Clinical and Aesthetic Outcomes in Conventional and Minimally Invasive Submandibular Gland Excision Approaches

Abstract

This narrative literature review is an evaluation of the cosmetic, perioperative, and clinical outcomes of alternative approaches to submandibular gland excision, including conventional and robotic transoral excision, as well as endoscopic and robotic retroauricular excision, in comparison with the traditional transcervical approach. We intend to compare the cosmetic impact of each procedure based on the type of scar and location. PubMed and Google Scholar identified studies published between January 2000 and May 2025 which reported the outcomes of submandibular gland excisions performed using a transcervical, transoral, or retroauricular technique. Our analysis showed that the transcervical approach provides reliable access but carries risks of marginal mandibular nerve injury (18.7%) and visible cervical scarring. Conventional transoral excision avoids external scars and reduces the risk to the marginal mandibular branch and hypoglossal nerves but increases the risk of lingual nerve neuropraxia, with transient sensory deficits reported in 25–43% of patients. Robotic transoral excision preserves the sublingual gland and may improve completeness of resection, but operative times remain prolonged (<190 min). Endoscopic and robotic retroauricular excision can hide scars in the hairline and provide improved cosmetic satisfaction scores, with limitations including longer operative times and learning curves. Alternative approaches to submandibular gland excision provide improved cosmetic outcomes with comparable safety profiles to the transcervical approach.

1. Introduction

Excision of the submandibular gland (SMG) is the treatment of choice for chronic sialadenitis, sialolithiasis, and SMG neoplasms. The most commonly used approach for excision has long been a transcervical approach (TCA) that leaves a visible neck scar and places the marginal mandibular branch (MMB) at risk for injury. Demand for scarless surgery, or concealed scars and lower cranial nerve morbidity, has spurred two alternative approaches, a transoral and retroauricular approach. Transoral SMG excision eliminates an external neck scar, while retroauricular routes hide the incision in the occipital hairline. Transoral SMG excision can be performed through conventional methods or with robotic assistance. Retroauricular approaches are performed with endoscopic or robotic assistance. Early data suggest a transoral robotic approach achieves a more complete resection compared to a conventional transoral approach (TOA). Endoscopic or robotic retroauricular resections are equivalent to transcervical outcomes, with improved cosmetic satisfaction [1,2,3]. However, research describing these alternative approaches is lacking. Furthermore, insufficient understanding of operative techniques, potential for nerve injury, and concerns regarding procedural costs reduce surgeons’ willingness to adopt these surgical approaches. This review synthesizes evidence on conventional and robotic transoral excision, as well as endoscopic and robotic retroauricular approaches. The analysis evaluates cosmetic outcomes, perioperative results, complications, clinical indications, and resource utilization to support patient-centered operative planning.

2. Materials and Methods

A narrative literature review of Google Scholar and PubMed to identify cosmetic, perioperative, and clinical outcomes of alternative approaches to SMG excisions was conducted to identify studies published between January 2000 and May 2025. Search strategies combined MeSH terms and keywords for “submandibular gland” and “excision”, with additional terms specifying the approach, including “transoral”, “intraoral”, “retroauricular”, “hairline”, “cosmetic”, or “aesthetic”. Studies that reported cosmetic and clinical outcomes of SMG excisions performed through either transoral, transcervical, or retroauricular techniques were included for review; all other studies that did not explicitly describe the removal of the SMG, such as sialolithotomies or parotidectomies, were excluded. Additional articles were gathered by using the reference sections of these articles. Three reviewers screened all titles and abstracts and extracted full-text copies of potentially relevant articles. No formal quality assessment or quantitative synthesis was conducted, as this review aimed to provide a descriptive overview rather than a systematic evaluation. Disagreements between any two reviewers was mediated and resolved via discussion by the third reviewer.

3. Results

3.1. Conventional Transoral Approach

The TOA for SMG excision begins with an incision parallel to the Wharton canal from the anterior papillary region to the posterior retromolar trigone. While dissecting the papillary region to the SMG, the lingual nerve is identified and preserved, and the submandibular ganglion is separated from the SMG to free the lingual nerve from the gland. After the lingual nerve is retracted, external digital pressure is applied to the SMG to enhance surgical exposure, and the SMG is dissected from the inferior muscles. Throughout the process, branches of the facial artery and vein are identified and ligated to prevent hemorrhage or postoperative hematoma, and after the en-bloc excision, the hypoglossal nerve is identified. The floor-of-mouth mucosa is then closed [4].

Both aesthetic concerns of prominent transcervical scarring and nerve paresis associated with the TCA for SMG excision led to the development of the conventional transoral approach. The MMB of the facial nerve, the lingual nerve, and the hypoglossal nerve are all encountered using the TCA [5]. Neuropraxia can result in both functional and cosmetic impairments. A study conducted by Springborg et al. showed 2.7% (3/113) of patients undergoing the TCA had long-term MMB paresis, less than the 17.6% (3/17) of patients who experienced temporary MMB paresis, as reported by Gülșen and Ulutaș [4,6].

Springborg et al. additionally identified 2.5% (3/139) of patients reporting an unsatisfactory cosmetic result from TCA for SMG excision [6]. Unlike TCA, the TOA inherently avoids aesthetic concerns as the incision and the scar are within the oral mucosa. Additionally, the approach can circumvent injury to the MMB and the hypoglossal nerve. However, studies have shown an increased risk of lingual nerve paresis using the TOA. In 2000, Hong et al. described the TOA for submandibular excision in 31 patients [7]. The mean operation time was 80 min for the first five cases, 60 min for the middle five cases, and 40–50 min for the last five cases [7]. Early postoperative complications included 81% (25/31) of patients with temporary abnormal tongue sensation, 68% (21/31) with limited tongue mobility, 3% (1/31) with hematoma formation, 3% (1/31) with abscess formation, 3% (1/31) with temporary abnormal taste, and 3% (n = 1) with subtotal SMG excision [7]. Ten percent (3/31) of patients developed long-term limited tongue movement as a result of scar contracture, though these were only found on physical exams and were not accompanied with subjective complaints [7]. Lingual nerve neuropraxia resolved within a mean of 1 month [7]. Patients were all discharged within 6 days [7].

In 2020, Gülșen et al. compared 31 patients undergoing SMG excisions, 14 via TOA and 17 via TCA [4]. The mean operating time for TOA was 85.3 ± 17.8 min, significantly longer than that of TCA (40.8 ± 7.3 min, p < 0.001) [4]. However, mean hospitalization time was shorter for TOA (28.5 ± 4.6 h) than for TCA (49.4 ± 6.9 h, p = 0.021) [4]. Gülșen et al. reports that 35.7% of TOA patients (5/14) experienced temporary abnormal taste and tongue sensation, which is comparable to the approximately 25% reported by Kauffman et al. and the 43% reported by Weber et al. [4,8,9]. There was also one case of hematoma in the TOA group [4]. The mean pain score within the first 3 days after surgery was significantly lower for the TOA cohort when compared to the TCA cohort; however, the mean numerical rate scale score of eating and difficulty swallowing was higher for those 3 days in the TOA cohort [4].

Kauffman et al. described a case series with nine patients undergoing a conventional TOA to SMG excision. Four patients experienced complications, with one patient experiencing delayed healing with floor-of-mouth mucosal breakdown, two patients experiencing temporary mild tongue paresthesias, and one patient having a failed transoral attempt due to the size of the gland and scarring [8]. None of the patients experienced hematomas or permanent complications, and the mean hospital length of stay was 2.4 days [8].

Akbay et al. recorded perioperative complications in 19 patients treated with TOA [10]. In this case series, the mean surgery time was 77.63 ± 18.88 min [10]. The authors attributed the difficulties in dissection to limited exposure, as the mandible constricts perioperative view flexibility [10]. This difficulty led to fragmented dissection in two cases of chronic sialadenitis and facial artery injury in two separate cases [10]. Akbay et al. also reported temporary tongue paresthesias in 5.3% of patients (1/19) who had partial adhesion between the Wharton canal and the lingual nerve [10]. The dissection of the nerve from the canal likely contributed to the paresthesias experienced [10].

Curtis et al. conducted a retrospective case series of 30 patients who underwent transoral SMG excision for benign conditions between the ages of 7 and 80 years old [11]. They found that the approach did not pose any additional risk and the transcervical scar was avoided on all patients. The average operative time was 95 min; 10% of patients (3/30) experienced transient tongue paresthesias and 6% of patients (2/30) experienced infection, all of which resolved. No hematomas or seromas were identified. Notably, no patients experienced hypoglossal or MMB dysfunction [11].

3.2. Robotic Transoral Approach

In 2019, Lin et al. noted that the use of the conventional TOA often included removing the sublingual gland to visualize the submandibular gland, citing different institutions’ methods, which described removal of the sublingual gland [7,12,13,14,15,16]. In an effort to preserve the sublingual gland, Lin reported a case study using the robotic transoral approach in a 32-year-old female with pleomorphic adenoma [12]. Using the da Vinci Si robotic system, surgeons dissected laterally at the deep lobe of the submandibular gland and stripped the lingual nerve off of the gland using bipolar coagulation and sharp dissection. Post-surgical drainage lasted 4 days, with a total volume of 97 mL [12]. Their study also reported mild edema of the mouth and submandibular region, as well as ipsilateral tongue numbness, all of which significantly improved by week 4 after surgery. The operation lasted 190 min, with total robotic time from docking to removal lasting about 160 min. Estimated blood loss was 20 mL [12]. In another case report described by Prosser et al., a 51-year-old female with 15-year recurrent sialoadenitis was operated on using the da Vinci Surgical Robotic System [17]. The total operation time was 135 min, with a robotic time of 95 min [17]. Estimated blood loss was 10 mL, and the patient was discharged on day 1 after surgery [17]. The patient experienced transient lingual paresthesia, which resolved within 4 weeks [17].

3.3. Endoscopic Retroauricular Approach

The retroauricular approach has been described to minimize postoperative complications. The conventional retroauricular approach describes an incision placed behind the ear at the hairline. The language varies among different papers, with some defining the retroauricular approach as at the retroauricular sulcus and hairline, while others use the term retroauricular to encapsulate an occipital hairline or trans-hairline approach. Though these incisional lines only vary by a few centimeters, a retroauricular sulcus scar tends to be more visible than an occipital hairline scar, particularly for patients with shorter hair. We will discuss both approaches in the coming sections, using the term “retroauricular” to encapsulate both approaches.

Much like the TOA, the retroauricular approach is aesthetically favorable, as the scar is hidden within the hairline. This approach often utilizes an endoscope for visualization, given the incision is further from the submandibular gland compared to the TOA or TCA. He et al. compared an endoscopic retroauricular hairline approach (n = 74) to the conventional TCA (n = 67) [18]. Hospitalization stay for the former cohort was 5 days ± 1 (p < 0.001), and the visual analog scale score was 8.8 ± 1.3 (p < 0.001) [18]. Four patients had complications; 2.7% patients (2/74) had scar hyperplasia, 1.4% (1/74) had temporary auricular paresthesias, and 1.4% (1/74) had postoperative hematomas [18]. Intraoperative bleeding was significantly lower in the endoscopic retroauricular group (12.5 mL ± 10, p < 0.001) [18]. He et al. described a learning curve with the endoscopic retroauricular approach. The first 29 cases had an average operative time of 110.0 min ± 27.5 (p < 0.001), while the next 44 cases had an average operative time of 85.0 min ± 20, which was not statistically significant compared to the TCA group [18]. Similarly, Rui et al. compared the endoscopic hairline approach (n = 19) with the conventional hairline approach (n = 19) and found that the operation time was longer for the endoscopic approach, with an average operative time of 55.5 ± 23.8 min (p < 0.05) [19]. However, the endoscopic approach had a shorter hospital length of stay (2.8 ± 1.1 days, p < 0.05) compared to the conventional hairline approach [19]. Using the University of Washington Quality of Life Questionnaire (UW-QOL) appearance score, the endoscopic approach had a score of 92.38 ± 7.58 points, which was significantly higher than that of the conventional hairline approach (80.31 ± 12.15 points), despite the fact that incision lengths were not significantly different [19,20].

One study by Yang et al. included 25 patients who underwent endoscopic SMG excision, with a mean follow-up of 10 months. There were no reported complications, including wound infections, hematomas, skin necrosis, nerve injury, hair loss, or recurrence. Patient satisfaction surveys were used to quantify cosmetic results. Fifteen patients reported being very satisfied and ten patients reported being satisfied, for a mean score of 9.2 ± 1 out of 10 [21]. Unlike Yang et al., Neves et al. reported temporary MMB paresis in 10/23 patients (43%) and seroma in 1/23 (4.3%) in a study of 23 patients who underwent an endoscopic retroauricular approach [3]. Neves et al. also reported longer operation times compared to the TCA (128.0 ± 27.60 min, p < 0.001) [3]. However, because of the smaller sample size, this may be attributed to the complexity of the approach and a learning curve. Operative time may decrease as surgeons become more familiar with this approach.

3.4. Robotic Retroauricular Approach

In 2013, Lee et al. explored the feasibility of the robot-assisted retroauricular approach in five patients undergoing SMG excision and reported no intraoperative complications in this cohort [22]. Mean operative time was 62.4 ± 6.3 min and the incision length was 8.1 ± 0.4 cm [22]. In the following year, Lee et al. directly compared the endoscope-assisted (n = 22) vs. robot-assisted (n = 13) retroauricular approaches and found complication rates to be comparable between the two [2]. Lee et al. reported that the robot-assisted cohort (n = 13) experienced 1/13 (7.7%) cases of temporary lip paresis, 1/13 (7.7%) cases of auricular paresthesia, and 1/13 (7.7%) cases of postoperative seroma [2]. The endoscope-assisted cohort had 2/22 (9.1%) cases of temporary lip paresis, 2/22 (9.1%) cases of auricular paresthesias, and 1 (4.5%) case of scar hypertrophy [2]. Intraoperative bleeding was similar between the two cohorts. The endoscope-assisted cohort experienced a mean of 11.0 ± 3.3 mL of intraoperative bleeding compared to the robot-assisted group (9.7 ± 3.1 mL, p = 0.25) [2]. Hospitalization length of stay was also comparable, at 3.8 ± 1.1 days and 3.5 ± 0.9 days (p = 0.45) for the endoscopic and robotic cohort, respectively [2]. In total, there was no difference in incision length, total operation time, intraoperative bleeding, amount or duration of postoperative output surgical collection in the drain, hospital stay, or cosmetic satisfaction. Though these similar outcomes do not suggest any significant difference, the study argues that there still exists a technical advantage to the robotic approach that cannot be quantified, such as precise manipulation of the robotic arms in a tight space [2].

In a case series published in 2017, Yang et al. compares a cohort of 24 patients who underwent robot-assisted, retroauricular excision of submandibular gland using either the Xi or Si da Vinci Surgical Robot system [23]. There were no reported postoperative complications, including nerve injuries, hematomas, or infections. The trans-hairline incision lengths were 4.1 ± 0.4 cm, compared to the retroauricular incision lengths reported by Lee et al., which were approximately 8 cm in both cohorts [22,23]. Though the incision mark in the occipital hairline approach is hidden, the shorter length can contribute to higher cosmetic satisfaction, especially for patients with short or thin hair [23].

4. Discussion

The goal of this review was to evaluate the available evidence on cosmetic and clinical outcomes of submandibular gland excision performed through transoral, transcervical, and retroauricular approaches as shown in

Table 1. Outcomes and risks of surgical approaches.

Approach	Operative Time	Blood Loss	Length of Stay	Cosmesis	Nerve Risk Profile
Transcervical [4]	~40 min (e.g., 40.8 ± 7.3 min)	Not reported	≈49.4 ± 6.9 h; 3–5 days in some series	Visible cervical scar	Marginal mandibular branch paresis ~17–19% immediate postop; rare XII injury
Transoral (Conventional) [4,7,10]	60–90 min; improves with experience	Not reported; facial artery injury in 2 cases	28.5 ± 4.6 h to 2.4 days; up to 6 days in early series	No external scar	Lingual neuropraxia common (25–81%), mostly transient; marginal mandibular branch and XII rarely affected
Transoral (Robotic) [12,17]	135–190 min (case reports)	10–20 mL	Post-op day 1 discharge reported	No external scar	Transient lingual paresthesia; tongue edema with prolonged retraction
Retroauricular (Endoscopic) [18,19]	55–128 min; learning curve improves after ~30 cases	≈12.5 ± 10 mL	2.8 ± 1.1 to 5 ± 1 days	Hairline or retroauricular scar; high UW-QOL appearance scores (~92)	Marginal mandibular branch paresis reported up to 43% in one series; lingual and XII injuries uncommon
Retroauricular (Robotic) [2,22]	≈62 ± 6 min; similar to endoscopic in one series	~10 ± 3mL	3.5 ± 0.9 to 3.8 ± 1.1 days	Trans-hairline incisions < 5 cm; retroauricular ~8 cm	Low rates; transient auricular numbness or mouth corner deviation in small percentages

. Within the included studies of our analysis, we compared the advantages and disadvantages of each technique, focusing on risk of nerve injury, operative time, bleeding risk, cosmetic appearance, and cost.

Cosmetic preference remains one of the primary driving factors when selecting the best surgical procedure. Despite this, most studies often did not look at cosmetic satisfaction scores; instead, they would generally note that the procedure improved cosmetic appearance. The patients were satisfied with the aesthetic outcomes because incisions in all four approaches were hidden. Rui et al. and Yang et al. reported cosmetic satisfaction scores for the endoscopic retroauricular approach, noting 92.38 ± 7.58 points and 9.2 ± 1 out of 10, respectively [19,21]. When comparing the trans-hairline versus the retroauricular incision for robotic surgery, Yang et al. noted that the incision mark for the trans-hairline incision was less than 5 cm [23]. In contrast, the retroauricular incision required an approximately 8 cm incision [22,23]. Though no definitive study compares the cosmesis of these two similar incision marks, the trans-hairline incision may lead to higher cosmetic satisfaction due to the shorter incision and increased ability to hide the incision beneath the hairline. In addition, abnormal tongue movement was noted in conventional TOA and floor-of-mouth edema was noted in the robotic intraoral approaches, which may contribute to perceived cosmetic satisfaction by the patient, though these complications were only transient or physical, not subjective findings.

Nerve injury and oncologic diagnoses are other important factors that play an important role in surgical decision-making. When comparing the risks of nerve injuries across the four operative routes, the hypoglossal nerve is at a higher risk of injury during TCA due to its proximity to the SMG, but it is rarely affected, whereas other alternative approaches largely circumvent the risk. However, the lingual nerve was at a higher risk of injury in both TOA procedures and was reported at a higher rate than the retroauricular approach. Temporary MMB injury was reported by Neves et al. in 43% (10/23) of patients treated via the endoscopic retroauricular approach [3].

From an oncologic standpoint, there is no evidence TOA corridor provides the same margin-negative resection rate as the standard transcervical approach, since no information about oncologic margin equivalence data is provided. The published series shows TOA is generally used for benign or inflammatory conditions of the SMG and is generally contraindicated in malignant or “likely malignant” conditions [4,5,8]. There are certain characteristics that a patient should have in order to maximize the chances of a successful operation, such as the lack of trismus or mandibular hypoplasia and normal tongue mobility [8]. The retroauricular seems like a viable approach for both benign and malignant conditions, as shown by the success in the study conducted by Neves et al. [3]. TCA remains the preferred approach for malignancy, prioritizing oncological outcomes over cosmetic appearances, particularly when neck dissections might also be required.

Furthermore, improvements in operation time were noted in the conventional TOA and the endoscopic retroauricular approach. Hong et al. reported that the first five cases of conventional TOA had an average of 80 min, while the last five cases lasted an average of 40–50 min [7]. Other studies broadly described the average operation time as 85.3 ± 17.8 min and 77.63 ± 18.88 min [4,10]. He et al. reported that the first 29 cases in his endoscopic retroauricular cohort had a mean operation time of 110.0 min ± 27.5, with the latter 44 cases lasting 85.0 min ± 20 [18]. Neves et al. similarly reported on 23 patients operated on via an endoscopic retroauricular approach who averaged 128.0 ± 27.60 min in surgery [3]. It is difficult to pinpoint exactly when surgeons become (or exactly how many procedures provide surgeons with enough experience to feel) comfortable with these new procedures, but the current literature supports that the learning curve levels off at around 30 patients for the endoscopic retroauricular approach [18].

Though data is limited, the longest operation time was described in the robotic transoral case report by Lin et al., with the total operation time reaching 190 min [12]. The shortest operation time, 62.4 ± 6.3 min, was reported using the robotic retroauricular approach [22]. Timing will vary depending on surgeon comfort.

In terms of bleeding risk, there was no notable difference among the robotic transoral, endoscopic retroauricular, and robotic retroauricular approaches, with blood loss ranging between 10 and 20 mL. There were no numbers of blood loss reported in conventional TOA; however, Akbay reported 2 of 19 cases (10.5%) with facial artery rupture due to blunt dissection. This complication will likely demonstrate higher rates of bleeding if blood loss is quantified in future studies [10]. Additionally, rates of hematomas are only reported in the conventional TOA and endoscopic retroauricular approaches. Yang et al. reported no cases of hematoma formation in the robotic retroauricular approach (n = 24) and there was no mention of hematoma formation in either of the two case reports that used the robotic TOA [12,17,23]. Seromas are only mentioned in the endoscopic and robotic retroauricular approach, at a rate of 4.3% and 7.7%, respectively [2,3]. Gülșen et al. reported the shortest hospitalization stay for a conventional TOA cohort, at 28.5 ± 4.6 h [4]. The longest stay was reported by He et al. (n = 74) for an endoscopic retroauricular cohort, at 5 ± 1 day [18]. Rui et al. (n = 19) and Lee et al. (n = 22) reported 2.8 ± 1.1 days and 3.8 ± 1.1 days for endoscopic retroauricular cohorts as well [2,19]. The robotic retroauricular cohort had comparable lengths of stay, at 3.5 ± 0.9 days [2].

Economic background can also influence a patient’s decision in a resource-constrained or fee-for-service environment. The robotic approach is generally more expensive than the endoscopic approach, not only because of the robot used but also the training needed to operate it. Additionally, longer operation times or longer hospitalization times will lead to increased costs. There has been no direct analysis of the direct or indirect costs of the procedures; however, the limited information we have about the robotic transoral approach may deter patients from opting in, especially when they take into consideration the long robotic operation time reported. A direct analysis would be needed to compare the average total operation time, hospitalization time, and complications to determine the most economically feasible route. Given the current evidence, we can say that the endoscopic or conventional approaches will be more cost-effective than either of the robotic approaches [3].

Overall, the surgical approach should be chosen with factors in mind, such as gland size, stone location, hairline, and surgical risk profile. The TOA is generally used for small-to-moderate, non-fibrotic glands and should not be considered in patients with tumors larger than 20 mm in diameter [4,5]. For larger-size glands or even complete gland removal, retroauricular is the best approach, since it provides a wider lateral corridor and better bleeding control, unlike TOA, which was shown to have bleeding challenges in complex cases [7,10,15]. In terms of stone location, TOA would be ideal for patients with distal duct or hilar obstruction, as this technique provides direct access to Wharton’s duct and hilum. Patients with multiple intraparenchymal stones or those requiring wide excision of the gland are better suited for the retroauricular approach, as it provides lateral exposure and better instrument reach [2,3,18,21,22,23]. Another important factor is the hairline. In patients with short hair or alopecia, a TOA would be beneficial in comparison with the retroauricular approach since it is truly scarless. The retroauricular approach ends in an incision hidden in the postauricular sulcus and could still be seen in some patients. Lastly, another important factor is patients with prior surgical history; especially in those with duct surgery or ranula procedures, a retroauricular approach would be favorable. TOA should only be considered in patients with adequate mouth opening and no trismus [8,10].

5. Limitations

Our methodological limitations are largely due to only two databases searched, PubMed and Google Scholar. This may have led to the exclusion of other relevant papers that could have contributed to developing the landscape of the current literature. Future studies may be advised to adhere to PRISMA guidelines and conduct a sweeping review of all the literature. Additionally, there is a paucity of research on the robotic TOA, as only two case reports have been reported on SMG excision.

6. Conclusions

In general, surgeons should counsel patients based on cosmesis, efficiency, neuropraxia risks, and expected costs. While the TCA remains the most widely used, retroauricular and transoral approaches provide alternatives that may improve cosmetic outcomes. However, oncological safety remains an important factor at the time of choosing a transoral approach and continues to be a barrier to these innovative operative advantages. The choice of approach should reflect patient preferences, the surgeon’s familiarity with each technique, and the patient’s diagnosis.