Robotic infraclavicular approach for minimally invasive neck dissection in head-neck cancers

Abstract

Background:

In the recent years, there has been a rapid increase in the use of robot assisted neck dissection (RAND) as an alternative method for conventional neck dissection. Several recent reports have emphasized upon the feasibility and effectiveness of this technique. However, substantial technical and technological innovation is still essential in spite of the availability of multiple approaches for RAND.

Materials and Methods:

The present study describes a novel technique, i.e., Robotic Infraclavicular Approach for Minimally Invasive Neck Dissection (RIA MIND) used in head and neck cancers with the help of Intuitive da Vinci Xi Surgical System.

Results:

After RIA MIND procedure, the patient was discharged on the third post operative day. Also, the total wound size was less than 3.5 cm which enhanced the patient recovery time and required minimal post operative care. The patient was further reviewed 10 days after the procedure for the removal of sutures.

Conclusion:

RIA MIND technique was effective and safe for performing neck dissection for oral, head and neck cancers. However, additional detailed studies will be required for establishing this technique.

INTRODUCTION

Globally, head-and-neck cancers (HNCs) are the sixth most common cancer and represent a major form of cancer in India.[1,2] Standard surgical treatment in HNC patients includes conventional, open lateral neck dissection that leads to surgical morbidity and unavoidable long scar on the neck.[3,4] The growing emphasis on aesthetics of the patient paved the way for the advent of minimally invasive surgery (MIS) such as endoscopic minimally invasive neck dissection (MIND) coupled with enhanced imaging and positioning technology. This coupling not only helped in enhancing aesthetics also in achieving remote access to the head-and-neck tumours, where complete resection with minimal damage to the surrounding tissue was achieved with reduced surgical morbidity.[5] However, the application of MIS in HNCs surgery has been avoided owing to concerns about two-dimensional visualisation, damage to vital structures and it is often criticised for being non-ergonomic, poor replicability and restricting the technique to few masters because of long learning curve.[6]

The use of surgical robot to perform neck dissection overcomes the shortcomings of MIND. Robot has advantage in the restricted space that is available in the neck while performing MIS offering better ergonomics and vision to the surgeon.[4] With these advantages, performing MIND has become a lot easier.[7] Currently, there are two major kinds of approaches for RAND, namely modified facelift (MFL) or retroauricular (RA) incision and transaxillary (TA) approach.[5] Some have used a combined TA and RA incision.[5] There are a few reports of transoral (TO) approach as well.[5] Various approaches for RAND and their surgical intervention details are depicted in Table 1. Innovation in access is an obvious evolution in any surgery, once proficiency in safety standards is achieved. Therefore, in the current paper, we describe a novel technique of Robotic Infraclavicular Approach MIND (RIA-MIND) using da Vinci Xi Surgical System without using any specialised retractors.

Table 1.

Various approaches for robot-assisted neck dissection and their surgical intervention details

Technique	Position	Incision	Flap raising	Neck dissection	Study group
TA	Shoulders - raised Neck-turned to opposite side Arm - abducted 80 degrees	Axillary-10-12 cm long	Subcutaneous followed by subplatysmal (external retractor required)	All levels are possible Difficulty in level I, IIB, lower IV, VA Requires re-docking and repositioning	Lee and Youn[8] Kang et al.[9] Kang et al.[10] Shin et al.[11]
TARA	Shoulders - neutral patient Position - supine Neck - slightly extended, turned to opposite side Arm - abducted 45 degrees	Axillary-7 cm and RA sulcus incision extending in front of tragus, around pinna and into occipital region posterior to hairline (7-10 cm)	Two flaps, both subplatysmal External retractor required	Robotic assisted and dissection under direct vision (level II, Va, upper part of level III)	Kim et al.[12]
RA/MFL	Shoulders - neutral patient Position - supine Neck - relaxed, no extension	Only RA/MFL incision as described above (7-10 cm)	Subplatysmal External retractor required	Part of the neck dissection -level Va, IIB and lateral aspect of level IIa and upper level III is done under direct vision and part of it is robotic assisted (level I, medial aspect of IIa and III, level IV, Vb)	Byeon et al.[13] Bomeli et al.[14] Terris et al.[15] Kim et al.[16]
TO	Neck slightly extended	Gingivobuccal sulcus incisions (1 mm × 12 mm and 2 mm × 8 mm) for port insertion	Subplatysmal Flap held up using CO2 insufflation	Central neck dissection only. No data available for lateral neck dissection via robotic approach	Richmon et al.[17] Richmon et al.[18]
RIA-MIND	Shoulders - raised Neck - turned to opposite side Arms by the side	Infraclavicular ports. 3 mm × 8 mm robotic. 1 mm × 5 mm assistant port. Combined length of all 4 port incisions is 3.5 cm	Subplatysmal Flap held up using CO2 gas insufflations	Level I-V, all dissected via this approach	Current approach

TA: Transaxillary, RA: Retroauricular, TARA: TA combined with RA, MFL: Modified facelift, TO: Transoral, RIA-MIND: Robotic infraclavicular approach minimally invasive neck dissection, CO₂: Carbon dioxide

MATERIALS AND METHODS

Patient selection

Prior approval from the institutional review board was obtained and written informed consent was obtained from the patient. This study was conducted in accordance with the standards ethical committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2000. Patients with tongue, buccal mucosa or lip primary and N0 or N1 neck node were considered for RIA-MIND. Further, patients requiring microvascular or pedicled flaps and/or mandibulectomy were not considered for this procedure. As the current approach offers accessibility to all levels (I to V) level of lymph node dissection was not a concern.

Patient preparation

Before surgery, the patient underwent dental examination and necessary procedures. Further, hair over the chest was clipped on the operative table to place the ports in infraclavicular region.

Surgical technique

Patient position and exposure

Under general anaesthesia, the patient was positioned supine on the operating table, where the apex of the head was stabilised on a soft head ring and the neck was extended using a sandbag under the shoulders and the arms were tucked to the sides of the patient. Further, draping was done with adequate exposure to important surface anatomical landmarks of the neck and supramammary area that include superiorly, mental protuberance, lower border of ramus of the mandible, angle of the mandible and the mastoid processes on either side. Inferiorly, the sternal notch and manubrium and on either side of the entire clavicle till the acromion process were exposed, and laterally both the trapezius borders were exposed and the thyroid notch was obviously visible.

Incision and working space creation

The RIA-MIND procedure involves 4 incisions, where a primary 2-cm incision was placed 2–3 cm below the clavicle in midclavicular line which was 1 cm lower compared with MIND approach using laparoscopic technique which we have published earlier.[19] The incision can be taken further down based on the dressing habit of the patient, so that the scar hides well under the garment. Further, the incision was deepened to the pectoral facia, medially to suprasternal notch and laterally till the acromion process by blunt dissection using finger and diathermy and space was created towards neck to enter the subplatysmal plane of neck as cranially as possible through the wound.

Port placement

The RIA-MIND requires the placement of 4 ports including 3 robotic ports (Intuitive da Vinci Xi Surgical System) that are placed in the infraclavicular region and 1 (5 mm) assistant ports shown in Figure 1. In order to achieve sufficient triangulation, ports were placed apart from the primary port where the distance between the ports would vary based on the built of an individual. First, on either side of the primary incision, 8-mm robotic ports were placed and this was followed by the insertion of 5-mm assistant port which was placed about 4 cm below and in between the camera and the lateral port on the side of the neck dissection. Later, an 8-mm camera port was placed through the primary incision, and the gaps on either side of port were sutured to obtain an airtight closure. Further, the space was insufflated with carbon dioxide (CO₂) at a pressure ranging between 6 and 10 mmHg and flow rate (6–10 ml/min). The lowest possible pressure that maintains the working space was used.

Figure 1.

Patient position and port placement

Docking and instrumentation

After centring the laser light on the camera port from the ipsilateral side of the neck dissection Da Vinci Xi robot was docked. The patient cart anatomy settings were configured to ‘thoracic’ with cart location set to ipsilateral. Monopolar scissors and bipolar forceps were inserted through the working ports and camera port was mounted on the central port. The working ports on either side are synced with surgeon’s right and left arm, respectively. The targeting step is skipped as it will not help in the process and the fourth arm of the robot was retracted back and left unused. The authors prefer to use fenestrated bipolar and hot shears as their choice of instruments for RIA-MIND.

Flap rising

After mounting and synchronising the instrument to the surgeon’s console, subplatysmal flaps were raised. Sometimes, the initial flap raising can be done using laparoscopic instruments to achieve sufficient working space for docking the robotic instruments. The first landmark to be identified was the sternal notch and sternal head of the ipsilateral sternocleidomastoid (SCM) muscle [Figure 2a]. Later, flap raising was continued along the SCM identifying the subplatysmal plane which lies superficial to it. Further, subplatysmal flap was raised superiorly till the inferior border of the ramus of the mandible and mentum, laterally till the anterior border of the trapezius (if modified radical neck dissection is needed), medially up to the midline. In order to prevent button hole or full-thickness burns while raising the flaps an on-table team constantly works without laxity and also in tandem with the operating surgeon to assure in reaching the landmarks. Further, care is taken to preserve the greater auricular nerve (if feasible) while raising the flap posterolaterally and marginal mandibular nerve superiorly. Additional care was taken to avoid button holing of flap while raising the flaps in midline and extreme laterally where the platysma is deficient.

Figure 2.

Intraoperative view of flap raising and level I dissection. (a) Rising the subplatysmal flap Note the ipsilateral SCM muscle, (b) Marginal mandibular nerve adjacent to facial artery and vein, (c) Level Ia tissue dissection between the two anterior digastric bellies, (d-f) Level Ib dissection completed-instrument pointing the mylohyoid, submandibular ganglion and lingual nerve respectively. AD: anterior digastric; PD: posterior digastric, SCM: Sternocleidomastoid

Neck dissection

The dissection of the nodal stations starts from the lower stations. In case of limited dissection of level I-III, we start the procedure with level III and so forth. In order to approach the levels III and IV two heads of SCM were split and the next structure to be identified was the omohyoid muscle more specifically its intermediate tendon which is the best landmark for identifying the internal jugular vein (IJV) owing to its proximity to its anterolateral wall beside this it also marks the boundary between the levels III and IV nodal stations [Figure 3a].[20,21] Later, omohyoid muscle was dissected off from the IJV and following this the rest of the nodal dissection was commenced. Our technique of neck dissection across IJV involves the longitudinal splitting of the investing layer of prevertebral fascia on the IJV along the anterior border and opening it like a book moving all the tissues posteriorly and superiorly. The assistant can retract and suction the fluid and keep the surgical field open and clean. The tissue around the IJV in levels III and IV can be dissected in this manner [Figure 3b]. The ansa cervicalis and its branches to the strap muscles can be preserved if possible and the omohyoid can be preserved (if a supraomohyoid neck dissection is planned) or divided depending on the extent of dissection planned. Following this, posterolateral dissection around the IJV was started by slit and cut technique using robotic scissors. Care was taken to avoid damage to the vagus nerve during this dissection. Later, level II dissection was performed by retracting the entire SCM laterally by the assistant. The posterior belly of digastric is identified and it serves as the upper limit of the dissection. Spinal accessory nerve (SAN) was identified in the posterosuperior end and the level IIb dissection was completed [Figure 3c and d]. Care was taken in identifying and preserving the SAN in this region as it exits the jugular foramen and enters the sternocleidomastoid muscle. All the fibrofatty lymphatic tissue from behind the IJV and lateral to it and over the prevertebral fascia till the posterior border of the SCM was dissected in a medial to lateral and carnio-caudal fashion and taken en bloc. In case if level V dissection was planned, the dissection was done in caudocranial fashion [Figure 3e] starting the procedure by incising the cervical fascia just above the clavicle and then proceeding craniocaudally taking the level II-V tissues en bloc. SAN is identified and preserved in the posterior triangle by identifying the nerve as it exits from the posterior border of SCM muscle.

Figure 3.

Intraoperative view of level II-V dissections. (a) Splitting the SCM with identification of omohyoid muscle, (b) Level III and IV dissection posterior to IJV, (c) SAN identification and level IIb dissection, (d) SCM lateral retraction and Level IIa dissection, (e) Level V dissection – note the prevertebral fascia and the transverse cervical artery in the posterior triangle. AD: anterior digastric, PD: posterior digastric, CFV: Common facial vein, SAN: Spinal accessory nerve, IJV: Internal jugular vein, SCM: Sternocleidomastoid

The dissection of level I starts by incising the deep fascia along the inferomedial border of mandible. The marginal mandibular nerve was identified and preserved [Figure 2b]. Later, the facial vessels were ligated just below the mandible and the dissection was continued till the mentum to dissect the level 1a nodes. Further, deep cervical fascia and adjacent fibrofatty tissue were dissected from the anterior border of upper 3^rd of SCM and rolled upwards. Following this, digastric muscle was identified and dissection was continued till the entire length of the muscle. which marks the beginning of the dissection of level Ib. Next, facial vessels were clipped and divided at the level of posterior belly of digastric [Figure 2c], and the common facial vein was clipped and divided first which aids in defining the superior border of the posterior belly of the digastric muscle. Following this, facial artery was clipped and divided. The submandibular gland was reflected superiorly and dissected off from the mylohyoid muscle [Figure 2d]. This manoeuvre aids in identifying the hypoglossal nerve, lingual nerve and the submandibular duct and ganglion [Figure 2e and f]. Later, the lingual nerve was separated from the submandibular ganglion. The submandibular duct was divided at the edge of the mylohyoid muscle and the rest of the gland and fibrofatty tissue was freed and the level Ib tissue was placed in the area.

Specimen retrieval was done using endobag and retrieved from the primary port after confirming hemostasis and a suction drain was placed and brought out of the assistant port. Haemostasis was confirmed using the suction and deflating the pneumonia intermittently. This deflation test is important as the veins collapse under pressure and may start bleeding in post-operative period.

RESULTS

Post-operative care and outcomes

The patient was discharged on the 3^rd post-operative day as this procedure causes minor wounds requiring minimal post-operative care. Further, the patient was reviewed on the 10^th post-operative day for the removal of sutures.

DISCUSSION

To the best of our knowledge, this is the first report on RIA-MIND in HNCs using Intuitive da Vinci Xi Surgical System. Irrespective of the approach, trans-axillary (TA), combined trans-Axillary and retro-auricular (TARA), retro-auricular (RA) or RIA-MIND, the principles of optimum neck dissection remain the same and the key is to access the subplatysmal plane and to identify the ‘anatomical lighthouses’.[22] In case of RA/MFL approach, entry to subplatysmal plane is made from the posterolateral region,[16] while in TA approach, the subplatysmal plane is reached along the lateral border of pectoralis fascia.[10] In the current infraclavicular approach, entry to the subplatysmal plane is made by incising the fascial attachments of the clavicle, but the key here is to spot a loose connective tissue layer called the superficial cervical fascia present between the platysma and the SCM, near the origin of its sternal head.[23] All the anatomical lighthouses that dictated our dissection at every level are depicted in Table 2.

Table 2.

Anatomical lighthouses that dictate the current approach

Anatomical lighthouses in RIA-MIND technique
The lighthouse	Outcome
Lower border of mandible	Identification of marginal mandibular branch of facial nerve
Anterior border of SCM	Leads to posterior belly of digastric
Common facial vein ligation	Defines posterior belly of digastrics fully
Posterior belly of digastric	Marks the beginning of level Ib dissection in a posterior to anterior direction by lifting off the submandibular gland
Anterior belly of digastric	Marks the beginning of level Ia nodal dissection
Omohyoid muscle	Helps identify IJV
	Marks the boundary between level III and IV nodes
Erb’s point	Helps to identify SAN which enters posterior triangle 2 cm superior to this point
Prevertebral layer of deep fascia ‘Holy layer’	Protects the phrenic nerve and brachial plexus in level V dissection

SAN: Spinal accessory nerve, RIA-MIND: Robotic infraclavicular approach minimally invasive neck dissection, SCM: Sternocleidomastoid, IJV: Internal jugular vein

Remote access thyroidectomy techniques were criticised for their extensive flap dissection to reach the surgical field of interest.[18] Contrastingly, access routes of the present MIND techniques are close to the surgical field of interest. Further, avoidance of long curvilinear and Y-shaped incisions avoids the scars and facilitates the readiness for adjuvant therapy soon after surgery. As mentioned earlier, the ports are inserted close to dissection area, i.e., infraclavicular which facilitates minimising the morbidity presently associated with flap rising. Further, space needed for primary port insertion was preferably done by blunt finger dissection to minimise incision size and inserting any retractors can enlarge the size of incision. Attention should be given to raise the flap 2–3 cm caudal to the incision which facilitates a lift to the robotic ports after gas insufflation. This little detail not only prevents the fulcrum effect of the clavicle but also assists the instrument to reach difficult nodal stations such as levels IIB, IV and V.

After securing the ports and insufflating gas, depending on the surgeon’s preference, the flap was raised either by robotic or laparoscopic instruments. The extent of flap rising in the neck remains more or less the same in all approaches with an exception in TO approach where the flap rising is limited to the central compartment.[10,12,16-18,24]

The hallmark of our technique includes flap elevation by insufflating CO₂ gas that avoids the use of specialised retractors to hold the flap up thereby reducing additional investment and facilitates dissection by inserting ports (as opposed to an incision) like in any MIS. The lack of specialised retractor has been a restricting point for starting robotic head-and-neck programs in many centres across the world. The initial insufflating pressure was set to 6 mmHg and baseline end-tidal CO₂ (ETCO₂) was noted and the pressure was raised if the flap dissection is cumbersome or working space is inadequate for flap rising or dissection. The optimum pressure proposed for proper view of the surgical field without adverse metabolic and hemodynamic changes in humans undergoing the RAND is under 10 mmHg.[25] In the current approach, we managed to complete our dissection at a pressure ranging between 6 and 8 mmHg and any rise in ETCO₂ beyond the range can be managed by desufflating or achieving haemostasis in case any blood vessel has been breached (usually a superficial vein or rarely IJV).

In the current scenario, we followed an ‘en bloc’ approach for the removal of lymph nodes at all levels.[26] Only level I was disconnected in case it was found difficult to manoeuvre the specimen. Unlike TA approach where the clavicular head of SCM is divided to visualise the junction of IJV with subclavian vein,[10] in RIA-MIND SCM is split or retracted laterally to facilitate levels II, III and IV dissection and retracted medially to complete level V dissection in a craniocaudal fashion. However, the authors see no disadvantage in cutting the SCM as there is no risk of flap necrosis in RIA-MIND.

Another important aspect of RAND is the on-table manoeuvres to facilitate nodal dissection which includes changing the position of patient, repositioning the external retractor and redocking the patient cart as in TA approach and using the assistant arm of the robot for performing nodal dissection under direct vision as in RA/MFL approach. However, RIA-MIND approach is not plagued by these manoeuvres and facilitates optimum neck dissection. Further, the CO₂ insufflation holds the flap up, thereby limiting the use of expensive external retractors that are generally used in gasless techniques.[9,10,12,14] This ensures the quality of the skin flap as gas does not produce any undue pressure on the flap. In the current approach, much of the traction-counter traction is done by the left-hand instrument of the operating surgeon supplemented by the help of the on-table assistant, which in turn avoids the use of third robotic arm and facilitates dissection. Besides this, all the nodal positions can be visualised from the same camera and port positions; therefore, entire surgical field from caudal to cranial can be seen which is a familiar anatomy for all open surgeons.

Earlier, Weinstein et al. were successful in introducing and transitioning transoral robotic surgery from canines to humans. Further, they successfully performed supraglottic laryngectomy in 3 patients with a median operating time of 120 min without any perioperative complications and mortality.[27] Later, Kim et al. described robotic neck dissection in oral and laryngopharyngeal cancers using a combined TARA approach.[12] Since then, many studies have described the long-term (5 years) functional, cosmetic, operative and oncologic outcomes of RAND in HNCs.[4,28,29]

Further, the robotic TA approach is well established in the management of well-differentiated thyroid cancer.[9] However, when it comes to neck dissection involving head-and-neck squamous cell cancers, Shin et al. reported difficulty in accessibility issues when dissecting levels I, IIb and lower IV nodal stations.[11] Similarly, Kang et al. reported similar difficulty in accessing level Va by TA approach and suggested that before proceeding with level IIA dissection through TA approach, the external retractor and robotic axis and operating table should be repositioned accordingly to aid in better visualisation of the dissection area.[10] Besides this incision size in TA approach was increased from 7 to 10 cm to facilitate the entry of the additional retractor arm along with working arms which brings both retractor blade and working arm close enough.[9] Perhaps, this was the guiding principle behind the evolution of the TARA and RA/MFL approaches. Further, the long incisions especially in the RA region have 3.7% risk of flap necrosis. In comparison, RIA-MIND requires 3 small subclavian incisions measuring in total 3.5 cm reducing the wound-related complications.

In RA/MFL approach, the dissection of level IIB and VA along with the lateral aspect of level IIA and upper part of level III was done under direct vision owing to the proximity of these levels to the incision and the rest of the nodal dissection (level I, medial aspect of level IIa, III, level IV and level Vb) was done using robotic assistance.[16] TARA approach involves employing the best of both TA and RA/MFL techniques; therefore, it offers a comprehensive neck dissections from level I to V, especially in HNCs.[12] Further, TARA approach allows the dissection of levels IIB and Va (extended to levels IIa and III) under direct vision and specimen can be easily handled with the assistant arm placed through the RA incision. However, the indication for the TARA approach became limited to the selective neck dissection (SND) including levels II–IV, II–V, or I–V, whereas SND levels I-III were handled by RA/MFL approach alone.[12] Currently, levels I-V nodes can all be dissected through RA approach with surgical and functional outcomes compared to the conventional open technique. This is facilitated by the manipulation of external retractor, dissection under direct vision where feasible and the versatility of the robotic working instruments.[28,30,31]

CONCLUSION

We believe that RIA-MIND technique appears to have eased the difficulties that are encountered in its contemporary techniques using familiar anatomical orientation, CO₂ insufflation and facilitates one-stop completion of neck dissection of all neck levels (I-V) without necessitating the need to raise long flaps and employing expensive external retractors to keep the flap elevated. Above all, robotic assistance eases the neck dissection comfortable and ergonomic to the surgeon. Further, RIA-MIND might be an alternative means of surgery that needs to be provided with further research and evidence.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.