. Third, the number of patients and events in our recurrence analysis was small. This introduced a risk of type II (beta) error and limits the interpretation of recurrence rates. Fourth, the reliability of reporting on intraoperative variables, such as tumor rupture or spillage, is likely poor due to inconsistent documentation across studies.
Other notable limitations include that the studies included in our analysis were predominantly retrospective (n = 39/48), which introduces an inherent bias. Moreover, most studies included in our analysis had small sample sizes (mean: 36, 95% CI: 16.73-52.73). Thus, there is a need for more large-scale prospective level 1 evidence to guide evidence-based findings. Second, significant heterogeneity was noted in most statistical analyses. This heterogeneity might be attributed to differences in sample size, histopathological diagnosis, and surgical techniques. However, a random-effects model was used to pool these results and provide conservative estimates. Third, we used the prestyloid and poststyloid spaces to assess the PPST location; however, PPS anatomy is more complex and can be further divided into the pharyngeal mucosal space, the masticator space, and the parotid space.
1
Finally, we did not retrieve records published in languages other than English, which could have inadvertently led to selection bias.

Conclusion
The current meta-analysis showed that transoral surgical approaches resulted in fewer postoperative neurologic complications, less intraoperative bleeding, and shorter hospitalization duration than conventional EAs, especially for tumors located in the prestyloid compartment. These novel findings can inform surgical decision-making in selected patients, particularly when performed in centers with expertise in advanced transoral surgery while maintaining careful attention to individualized, patient-centered care.

Supplemental Material

sj-docx-1-ohn-10.1177_19160216251414008 – Supplemental material for Clinical Outcomes with Transoral and External Surgical Approaches for Resection of Parapharyngeal Space Tumors: Systematic Review and Meta-Analysis

Supplemental material, sj-docx-1-ohn-10.1177_19160216251414008 for Clinical Outcomes with Transoral and External Surgical Approaches for Resection of Parapharyngeal Space Tumors: Systematic Review and Meta-Analysis by Arshbir Aulakh, Mankirat Gillon, Brendan D. McNeely, Cameron Bakala and Ciaran Lane in Journal of Otolaryngology - Head & Neck Surgery