Lymph Node Dissection
Daniel C. Thomas and Dale Han
Lymphadenectomy is indicated for the treatment of macroscopic or microscopic melanoma metastatic to associated nodal basin draining the primary lesion in patients who have no evidence for distant disease. Patients with palpable nodes may present with macroscopic nodal metastasis at the time of initial primary evaluation or may develop macroscopic disease later as a nodal recurrence. Patients with palpable nodes often undergo needle biopsy to confirm metastatic melanoma in the enlarged lymph node. Microscopic nodal metastasis may also be detected in clinically node-negative patients through sentinel lymph node biopsy (SLNB). When lymphadenectomy is performed for macroscopic nodal disease, it is termed a therapeutic lymph node dissection (TLND), whereas lymphadenectomy performed for microscopic nodal disease detected by SLNB is termed a completion lymph node dissection (CLND).
The goals for performing lymphadenectomy are to improve regional disease control and to improve survival. Although the presence of lymph node metastasis is a predictor for worse survival as defined by the seventh edition of the American Joint Committee on Cancer (AJCC) staging system for melanoma, there is no conclusive evidence showing that lymphadenectomy definitively imparts a survival benefit (1). The data from the Multicenter Selective Lymphadenectomy Trial (MSLT)-I suggest that survival may be improved for positive SLN patients who have an immediate CLND when compared with patients who have a delayed lymphadenectomy for a macroscopic nodal recurrence (62.1% vs. 41.5% 10-year melanoma-specific survival, respectively; HR: 0.56; 95% CI: 0.37–0.84; p = .006) (2). It should be noted that this survival benefit was only seen when all node-positive patients (positive SLN patients and patients with a macroscopic nodal recurrence) were analyzed together as opposed to evaluating the original treatment arms. When the initial treatment arms (wide local excision with SLNB followed by CLND for a positive SLN versus wide local excision with nodal observation followed by TLND for a nodal recurrence) were compared, disease-free survival was significantly improved in the arm treated with SLNB followed by CLND. However, there was no significant difference in melanoma-specific survival or overall survival (3). In addition, the results of other studies suggest no significant difference in survival between positive SLN patients who did and did not have a CLND (4). Currently, the effect of CLND on survival is extensively debated, and no definite conclusions can be made as to whether performing a lymph node dissection affects survival in patients with melanoma.
The main reason for performing lymphadenectomy in melanoma patients with nodal disease is for regional disease control, particularly in patients with macroscopic nodal metastases. For patients with microscopic nodal disease (positive SLN), 15% to 20% of cases will have further disease in the CLND specimen, and lymph node dissection is recommended to remove these additional involved nodes (2,3). The rate of regional nodal recurrence in positive SLN patients who do not undergo CLND is approximately 15% to 20%, which is reduced to <5% after CLND for a positive SLN (4,5). In addition, an analysis from MSLT-I suggested that treatment of disease at the microscopic level through CLND is associated with a lower rate of lymphedema compared with a TLND performed for a macroscopic nodal recurrence. Possible reasons for this may be that more nodal tissue may be involved with tumor and more tissue may need to be dissected for macroscopic disease, thereby increasing the amount of lymphatics that are disrupted. These results suggest that CLND allows for regional disease control with decreased morbidity.
However, approximately 80% to 85% of positive SLN patients have no further nodal disease found in the CLND specimen, and some argue that these patients are needlessly exposed to the morbidity of CLND with little additional benefit (6,7). Others argue that lymph nodes in a CLND specimen are not evaluated as rigorously as for a SLNB, and the rate of finding additional nodes with metastatic melanoma in a CLND specimen may be higher. Several studies have evaluated clinicopathologic factors that may predict an additional positive nonsentinel lymph node (non-SLN) in a CLND (8–11). The ability to clinically predict which positive SLN patients are at higher risk for additional nodal disease would allow offering CLND to patients who may benefit the most from this procedure. Multiple factors have been assessed, but none are consistently reported in the literature as prognostic; these predictive markers are extensively debated. Several classification systems to predict non-SLN disease have been reported, although none are universally accepted (9–15). The most promising factor to predict which patients may not need a CLND is a subcapsular nodal deposit in the SLN measuring <0.1 mm in maximal diameter, which predicts for an approximately 5% chance for additional nodal disease (10). However, other studies have not shown the same results (9,13). Currently, there is no consensus and no reliable way to predict which patients with a positive SLN will have additional nodal disease. As a result, the standard recommendation in current guidelines is to offer CLND to all positive SLN patients, and to offer TLND for all patients who develop macroscopic nodal disease to allow for regional disease control (1,16).
Although there is less debate about the utility of TLND for providing regional disease control in patients with macroscopic nodal metastases, the development of new immunotherapies and targeted therapies have now prompted discussion about the merits of these systemic therapies in patients with macroscopic nodal disease, the timing of TLND if systemic therapies are used, and the role of TLND in patients who demonstrate a complete response.
For patients with microscopic nodal disease, part of the debate over whether to perform CLND for a positive SLN arises from two schools of thought in terms of the biology of the sentinel nodes (17). One school describes a model in which sentinel nodes serve as an “incubator” through which melanoma metastasizes in an orderly process. In this concept, metastasis first occurs in the SLN, which serves as the “gateway,” followed by spread to other nonsentinel nodes in the draining nodal basin and then potentially to distant sites. The second school describes a model in which sentinel nodes serve as an “indicator” of disease that has potentially already spread microscopically to distant sites. This latter model is based on the fact that approximately 15% to 20% of cases will develop distant metastases without the development of disease in the draining nodal basin. The “incubator” model of SLN metastasis would suggest that further spread of melanoma can be halted through surgical treatment of nodal disease. In contrast, the “indicator” model would suggest that positive SLN patients already have microscopic distant disease and systemic therapy is required. There is active debate as to which model accurately represents the biology of SLN disease, but it is likely that both models are correct and that SLN disease correlates to each model in specific subsets of patients.
The ongoing MSLT-II trial was developed to clarify the role of CLND in patients with a positive SLN. MSLT-II will analyze prospective survival data in positive SLN patients comparing patients who are randomized to undergo CLND with patients who are randomized to nodal observation. Accrual has completed, but the analyses are still pending and will not be reported for several years. The European Organization for Research and Treatment of Cancer–Minimal Sentinel Node Tumor Burden study (EORTC–MINITUB) is another trial that is looking at survival in patients who undergo nodal observation versus CLND. The results of the German Dermatologic Oncology Group (DeCOG)-SLT trial were reported at the 2015 American Society of Clinical Oncology Annual Meeting (18). DeCOG-SLT is a phase III study which randomized positive SLN patients to either CLND or nodal observation. After a median follow-up of 35 months, this trial showed no differences in 3- and 5-year recurrence-free survival, distant metastasis-free survival, and melanoma-specific survival between positive SLN patients who were randomized to nodal observation and positive SLN patients who had CLND. However, it is important to note that the authors did not present data in terms of regional disease control. It is possible that certain subgroups of positive SLN patients may not benefit from CLND, but at this time, current guidelines recommend CLND for all positive SLN patients, and caution is warranted against abandoning the standard of care based on currently limited data.
CERVICAL LYMPH NODE DISSECTION
Regional metastases of head and neck melanomas often follow multidirectional and unpredictable lymphatic drainage patterns, and cervical lymphadenectomy warrants specific technical and anatomic considerations. Cervical lymph nodes are divided into six levels (I–VI) defined by certain anatomic boundaries as shown in Figure 8.1 (19,20). Level I encompasses the submental triangle (IA), which is bound by the anterior bellies of the digastric muscles and the hyoid bone, and the submandibular triangle (IB), which is bound by the anterior and posterior bellies of the digastric muscles, the stylohyoid muscle, and the body of the mandible. Level II, the upper jugular group, includes the lymph nodes surrounding the upper third of the internal jugular vein and spinal accessory nerve extending from the skull base to the level of the carotid bifurcation. Level III, the middle jugular group, extends from the carotid bifurcation to the omohyoid, with the anterior and posterior boundaries being the lateral border of the sternohyoid muscle and the posterior border of the sternocleidomastoid muscle (SCM), respectively. The lower jugular group, level IV, is located at the lower third of the internal jugular vein from the inferior border of the cricoid cartilage to the clavicle. The anterior and posterior borders of level IV also consist of the sternohyoid muscle and the posterior border of the SCM, respectively. Level V nodes, termed the posterior triangle group, extend anteriorly from the posterior border of the SCM and posteriorly to the anterior border of the trapezius muscle. Superiorly, this group is bordered by the skull base, and the inferior border is the clavicle. The anterior compartment group, level VI, comprises the midline neck superiorly from the hyoid bone to the suprasternal notch inferiorly, and is bordered laterally by the medial border of the carotid sheaths.
The full extent of cervical lymph node dissection is debated. The extent of dissection depends on the site of the primary lesion, the presence of clinically evident nodal disease, and identification of the involved lymph node basins. Several methods have evolved to address nodal metastasis of the neck. A century ago, George Crile introduced the operation for en bloc resection of the regional lymph nodes in the neck, together with the internal jugular vein, the spinal accessory nerve, and the SCM (21). He wrote that he included the “sternocleidomastoid so as to afford a better exposure of the parotid group of glands, thereby protecting against injury of the facial nerve or the jugular.” Hayes Martin then popularized the radical neck dissection with his “commando” operations, combining cervical lymphadenopathy with mandibular resection for oral cancers (22). Ettore Bocca subsequently advocated for a “conservative neck dissection” with preservation of the internal jugular vein, spinal accessory nerve, and the SCM for oropharyngeal cancers in patients with N0 necks (23). Stephan Ariyan then reported on his technique of a “functional radical neck dissection” in patients with N1 and N2 necks, half of whom had metastatic melanoma (24). These later techniques reduced the morbidity of the earlier radical neck dissection and formed the basis for creating the modified radical neck dissection (MRND). However, the extent of neck dissection continues to be debated with some advocating for MRND in order to aggressively gain regional disease control for patients with known nodal metastases. Others recommend a less extensive neck lymph node dissection called selective neck dissection (SND) and suggest that SND is associated with less morbidity but allows for comparable regional disease control (25,26).
Modified Radical Neck Dissection
Lesions of the upper neck and posterior scalp are treated with MRND. This entails dissection of levels II to V and preservation of the spinal accessory nerve, SCM, and internal jugular vein. These three structures are preserved unless there is direct involvement with metastatic disease (26). Multiple types of incisions have been described (see Figure 8.2) such as Conley’s incision which consists of a broad curvilinear incision beginning below the mastoid, traversing over the SCM to the clavicle, and coursing anteriorly. Any prior SLNB site should be incorporated into the incision. Subplatysmal skin flaps are created to the level of the mandible medially, clavicle inferiorly, and to the anterior border of the trapezius muscle posteriorly. Flaps are created with attention to protecting the greater auricular and spinal accessory nerves. Of note, Erb’s point is located on the posterior portion of the SCM at the juncture of the middle and upper thirds of the SCM, from which the cutaneous branches of the cervical plexus exit. The spinal accessory nerve is located approximately 1 cm above Erb’s point. The SCM is isolated and separated from the strap muscles. The SCM is then retracted medially to allow for dissection of the level V lymph nodes. The nodal specimen is dissected away from the trapezius muscle with care to identify and protect the spinal accessory nerve. Dissection continues medially to the anterior border of the SCM, where the carotid sheath, internal jugular vein, and the vagus nerve (which lies deep and between the carotid and internal jugular vein) are identified. The nodal tissue along the internal jugular vein is dissected from inferior to superior along the SCM to include the lymph nodes of levels II, III, and IV. Of note, a level I lymph node dissection may also be indicated depending on the location of the primary lesion, particularly lesions located along the lower lip and jaw. With the dissection completed, the lymph node specimen is removed. A closed-suction drain is placed and the platysma is reapproximated. The subcutaneous tissue and skin are then closed primarily.
Approximately 30% of melanoma lesions of the face, lips, ears, and anterior scalp drain to the parotid lymph node basin (26). For this reason, these anterior lesions may require parotidectomy in addition to cervical lymphadenectomy for optimal regional disease control in cases that have a positive SLN or macroscopic nodal involvement. Some advocate performing a parotid SLNB, suggesting that the procedure could potentially spare patients the morbidity of a parotidectomy. If there is no drainage by SLNB to the parotid nodes or if the parotid SLN is negative for metastatic melanoma, parotidectomy is not indicated (27). However, parotidectomy is recommended in patients who have a positive parotid SLN and in patients with clinical involvement of the parotid.
When a parotidectomy is performed, a superficial parotidectomy is recommended unless there is involvement of the deep segment of the parotid gland. A superficial parotidectomy involves an incision carried from the anterior tragus down to the lobule of the ear, and posteriorly over the mastoid process curving down the superior cervical crease to below the angle of the mandible (28). The tail of the parotid gland is dissected free from the SCM anteriorly to the posterior belly of the digastric muscle. The facial nerve is identified and preserved within the tympanomastoid fissure coursing between the digastric muscle and the external auditory canal (29). Careful dissection of the facial nerve from the parotid gland continues anteriorly, with attention paid to preserving all major divisions of the nerve. Once free from the facial nerve, the superficial segment is ligated at the isthmus, allowing the deep segment to remain below the facial nerve.
Selective Neck Dissection
SND can also be utilized in patients with known nodal metastases (30). As opposed to MRND, in which multiple lymph node levels are resected in a standard fashion, SND consists of removing only specific lymph node basins that are most likely to drain a primary site. As with MRND, the goal of SND is to gain regional disease control and remove all clinically evident nodal disease. The selection of which lymph node levels to include in a SND is based upon anatomic drainage patterns of the primary lesion and vary from site to site. For instance, a melanoma of the lower lip may require a SND of only level I nodes, whereas a posterior–lateral SND, including levels II to V, can be performed for retroauricular or suboccipital lesions (31,32).
AXILLARY LYMPH NODE DISSECTION
Melanomas of the upper extremities or trunk often drain to axillary lymph nodes. The goal of axillary lymph node dissection (ALND) in treating melanoma nodal metastases is to remove all lymphatic tissue including level I nodes, which are inferior and lateral to the pectoralis minor muscle; level II nodes, located posterior to the pectoralis minor muscle and inferior to the axillary vein; and level III nodes, which are medial to the pectoralis minor muscle (33,34).
A curvilinear incision is made from the lateral border of the pectoralis major muscle extending posteriorly below the hair-bearing area to the anterior border of the latissimus dorsi muscle. Any prior SLNB site should be incorporated into the incision. Skin flaps are then raised at the level of the deep subcutaneous fascia to the border of the pectoralis major muscle anteriorly and to the latissimus dorsi muscle posteriorly. The axillary nodal basin is then entered, and nodal tissue is dissected from the undersurface of the pectoralis major muscle. Dissection continues to the medial border of the pectoralis minor muscle. The nodal tissue between the pectoralis muscles and Rotter’s nodes are dissected into the specimen. Dissection continues along the posterior surface of the pectoralis minor muscle to its medial border where the medial pectoral nerve is identified and protected to preserve function of the pectoralis muscles. Inferiorly, the nodal tissue is dissected free where the latissimus dorsi muscle meets the chest wall (approximately the fourth or fifth rib). The serratus anterior muscle is then exposed and the nodal tissue is dissected off the serratus anterior muscle. Dissection continues superiorly toward the apex of the axilla until the axillary vein is exposed. The nodal specimen is then dissected off the axillary vein and the latissimus dorsi muscle. Care is taken to identify and protect both the long thoracic nerve and thoracodorsal neurovascular bundle, which are separated from the axillary contents. When dissecting the specimen off of the serratus anterior muscle, several intercostal nerve bundles, which provide sensation to the skin of the medial upper arm, are also encountered and often ligated. At this point, the level I and II nodes are free. With retraction of the pectoralis minor muscle, the level III nodes are exposed along the chest wall and axillary vein up to where the axillary vessels emerge from the chest wall. The level III nodes are then dissected into the specimen. With completion of the dissection, the lymph node specimen is removed. A closed-suction drain is placed in the axilla and the subcutaneous tissue and skin are closed primarily.
Current National Comprehensive Cancer Network guidelines recommend an anatomically complete ALND for melanoma patients with a positive SLN or with macroscopic nodal disease (1). Despite the routine inclusion of levels I to III nodes for ALND performed in melanoma patients, some advocate for dissection of only levels I to II in patients with microscopic disease found through SLNB. Proponents of levels I to III ALND argue its necessity due to variable drainage patterns and an increased risk of nodal recurrence if level III nodes remain in situ. Advocates of levels I to II ALND report that there is potentially less morbidity and that regional nodal recurrence and survival rates are equivalent to levels I to III ALND (35).
INGUINAL LYMPH NODE DISSECTION
Melanomas of the lower extremities or trunk often drain to the inguinal region and inguinal lymphadenectomy is performed for documented metastatic disease in these nodal basins. The goal of inguinal lymphadenectomy is to remove all lymphatic tissue in the superficial inguinal nodal basin with or without dissection of the deep pelvic nodal basin.
Superficial Inguinal Lymph Node Dissection
The superficial inguinal lymph node dissection begins with a “lazy-S”-shaped incision extending vertically from approximately 2 to 3 cm medial to the anterior superior iliac spine, into and parallel to the groin skin crease, then extending vertically down to the distal apex of the femoral triangle (36). Any prior SLNB site should be incorporated into the incision. A medial flap is raised to the adductor longus muscle border while a lateral flap is raised to the sartorius muscle border. A superior flap is raised above the inguinal ligament while an inferior flap is raised to the apex of the femoral triangle. The dissection begins by sweeping the suprainguinal nodal tissue off the external oblique aponeurosis inferiorly into the femoral triangle. Next, the overlying tissue and fascia of the sartorius muscle is mobilized medially, exposing the femoral vessels and nerve, which are identified and preserved. Similarly, the nodal tissue overlying the adductor longus muscle is mobilized laterally where the saphenous vein is encountered. Of note, there is no need to excise the fascia of this muscle as this increases lymphedema rates with no improvement in cancer control (37). For patients with microscopic nodal disease, the saphenous vein is generally preserved; however, in patients with macroscopic nodal disease, the saphenous vein is usually ligated and taken en bloc with the specimen. The tissue overlying the femoral vessels, which are skeletonized over their anterior 180°, is carefully resected from the apex of the femoral triangle to the inguinal ligament as shown in Figure 8.3 with attention to ligating small branches of the artery and vein. The dissection continues medially to the femoral vein, mobilizing the nodal tissue superiorly to the level of the saphenofemoral junction where the saphenous vein is either ligated as it enters the femoral vein (for macroscopic nodal disease) or dissected free from the specimen (for microscopic nodal disease). Dissection is performed to the level of the femoral canal where the specimen is ligated and removed. To protect the femoral vessels, a sartorius muscle flap may be used, although this may not be necessary if the skin flaps are not too thin and are fashioned at the level of Scarpa’s fascia. The sartorius muscle flap is created by dividing the sartorius muscle at its insertion on the anterior superior iliac spine. Care is taken to preserve as many perforating vessels as needed to maintain adequate muscle perfusion. The muscle is then rotated to cover the femoral vessels and secured to the inguinal ligament with sutures. A closed-suction drain is placed in the superficial groin and the subcutaneous tissue and skin are closed primarily.