Lung cancer remains the single most deadly cancer in the United States with an incidence of 170,000 per year and a 5-year overall mortality of 84%. This poor prognosis is secondary to the fact that most patients are diagnosed at higher stage (III and IV) on presentation.¹,² The best chance for cure is provided by surgical resection of early stage (stage I and II) non-small cell lung cancer (NSCLC). This chapter focuses on surgical management of stage II NSCLC. The treatment, and even the definition, of stage II lung cancer have undergone significant changes over the last decade. Data examining the role of chemotherapy have expanded treatment options for patients with stage II NSCLC through adjuvant and neoadjuvant regimens. Although surgical therapy still remains the mainstay of therapy, a complete understanding of the role of multimodality therapy is crucial in the management of stage II NSCLC.

This chapter will first review the definition of stage II NSCLC along with the proposed modifications to the staging system. The presentation, diagnosis, and staging of stage II lung cancers will be reviewed followed by a discussion of the surgical treatment. Pancoast tumors without nodal involvement, a distinct subgroup of stage II NSCLC, will be discussed separately secondary to their unique presentation and treatment approach.

The TNM classification for lung cancer, developed by Dr. Clifford Mountain in 1972, has been revised once in 1997. More recently, an international effort organized through the International Association for the Study of Lung Cancer (IASLC) has led to a further revision outlined in greater detail earlier in this book.

Stage II NSCLC is currently defined as tumors confined to the lung or obstructing a bronchus >2 cm distal from the carina (T1 to T2) with involvement of nodes within the ipsilateral visceral pleura (N1). Stage II also includes tumors adherent to surrounding structures such as chest wall, mediastinal pleura, diaphragm, or pericardium (T3) without nodal involvement (N0).¹,³ Stage II, therefore, includes T1N1, T2N1, and T3N0 tumors.

The new proposed system for stage II NSCLC will include T1a, T1b, and T2a lesions with N1 disease or T2b or T3 lesions without nodal involvement (N0). T1a are tumors <2 cm, T1b are tumors 2 to 3 cm in size, T2a lesions are 3 to 5 cm in size, and T2b lesions are tumors 5 to 7 cm in size. The term T3 will not only refer to tumors causing postobstructive lung collapse and/or tumors invading the surrounding structures, but will also involve intraparenchymal lesions greater than 7 cm in size or tumors with satellite nodules in the same lobe of the lung (prior T4). N1 definition will remain as nodes within the ipsilateral visceral pleura (stations 10 to 14.)¹,²,³ Thus, large tumors (>7 cm) with negative nodes, previously stage I, will now be considered stage II. Tumors with satellite nodules, which were previously considered T4N0 and therefore stage IIIB, will now be considered T3N0 stage IIB.

These changes to the stage groupings and the T descriptors will clearly have an impact on which patients are now considered resectable. In a recent validation, Kassis et al.³ studied the impact of the new staging system on 1154 pathologically staged patients. Stage assignments changed in a total of 202 patients (17.5%) using the new proposed system, (79 patients [6.3%] upstaged and 129 [11.2%] downstaged). However, it is important to note that no patient was upstaged into a nonsurgical stage (i.e., stage IIIB or IV). Downstaging from a traditionally nonsurgical stage to a surgical stage occurred in 59 patients previously staged IIIB and in 2 patients who had been staged IV. What remains to be determined, however, is what the impact of this redistribution of patients will be on recent adjuvant chemotherapy data. The recent studies demonstrating adjuvant chemotherapies’ effectiveness was performed on stage I to IIIA patients staged using the prior systems. Whether this new system alters those results is yet to be seen.

Clinical Presentation Risk factors for lung cancer include age, history of smoking, and history of lung cancer. Of the 170,000 new patients with lung cancer diagnosed each year in the United States, 90% are older than age 45, and 85% have a history of smoking.¹ When present, clinical symptoms of lung cancer include cough, weight loss, dyspnea, hemoptysis, chest pain, and hoarseness in decreasing order.¹,²,⁴ There is limited data regarding symptoms at presentation of a clinical stage II lung cancer. Therefore, we reviewed our prospective database at the MD Anderson Cancer Center, and found 280 clinical stage II patients over the last 7 years. Symptoms were reported in 69.6% of patients at presentation. The most common symptom for the entire group was cough. For those with T3N0 caused by chest wall invasion, chest wall pain was a very common symptom. The remaining patients presented with an incidental finding of a lung lesion on chest radiograph or a computed tomography (CT) scan.

Imaging Earlier chapters have addressed imaging in lung cancer (see Chapters 26 and 27), but some specific aspects of imaging related to stage II lung cancer are discussed here.

Chest radiograph has traditionally been the modality of choice for screening and evaluating lung cancer. The advantages of chest radiographs include widespread availability, lowradiation dose, and low cost. The disadvantages include low sensitivity and specificity, especially for small tumors. The role of chest radiographs in stage II lung cancer is confined to excluding more advanced stage through identification of a pleural effusion or an extrathoracic bony metastasis.

During the past 20 years, development of the CT technology has made CT scans of the chest a standard of care in the evaluation of lung cancer. Patients with suspected lung cancer should receive CT scan of the chest with inclusion of the liver and adrenals. The latest versions of high-resolution CT scanners provide 1.2-mm thick slice capabilities, and this resolution improves almost yearly. These scans carry a sensitivity of 98%, a specificity of 58%, and an accuracy of 77% in the diagnosis of lung cancer.¹ Specifically with regard to stage II lung cancer, CT scans can assess both the nodal and tumor status of a lesion. The T status including size, ipsilateral/ipsilobar satellite lesions (now considered T3 based on the IASLC staging recommendations), involvement of central lobar structures, and proximity to the carina are easily assessed by CT imaging. However, for the latter two, bronchoscopy remains the gold standard to confirm the lobar involvement and the distance from the carina to the tumor. Pleural and chest wall involvement, distinguishing characteristics of T2 and T3 tumors, respectively, are also addressed by CT scanning. CT scans have an 80% sensitivity and specificity at predicting chest wall involvement using established criteria such as obliteration of the extrapleural fat plane, length of tumor-pleura contact, high ratio of contact length to tumor diameter, and obviously, rib destruction. However, when these criteria are absent and the patient has no symptoms consistent with chest wall involvement (constant or episodic pleuritic chest pain), surgery remains the gold standard approach for elucidating chest wall involvement. If the presence of chest wall involvement is a significant factor to determine respectability (e.g., a patient with marginal performance status and limited cardiopulmonary reserve), a thoracoscopy may be necessary to establish chest wall involvement.

Hilar nodal involvement (N1 disease), a hallmark of most stage II patients, is also assessed by CT scanning although its accuracy is limited. In the absence of significant N1 nodal enlargement, many stage II lung cancers will be difficult to distinguish from stage I lung cancer based on CT alone. Several studies have evaluated CT criteria for predicting lymph node involvement. These studies have focused on the predictive ability of CT for N2 disease (mediastinal nodal involvement) and have a pooled positive predictive value of 0.56.⁴ Extrapolating from this data, one can estimate that a solitary peripheral lung cancer can carry 26% to 44% risk of lymph node involvement (hilar or mediastinal), despite a negative CT scan for nodal enlargement.¹ Data specifically assessing the diagnosis of N1 (hilar) disease is limited, because preoperative determination of N1 disease has limited clinical relevance. Currently, both stage I and II patients are offered surgical resection, and it is only the preoperative identification of N2 (mediastinal) disease that alters treatment. For this reason, most investigators have not addressed the diagnosis of N1 disease specifically. However, if appropriate application of lung sparing techniques, such as stereotactic radiotherapy, radiofrequency ablation, and cryoknife are to be made, then accurate information on the presence or absence of N1 disease will be mandatory to prevent early recurrence within the hilum.

Positron emission tomography (PET) and PET/CT scans have been the latest addition in the armamentarium for the evaluation and staging of lung cancer. PET scans are excellent at discerning metastatic disease; however, their poor spatial resolution makes evaluation of nodal disease less than optimal. Similarly, this poor spatial resolution limits PET’s contribution to the evaluation of the T status. PET/CT has greatly improved the spatial resolution of the PET scan and thus improved the nodal evaluation of the test, although T status is still best evaluated by CT scanning. Where PET/CT scanning excels is in its negative predictive value. A negative CT and PET scan of the mediastinum has a high sensitivity. A recent study by Cerfolio et al.⁵ addressed this issue, where 129 consecutive patients with NSCLC underwent PET and integrated PET/CT. Patients then underwent mediastinal node biopsy, and if negative, went on to pulmonary resection. Their study reported a negative predictive value of 99% for N2 nodal disease. Their positive predictive value was only 49%; however, the positive predictive value is hampered by endemic chronic inflammatory conditions (i.e., histoplasmosis) and acute inflammatory processes (i.e., postobstructive pneumonia secondary to a central lobar tumor). For this reason, PET or PET/CT positive N1 or N2 lymph nodes that are accessible should be biopsied to confirm their involvement, especially when the node is normal or only slightly enlarged on CT imaging.

Bronchoscopy Bronchoscopy, as indicated previously, has a prominent role in stage II disease. As mentioned earlier, it remains the gold standard for evaluation of lobar involvement and proximity to the carina. The treating surgeon should perform the bronchoscopy or be present when it is performed. This is especially true of central lobar T2 lesions that may require a sleeve resection. In this case, accurate assessment of the airway anatomy by the operating surgeon is required to assist with surgical planning.

Mediastinoscopy Evaluation of the mediastinum in stage II lung cancer is paramount. Although imaging has provided significant advances in noninvasive staging, there still remains a significant role for direct biopsy. Mediastinoscopy and its ability to assess mediastinal nodes are discussed in detail in Chapter 29. It remains a vital component in the evaluation of stage II disease. The obvious benefits of assessing ipsilateral and contralateral mediastinal nodes are well established. Many stage II tumors already involve hilar nodes and microscopic nodal involvement in the mediastinum can easily escape detection by imaging. However, an additional benefit of mediastinoscopy is its ability to assess the degree of mediastinal involvement. Mediastinal invasion by central tumors abutting the trachea or the proximal pulmonary artery can be difficult to assess by radiographic imaging. Mediastinoscopy is perfectly suited to evaluate this aspect of a tumor distinguishing a T4 tumor from a T3 or lesser tumor. However, mediastinoscopy, performed in this setting, can lead to scarring and distortion of normal-tissue planes. Although this degree of scarring is limited, it can increase the difficulty of subsequent planned complex resections such as a sleeve resection. If a sleeve resection or central dissection is anticipated, then the mediastinoscopy should be performed concurrently or within 2 to 3 days of the resection.

Endobronchial Ultrasound and Esophageal Ultrasound Similar to mediastinoscopy, endobronchial ultrasound (EBUS) allows sampling of mediastinal nodes for pathologic evaluation. Unlike mediastinoscopy, EBUS is less invasive and allows some hilar (level 10) nodes to be assessed. Esophageal ultrasound (EUS) coupled with transesophageal fine-needle aspiration (FNA) has the capability of reaching paraesophageal and inferior pulmonary ligament nodes (levels 8 and 9) as well as some of the nodes reached by EBUS (levels 7 and less frequently level 4). When combined, these two complimentary procedures have the capability of sampling seven nodal stations²,⁴,⁶,⁷,⁸,⁹,¹⁰ versus only four stations with mediastinoscopy.²,⁴,⁶,¹¹ Discussed in Chapter 29, EBUS is quickly becoming a powerful tool in the evaluation of patients with lung cancer. Its high specificity has been consistently reported, but its sensitivity, especially compared with mediastinoscopy, is still debated. The high resolution of some EBUS and EUS probes can even supplant the evaluation of mediastinal invasion performed by mediastinoscopy.

The complete evaluation of mediastinal nodal involvement and assessment of mediastinal invasion has a clear impact on the subsequent treatment of a lung cancer. Until recently, the preoperative identification of hilar nodal involvement had fewer treatment-related consequences. When the treatment of stage I or II lung cancers were confined to lobectomy or conventional radiation, the preoperative determination of N1 nodal disease was of prognostic value only. However, the use of lobectomy as the sole treatment of subcentimeter node negative peripheral tumors is no longer the sole modality offered. Lung sparing techniques including stereotactic radiotherapy, percutaneous treatments such as radiofrequency ablation, or techniques such as wedge resections or segmentectomies (see also Chapters 32, 37, and 43), may be appropriate options. In these cases, preoperative identification of N1 involvement will play a significant role in treatment decisions. Similarly, the use of neoadjuvant chemotherapy, although currently not extensively utilized for early stage lung cancer, may be an option in the future. Again, pretreatment identification of N1 disease may be paramount in making these treatment-related decisions. The role of EBUS and EUS will undoubtedly increase in the evaluation and confirmation of stage II lung cancer.

All patients with stage II NSCLC without signs of metastatic disease should be evaluated for surgery. Exercise testing and preoperative evaluation of the patient with lung cancer is applicable to patients with stage II disease. Considerations that are specific for stage II disease stem from the frequently central location of some of these tumors. The involvement of a lobar bronchus by either the primary tumor (T2) or a hilar lymph node (N1) is encountered in stage II disease. These tumors can be approached by a sleeve resection rather than a pneumonectomy, and this should be considered when calculating the postoperative pulmonary reserve. In addition, the distal collapse caused by large >7-cm tumors (T3) or the lobar obstruction requires a split perfusion scan to be performed when calculating the postoperative pulmonary reserve. Additional considerations applicable to Pancoast tumors are discussed later in this chapter.

Surgical Approach Possible approaches include an open thoracotomy or a minimally invasive approach. The open thoracotomy remains the dominant approach for stage II disease, and it can include various methods. These include the posterolateral thoracotomy with division of the latissimus muscle and ± the serratus muscle. Muscle-sparing techniques are also quite common and include a posterior, anterior, or axillary approach. In rare cases a sternotomy, a hemiclamshell incision can also be utilized. These incisions are useful for centrally placed tumors that abut the anterior mediastinum.

The minimally invasive approaches have included several approaches that are nothing more than an open thoracotomy with an additional port for the thoracoscope to a completely thoracoscopic procedure. The definition most widely accepted for minimally invasive pulmonary resection was outlined in the feasibility trial of video-assisted thoracic surgery (VATS) lobectomy (Cancer and Leukemia Group B [CALGB] 39802).¹² The study mandated no rib spreading; a maximum length of
8 cm of the access incision for removal of the lobectomy specimen; individual dissection of the vein, arteries, and airway for the lobe in question; and standard node sampling or dissection (identical to an open thoracotomy). All specimens were placed in an impermeable bag and removed through the access incision. Although other approaches and techniques that have been reported included minimal rib spreading and hilar ligation, the authors agree with the intergroup investigators in maintaining a rigorous adherence to the principles utilized in an open thoracotomy.

The use of a VATS lobectomy in stage II disease may be limited by the tumor size. To effectively remove a tumor without rib spreading, most authors feel that the primary must be less than 5 to 6 cm in greatest diameter. Of course, flat tumors larger than 6 cm can be removed if the specimen is appropriately oriented in the impermeable bag. Additional concerns about a VATS approach in stage II disease center on the dissection of enlarged hilar nodes. Although lymph node dissection via a VATS approach has been shown to be very effective,¹³ the intergroup investigators did comment on factors that required conversion to an open procedure. In 25% of the case (4 of 12), conversion was secondary to the difficulty dissecting lymph nodes from the vascular structures. Although this is not an absolute contraindication for a VATS approach, the surgeon should consider the size of the hilar nodes, the presence of calcifications or scarring from prior induction therapy, and the technical experience with a VATS approach.

Open Thoracotomy Although there are several types of open thoracotomy, the most commonly employed versions include a posterolateral thoracotomy, a posterior muscle-sparing thoracotomy, and perhaps an axillary thoracotomy. All are performed with the patient in a lateral decubitus position. Attention to the method of pain control should also be given with the most common approach being a thoracic epidural. Local delivery of analgesics with a self-contained pump and catheter positioned in the extrapleural space are also gaining acceptance.

Although each has slight advantages and disadvantages when compared with one another, they are all appropriate for resection of most stage II lung cancers. The posterolateral thoracotomy is the most commonly employed incision and is the gold standard with regard to hemithoracic exposure. Its primary drawback remains the pain associated with this large incision and the division of the latissimus dorsi muscle, which has implications in postoperative function and the muscle’s availability as a rotational flap. The posterior muscle-sparing thoracotomy uses the posterior two thirds of the posterolateral thoracotomy incision and mobilizes the latissimus dorsi and serratus anterior muscle anteriorly. Although the incision is slightly smaller, almost all procedures performed via a posterolateral thoracotomy can be safely approached through this incision. Exceptions, although not absolute, include complex sleeve resections and chest wall resections. The preservation of the latissimus dorsi muscle (and perhaps the speed of closure) is its primary advantage. The axillary thoracotomy utilizes either a vertical incision in the low axilla aligned with the anterior border of the latissimus dorsi muscle or a transverse incision centered on the anterior border of the latissimus dorsi muscle. The fibers of the serratus anterior muscle are separated exposing the third or fourth intercostal space through which the thorax is entered. The entry through the lateral aspect of the chest wall, where the ribs are most mobile, allows for distraction of the ribs with the least force being applied to the costovertebral junction. The main disadvantage is the limited exposure of lower lobe lesions.

For stage II tumors, each of these approaches can be used at the discretion and comfort of the surgeon. Patients with significant hilar disease or central lesions requiring sleeve resections may be best approached through a larger posterolateral thoracotomy. The limitations of an axillary thoracotomy make this approach the least utilized incision.

Video-Assisted Thoracic Surgery In the past, the definition of a VATS resection has been variable in the literature. The range of definitions has included everything from a limited posterior muscle-sparing thoracotomy with thoracoscopic visualization, to a completely thoracoscopic approach without any rib spreading. Most thoracic surgeons today will agree that a true VATS resection should include absolutely no rib spreading, although the conduct of the resection should duplicate an open resection. The technical considerations for this approach have been discussed in detail previously in this text.

Although the effectiveness of this technique for early stage malignancies seems apparent, there are some concerns when applying the technique to stage II malignancies. Enlarged hilar N1 nodes, especially where there is concern of arterial or bronchial invasion, can magnify the technical complexity of a VATS resection. Although reports of VATS sleeve resections and pneumonectomies exist, these require advanced skills and will be adopted more slowly. A more practical limitation of the VATS technique in stage II malignancies is the size of the tumor that can be physically removed. Tumors larger than 5 to 6 cm cannot be removed through the utility thoracotomy without rib spreading. Correspondingly, these large tumors may limit visualization of and accessibility to the hilar structures increasing the technical complexity of the resection. Finally, the utility of combining a chest wall resection with VATS lobectomy may be questionable. Reports of chest wall resections accomplished by VATS do exist. However, the main advantage of a VATS resection, which is of rapid recovery of function, may be lost when coupled with a chest wall resection.

Extent of Resection Although some debate exists regarding the extent of resection for stage I tumors, the options are less contentious for stage II disease. The frequent involvement of hilar lymph nodes in stage II disease mandates a resection that encompasses the nodal basin of the tumor. Although the new staging system includes some tumors that are without nodal involvement (T2bN0 and T3N0), the large size of these tumors (greater than 5 cm) and their aggressive features (chest wall invasion and interlobar satellite lesions) would make lung-sparing techniques difficult or less advantageous. For these reasons, there
is relatively uniform agreement that an anatomical lobectomy with complete hilar nodal dissection should be performed.

For the same reasons mentioned previously, most surgeons would also advocate a systematic mediastinal nodal dissection rather than a nodal sampling. However, definitive data demonstrating an advantage from complete nodal dissection on survival has yet to be established. The ACOSOG Z0030 trial randomizing clinical stage I patients to systematic node dissection versus node sampling have not yet reported long-term results. The early results exhibited no difference in morbidity associated with the performance of a systematic node dissection.¹⁴ Mediastinal nodal involvement was identified in 3.8% of clinical stage I patients. Considering the higher likelihood of occult N2 disease in clinical stage II patients and the lack of evidence for increased morbidity associated with complete lymph node dissection, we would endorse complete lymph node dissection in all stage II patients.

Pneumonectomy Pneumonectomy, or the complete removal of an entire lung, entails the division of the ipsilateral main pulmonary artery, both pulmonary veins and the mainstem bronchus. Ipsilateral hilar nodes are removed en bloc with the lung, and the mediastinal lymph nodes are removed as part of a systematic nodal sampling or dissection. A conventional pneumonectomy implies division of the vessels within the thoracic cavity. An intrapericardial pneumonectomy entails entry into the pericardium for division of any of the pulmonary vessels.

The definition of stage II disease includes several features that can require a pneumonectomy for complete resection. The most common is hilar nodal disease with concomitant invasion of the mainstem bronchus or the proximal pulmonary artery. This is especially true when the primary tumor resides in the lower lobes and the nodal disease involves the upper lobe bronchi or vasculature. Other reasons are large tumors (>7-cm tumors now classified as T3) that cross the fissure and involve the upper and lower lobes or tumors that obstruct the mainstem bronchus not amenable to a sleeve resection.

Bilobectomy As the name implies, a bilobectomy is the removal of two lobes and can only be performed on the right lung. The indication for such a procedure is a large tumor that extends across the fissure to involve a substantial portion of an adjacent lobe. Additionally, hilar nodal disease that invades the central bronchi of two lobes can also necessitate bilobectomy. The later indication is not uncommonly seen in stage II disease.

Lobectomy The definition of a lobectomy is, as expected, the removal of the entire lobe of a lung. The vascular and bronchial branches are individually identified and ligated along with the resection of any hilar lymph nodes surrounding these branches. Intraparenchymal lymph nodes associated with the lobe are removed en bloc with the specimen. Similar to a pneumonectomy, the mediastinal lymph nodes are removed as part of a systematic nodal sampling or dissection.

Lobectomy is the most common procedure for stage II lung cancer. The frequent central location of the tumor or the associated hilar nodal disease necessitates resection of the entire lobe. Even if the hilar nodal disease can be resected without sacrificing the lobar structures, most thoracic surgeons would not advocate a wedge or segmental resection of the primary tumor. A lobectomy should be performed to remove the entire nodal draining basin.

Wedges or Segmental Resection These techniques resect a portion of lung surrounding the lesion or a segment of the lung. Segments are anatomically distinct with 10 named segments in the right lung and 8 in the left lung. These segments have a discrete arterial, venous, and bronchial anatomy, and a true segmentectomy would involve their isolation and division similar to a lobectomy.

As discussed earlier, these lung-sparing techniques are not frequently an option in stage II disease. The need to resect the hilar lymph nodes and their parenchymal nodal basin usually necessitates a lobectomy at the least. However, in some cases, such as a chest wall lesion without nodal disease (T3N0), one of these lung-sparing techniques can be used especially in patients with marginal pulmonary function.

Sleeve Resections Sleeve resection, also referred to as bronchoplastic procedures, are techniques that preserve normal lung tissue when the lesions are small but located at or near the main bronchi. The technique involves resection of a portion of the main bronchus followed by an end-to-end anastomosis to preserve the lung tissue distal to the resection. Examples are shown in Figure 34.1.¹⁵ In stage II disease, common indications for sleeve resection are involved hilar nodes that invades the origin of the lobar bronchus or small central tumors within 2 cm of the carina without lymph node involvement (T3N0). Although these tumors can also be resected with a pneumonectomy, there is clear data that the morbidity and mortality following a sleeve resection is less than a pneumonectomy. The added oncologic benefit of a pneumonectomy has not been clearly demonstrated.

Chest Wall Resections Isolated chest wall involvement outside of the superior sulcus is also encountered in lung cancer. In the current staging system, a chest wall lesion without lymph node involvement is considered stage II disease, in the proposed changes to the staging system a T3 tumor with hilar nodal disease (N1) will also be considered stage II. Both of these tumors could be offered surgical resection. The goal of resection is an en bloc anatomical resection. Most investigators have identified completeness of resection as a prognostic factor in long-term survival.¹⁶,¹⁷,¹⁸ The depth of invasion seems to be less predictive of survival.

There are areas of controversy that exist regarding the treatment of lung cancer invading the chest wall. The first is the use of an extrapleural dissection for tumors limited to parietal pleural involvement. Several publications have decried the use of this technique reporting worse outcomes.¹⁹,²⁰ Although the series from Memorial Sloan Kettering²¹ failed to demonstrate a difference, the authors stated that patients with any suspicion of parietal pleural involvement were addressed with
a chest wall resection. Extrapleural resection was reserved only for those with filmy inflammatory adhesions to the chest.