When treating patients with non-small cell lung cancer (NSCLC), it is important to assign an accurate clinical or pathologic stage to the disease at the time of diagnosis. This adds value to the process of selecting the most appropriate therapy for the individual patient, whether surgical resection, neoadjuvant chemotherapy or radiotherapy, or definitive chemoradiation. The current cancer staging convention uses the basic descriptors originally proposed by Denoix: primary tumor (T), lymph node involvement (N), and tumor metastasis (M).¹ The contemporary classification system was adopted worldwide in 1997 after features of the 1986 combined American Joint Committee on Cancer and the International Union Against Cancer TNM staging system² were reconciled with the 1983 American Thoracic Society statement on cancer staging. The organization responsible for updating this system is the International Association for Lung Cancer Staging. The value of classifying NSCLC patients according to a uniform staging system that has prognostic implications based on stage grouping is difficult to overstate.

Clinical staging can be determined based on CT scanning, MRI, and CT/PET scanning. Pathologic staging requires biopsy, which can be obtained from cervical or anterior mediastinoscopy, during video-assisted thoracic surgery (VATS) approaches, less commonly by means of open thoracotomy, and most recently by fine-needle aspiration performed during endoscopic or endobronchial ultrasound. Lymph node involvement has important implications for surgical treatment strategies for lung cancer.^4–6 Patients without lymph node involvement (N₀) or those with limited involvement (N₁), which is usually determined at the time of surgery, are candidates for resection based on T and M status. Most patients with contralateral or supraclavicular disease (N₃) or T₄ involvement are not resectable (stage IIIB). The usual approach in stage IIIA patients with ipsilateral mediastinal nodal involvement (N₂) involves either neoadjuvant chemotherapy or chemoradiation, followed by resection, if appropriate, when N₂ status is determined prior to resection. In some cases, surgically detected N₂ disease can be discovered at the time of resection by means of lymph node dissection in conjunction with the pulmonary resection. There is no doubt that sampling of lymph nodes in some fashion is useful for staging and prognostic purposes. However, the extent of lymph node dissection is controversial, and the benefits of systematic sampling versus complete mediastinal lymph node dissection or extended lymph node dissection are still under review.^7–10

LYMPH NODE MAPPING

To unify the two most widely used systems of lymph node mapping in NSCLC, the American Joint Committee on Cancer and the International Union Against Cancer adopted a standardized method of classifying lymph node stations in 1996. This was outlined in 1997 by Mountain and Dresler based on the work of Naruke and the American Thoracic Society and the North American Lung Cancer Study Group.¹¹ The most notable difference between these systems was the boundary between peribronchial hilar (N₁) and paratracheal mediastinal (N₂) lymph nodes comprising stations 4 and 10, respectively. These are now separated by the pleural reflection between the visceral pleura (station 10) and the mediastinal pleura (station 4) (Fig. 69-1). The other definitions remained basically the same. The nodal station descriptions of mediastinal, hilar, and intrapulmonary are designed to be reproducible, and stage groupings are based on studies of prognostic factors, including metastasis.

Regional lymph node stations for lung cancer staging are shown in Fig. 69-2. Any double-digit station is classified as N₁. The anatomic definitions for lymph node mapping are shown in Table 69-1. Involvement of contralateral mediastinal or hilar stations, as well as ipsilateral supraclavicular or scalene nodes, is considered N₃ disease. Survival is inversely proportional to nodal involvement and was validated in a database analysis by Mountain and Dresler.¹¹

Reproduced with permission from Mountain CF, Dresler CM: Regional lymph node classification for lung cancer staging. Chest 111:1718–23, 1997.

The distribution and likelihood of lymph node metastasis based on lobar location of the cancer has been described.¹² In Cerfolio and Bryant's study of 954 patients, incidence and location of N₂ disease were distributed as shown in Fig. 69-3 based on location of the primary lung lesion. This knowledge can help advocates of mediastinal lymph node sampling to target the lymph node stations to be obtained at the time of resection.^13,14 Others perform radical lymphadenectomy routinely with each resection regardless of primary lobar location.¹⁵ This is thought to improve the yield of lymph node sampling, detect noncontiguous lymph node involvement (skip metastases), and improve survival.^8,10,16,17

SURGICAL TECHNIQUE

The technique of mediastinal lymph node dissection was first described in detail by Cahan and colleagues in 1951, when simple pneumonectomy was differentiated from radical pneumonectomy, which included hilar and mediastinal lymph nodes in continuity.¹⁸ Prognosis became linked to lymph node involvement, and lymph node mapping for cancer was established. The technique of mediastinal staging preceding surgical resection has been described.¹⁹ Mediastinal lymph node dissection at the time of resection will be reviewed herein. While originally performed by means of open thoracotomy, this method of lymph node dissection is now possible by means of VATS technique with a greater technical demand yet comparable results.^20,21 Median sternotomy also can be used for extended lymph node dissections.²² It is common to dissect the lymph nodes after pulmonary resection. However, if the decision to proceed with resection would be altered by positive N₂ involvement, the lymph node dissection is undertaken first. Positive findings may warrant chest closure and neoadjuvant treatment before resection or may limit the extent of intended resection.²³ The importance of proper specimen labeling using standard nomenclature cannot be overemphasized. Each lymph node station should be labeled accordingly and submitted separately to avoid confusion among specimens.

Right-Sided Lymphadenectomy

Nodal stations to be addressed from the right chest include stations 2 and 4, stations 7–9, and N₁ stations included in the lobar resection, 10 and 11. The boundaries to the paratracheal nodes (stations 2 and 4) include superior vena cava anteriorly, right brachiocephalic vein and subclavian artery superiorly, trachea and right mainstem bronchus medially, azygos vein and pulmonary artery inferiorly, esophagus and vagus nerve posteriorly, and mediastinal pleura laterally.

Resection of this nodal packet begins with exposure via incision in the fourth or fifth intercostal space. The lung is retracted inferiorly, and the pleura overlying the azygos vein is opened sharply using a parallel incision (Fig. 69-4). The azygos vein is mobilized and can be divided or resected for complete exposure. This can be done easily by using a vascular stapling device. The dissection then is performed from caudal to cephalad using an Allis or Babcock clamp to grasp the lymph node packet and sweep it off the medial structures. The inferior paratracheal station 4 nodes are carefully dissected at the tracheobronchial angle just above the pulmonary artery and pericardium. A flap of pleura is reflected further in a cephalad direction over the superior vena cava, avoiding the phrenic nerve anteriorly, and over the trachea posteriorly. The station 4 nodes are dissected en bloc from their paratracheal position behind the superior vena cava, which can be retracted anteriorly with a Cushing vein retractor or dental pledget. Small perforating veins draining anteriorly to the superior vena cava are identified and controlled, usually with clips or careful electrocautery. Care must be taken posteriorly to avoid the esophagus, vagus nerve, and membranous trachea along the posterior aspect of the station 4 nodal packet. The dissection is completed at the apex just above the level of the aortic arch toward the right subclavian artery, where the station 2 nodes are found. The recurrent laryngeal branch of the vagus nerve has its takeoff in this location and should be avoided.

The landmarks associated with the pyramid-shaped subcarinal nodes of station 7 include the tracheal carina superiorly and anteriorly, the esophagus posteriorly, the tracheal bifurcation into the mainstem bronchi laterally, and the pericardium and upper aspect of the inferior pulmonary vein inferiorly. Dissection at this level can be performed from either side, but the left is more challenging owing to the length of the mainstem bronchus and proximity of the aortic arch. The right-sided resection of station 7 subcarinal lymph nodes begins with retraction of the lung anteriorly (Fig. 69-5). The mediastinal pleura is opened posteriorly between the azygos and hilum. The nodal packet is teased out and grasped with a clamp near the lung parenchyma at the bronchus intermedius and superior to the inferior pulmonary vein. The dissection is carried proximally along the mainstem bronchus. Bronchial arterial branches are controlled with clips or careful electrocautery. Proper orientation of the airway must be maintained to avoid inadvertent injury of the left or right mainstem bronchus, and the esophagus is kept retracted posteriorly.

The remainder of the N₂ nodes includes stations 8 and 9. These can be approached from either side fairly equally. The paraesophageal nodes (station 8) are inferior to the carina along the esophagus, whereas the station 9 nodes are within the inferior pulmonary ligament (Fig. 69-6). Most lung resections include mobilization of the inferior ligament either for lower lobe resection or to improve postoperative lung expansion and avoid a residual apical space. Therefore, the approach to dissection of station 9 nodes is straightforward. The lower lobe is oriented such as to put the inferior pulmonary ligament under tension. The ligament then is carefully mobilized in a cephalad direction away from the diaphragm with a combination of blunt and electrocautery dissection. The ligament is composed of anterior and posterior folds, which soon become apparent during the dissection. The mobilization is directed toward the inferior pulmonary vein, with care taken to avoid injury to the adjacent esophagus. Within the areolar tissues overlying the esophagus are usually several station 9 nodes that are elevated and removed. Hemostasis is achieved with control of the small vessels within the ligament inferiorly. Once the esophagus is exposed, other nodes can be sampled from the paraesophageal station 8 position, if present. Care is taken to avoid the vagus nerves, which run along each side of the esophagus laterally at the subcarinal level, as well as bronchial or intercostal arterial branches at all levels.

Station 10 hilar lymph nodes on the right are located overlying the distal right mainstem bronchus, posterior to the pulmonary artery and inferior to the azygos vein (Fig. 69-7). Exposure is gained by opening the visceral pleura overlying the superior hilum anteriorly with care to avoid the phrenic nerve running along the superior vena cava and pericardium anteriorly. The interlobar station 11 nodes are found near the bifurcation between the right upper lobe and the bronchus intermedius. These are often referred to as "sump" nodes as a result of their interlobar pattern of drainage. Station 12 lobar nodes are encountered when the bronchus is exposed and transected, whereas segmental and subsegmental nodes (stations 13 and 14, respectively) are generally included in the lobectomy specimen.

Left-Sided Lymphadenectomy

Nodal stations to be addressed from the left chest include mediastinal stations 4–9 and N₁ stations included in the lobar resection, 10–14. The boundaries for station 5 and 6 nodes include the phrenic nerve anteriorly, the aortic arch and head vessels superiorly, the vagus nerve and descending thoracic aorta posteriorly, the pulmonary artery and pericardium inferiorly, the ascending aorta and trachea medially, and the mediastinal pleura laterally.

Resection of nodes at stations 5 and 6 begins with a fourth or fifth intercostal approach. The lung is retracted inferiorly, and the mediastinal pleura is opened over the left main pulmonary artery between the phrenic and vagus nerves, and a flap is created from the aortopulmonary window in a cephalad direction along the aortic arch (Fig. 69-8). The prevascular station 6 nodes are found in this location. They can be grasped with an Allis or Babcock clamp, with care taken to avoid the surrounding nerves. Bleeding is controlled with clips or ligatures to prevent an injury from electrocautery. The station 5 lymph nodes in the aortopulmonary window are found more posteriorly along the pulmonary artery near the ligamentum arteriosum. This vestige of the ductus arteriosus may be divided for mobilization of the aorta and pulmonary artery, if needed. The left recurrent laryngeal nerve, which branches from the vagus nerve, must be protected along its course under the aortic arch during dissection. Electrocautery must be minimized or avoided outright. When necessary, the left paratracheal nodes are found along the trachea medial to the ligamentum arteriosum (station 4) and further cephalad above the aortic arch between the left brachiocephalic vein and the left subclavian artery (station 2). These stations, along with station 3, alternatively may be approached posterior to the aortic arch and left subclavian artery by opening the pleura posteriorly and retracting the vessels anteriorly. Intercostal arterial branches need to be divided for this exposure. Care must be given to the recurrent laryngeal nerve as it courses cephalad along the trachea near the esophagus. This method is not widely used in North America.

The dissection for station 7 from the left is similar to the approach from the right. The lung is retracted anteriorly, and the posterior mediastinal pleura is opened along the groove anterior to the descending thoracic aorta, which is retracted posteriorly (Fig. 69-9). The left mainstem bronchus is exposed, and the subcarinal nodal packet is grasped with a clamp and removed. Dissection may be done in blunt or sharp fashion, during which time attention must be given to hemostasis of the bronchial and intercostal perforating vessels. Care also must be exercised in dissecting the subcarinal nodal packet off the main carina itself and the right mainstem bronchus. The esophagus is deep to the bronchus and is retracted posteriorly with the aorta, whereas the pericardium is found at the anteromedial extent of the dissection and inferiorly at the superior portion of the inferior pulmonary vein.

Samples from the paraesophageal (station 8) and inferior pulmonary ligament (station 9) nodes are obtained in similar fashion as on the right side once the posterior pleura is opened anterior to the descending aorta from the carina to the diaphragm. Station 10 on the left is found inferior to the pulmonary artery and posterior to the superior pulmonary vein at the distal left mainstem bronchus. Interlobar station 11 nodes are approached via the fissure with the left upper lobe reflected anteriorly or superiorly. The station 12 lobar nodes are found during dissection of the bronchus prior to lobectomy, and the segmental and subsegmental nodes (stations 13 and 14, respectively) are usually found within the specimen.

TECHNICAL PITFALLS

While most surgeons understand the importance of complete staging of the mediastinum, including assessment made during lung resection, there is variability in approach to nodal staging at different centers (i.e., lymph node sampling versus complete dissection). In part, this is due to the concern that complete lymph node dissection may lead to increased complications or mortality compared with selective lymph node sampling. Furthermore, some surgeons hold the belief that treatment outcome or survival may not be altered with the additional knowledge of pathologic lymph node involvement. Recently, the American College of Surgeons' Oncology protocol #20030 Study Group reported results from a randomized, prospective trial regarding morbidity and mortality after pulmonary resection to determine whether mediastinal lymph node dissection caused an increase in either morbidity or mortality.²⁴ Four-hundred and ninety-eight patients underwent lymph node sampling, and 525 patients had complete lymph node dissection. Patient characteristics were comparable, and all tumors were clinically resectable (T₁ or T₂) with N₀ or nonhilar N₁ nodes and no metastases (M₀). Exclusion criteria included patients who underwent wedge excision and those who received neoadjuvant treatment.

In the lymph node dissection group, mean operative time was 15 minutes longer (p < 0.0001), and the amount of chest tube drainage was greater (1459 mL versus 1338 mL, p = 0.056) than in the lymph node sampling group. Although the chest tubes were left in longer in the lymph node dissection group (5 days versus 4 days, p = 0.495), there was no difference in median length of hospital stay (6 days, p = 0.404). There was no significant difference in operative mortality between groups (p = 0.157), where 10 (2%) lymph node sampling patients died compared with 4 (0.76%) lymph node dissection patients. There were no significant differences in any specific complication between groups, including frequency of chylothorax, postoperative hemorrhage requiring reoperation, number of patients requiring postoperative transfusion, recurrent nerve injuries, air leaks, or bronchopleural fistulas. While the impact of complete mediastinal node dissection on long-term survival is not known from this study, the added dissection did not create significant additional complications when compared with lymph node sampling in this population.

Similar findings were reported by Lardinois and colleagues in a nonrandomized series of 100 patients with clinical T_1-3 and N_0-1 disease.¹⁰ These patients were divided into two groups based on approach: complete dissection versus systematic sampling. While dissection took longer than sampling, there was no significant difference between groups in terms of hospitalization length, morbidity, or overall survival. This study did show a significantly longer disease-free survival after dissection compared with sampling (60 months versus 45 months, p < 0.03) and a significantly higher local recurrence rate after sampling compared with dissection (13% versus 45%, p = 0.02) in patients with stage I disease.

In patients with more advanced disease, there is controversy as to whether or not neoadjuvant treatment is associated with an increased risk of postoperative complications after pulmonary resection accompanied by lymph node dissection.^25,26 There are certain consequences to take into account when carrying out a lymph node dissection in the neoadjuvant setting. First, the timing of surgery is a factor. The patient's performance status must be adequate to tolerate the operation and withstand the postoperative stress associated with recovery. Chemotherapy nadirs should be taken into account, and radiation doses should be monitored. Traditionally, the timing of resection should be no sooner than 4–6 weeks after neoadjuvant treatment to permit the intended response to take effect and avoid acute inflammatory changes and yet not so far removed that postradiation fibrosis changes have begun. The lymph node dissection can be much more tedious after neoadjuvant treatment because, by definition, patients receiving neoadjuvant treatments have bulkier disease to begin with, as well as lymph node sclerosis from the treatment effect. The surgical approach generally is the same as in patients treated primarily, but identifying the proper dissection plane between the nodes and the pulmonary vasculature is often challenging. Furthermore, additional attention should be paid to the bronchus. This is especially true because radiation treatment and nodal dissection can alter bronchial blood supply.^27–29 These associated ischemic changes merit buttressing the bronchial stump to prevent bronchopleural fistula.^30,31 This can be done using any of a number of flaps, including an intercostal muscle pedicle, pericardial or pleural patch, or thymic fat pad.

As with any minimal access surgery, there is a learning curve associated with proficiency in VATS lobectomy. The same holds true for the associated lymph node removal, be it sampling or dissection. However, when the surgeon is versed in open techniques, lymph node dissection by VATS is feasible.³² Innovations in thoracoscopic equipment and technique have enabled minimally invasive thoracic surgery to advance safely. Using the then recently developed devices such as reticulating endoscissors, miniretractors, endoclips, and Harmonic scalpels, Kaseda and colleagues performed 36 VATS lobectomies with lymph node dissection.²⁰ The average number of lymph nodes resected (24, ranging from 10 to 51) was comparable with those from open thoracotomy (22, ranging from 16 to 36). More recently, systematic node dissection by VATS was compared with open thoracotomy in 350 patients undergoing pulmonary resection.²¹ Although the study was nonrandomized and the cohorts were not entirely matched, there were no significant differences in the number of nodes dissected between the VATS and open thoracotomy groups. Operative mortality and morbidity were comparable.

SUMMARY

Determination of lymph node status is a critical component to the staging and treatment of non-small cell lung cancer. Whether or not the surgeon performs systematic lymph node sampling or complete lymph node dissection, some level of nodal sampling is important for comprehensive staging. This practice helps to determine prognosis and may affect survival without added surgical morbidity or mortality. A working knowledge of the lymph node anatomy is necessary. Complete subcarinal and inferior pulmonary ligament dissections should be routine. For right-sided lesions, dissection of the paratracheal nodal packet is necessary. On the left side, dissection of the aortopulmonary and paraaortic stations is compulsory. In experienced hands, lymph node dissection can be done using VATS techniques. Despite advances in endoscopic access to lymph nodes for staging, surgeons are encouraged to complement pulmonary resection with lymph node dissection for comprehensive oncologic surgery.

EDITOR'S COMMENT

The authors have done an excellent job of describing the technique of radical lymphadenectomy. However, I would generally disagree with the view that this procedure should be performed as part of a standard routine. There is no doubt that it provides more lymph node tissue for analysis and may yield better overall accuracy in lymph node staging, but despite the recent findings of the American College of Surgeons' Oncology Study Group, a clear survival advantage has not yet been proved. Careful systematic lymph node dissection is the routine approach at our institution.

–MJK

REFERENCES