Paraneoplastic syndromes occur in approximately 10% of patients with lung cancer (Table 66-1). Such syndromes may be the initial presenting complaint triggering an evaluation but also can develop late in the course of disease. Paraneoplastic syndromes are unrelated to direct invasion or distant spread of tumor and in and of themselves do not preclude curative-intent therapy. Endocrine syndromes associated with lung cancers often are characterized by tumor production of biologically active hormones. Lung cancer is the most common cause of cancer-associated hypercalcemia, hyponatremia, and syndromes involving ectopic production of adrenal corticotropic hormone (ACTH).

Table 66-1 Paraneoplastic Syndromes Associated with Lung Cancer

Endocrine	Syndrome of inappropriate secretion of antidiuretic hormone (SIADH)/hyponatremia; hypercalcemia; ectopic adrenocorticotropic hormone (ACTH) syndrome; Cushing syndrome; hyperglycemia; hypoglycemia; hyperthyroidism; carcinoid syndrome; gynecomastia; elevated growth hormone; elevated follicle-stimulating hormone (FSH); galactorrhea
Musculoskeletal	Hypertrophic osteoarthropathy (HOP); clubbing, myopathy; dermatomyositis; polymyositis
Neurologic	Lambert-Eaton myasthenic syndrome (LEMS); encephalomyelitis/subacute sensory neuropathy; cerebellar degeneration; opsoclonus-myoclonus; autonomic neuropathy; retinopathy; mononeuritis multiplex; peripheral neuropathy; myopathy
Skin	Acanthosis nigricans; pruritus and urticaria, erythema multiforme; erythroderma; exfoliative dermatitis; hyperpigmentation
Hematologic	Anemia, thrombocytosis, leukocytosis, hypercoagulable state

Hypercalcemia of malignancy is the most frequent of these syndromes and is seen most commonly in patients with squamous cell carcinoma. Hypercalcemia in lung cancer patients usually is related to the ectopic production of parathyroid hormone–related peptide (PTHrP), rather than to the osteolytic effects of bone metastases. Early symptoms—thirst, polyuria, fatigue, constipation, nausea—are nonspecific. Mental status changes, lethargy, and even coma may accompany more severe hypercalcemia. The syndrome of inappropriate secretion of antidiuretic hormone (SIADH) most commonly occurs in patients with small cell lung cancer (SCLC). In some SCLC series, hyponatremia related to SIADH is present in up to 10% to 15% of patients. As with hypercalcemia, the symptoms—weakness, nausea, headache—often are nonspecific; more severe hyponatremia may have serious consequences, including confusion, seizures, and coma. Ectopic ACTH syndrome is seen most commonly with SCLC. The classic features of Cushing syndrome—truncal obesity, myopathy and muscle weakness, diabetes, hypertension, hirsutism—often are absent. In most cases, tumor cells express a precursor hormone that is cleaved to ACTH. A minority of patients produce corticotropin-releasing hormone, which stimulates ACTH production in the pituitary. The distinction between these two processes can be determined by a dexamethasone suppression test.

Nonendocrinologic extrapulmonary syndromes include musculoskeletal abnormalities, most commonly asymptomatic digital clubbing, which can be seen in isolation or in the setting of hypertrophic pulmonary osteoarthropathy (HPO). The pathogenesis of HPO is unknown, but the disorder consists of a proliferative periostitis characterized by symmetric, painful arthropathy that typically involves the ankles, shins, knees, wrists, and elbows. The diagnosis usually is confirmed by the identification of new periosteal bone formation on plain radiographs of the long bones, or by demonstration of diffuse long bone uptake of radionuclide on bone scan or positron emission tomography (PET) imaging. HPO is more commonly seen with adenocarcinoma. Hematologic dyscrasias commonly occurring in patients with lung cancer include anemia, which may compound fatigue and dyspnea; thrombocytosis; and leukocytosis. Eosinophilia is seen only rarely. Lung cancer is the most common cause of hypercoagulability associated with malignancy, which usually declares itself as deep venous thrombosis or thromboembolism, or with classic Trousseau syndrome (migratory superficial thrombophlebitis). Lung cancer, and specifically SCLC, also is the most common cause of a clinically diverse group of paraneoplastic neurologic syndromes (see Table 66-1).

The commonality with several of these syndromes is that they appear to be driven by autoimmune mechanisms. The most common of these is the Lambert-Eaton myasthenic syndrome (LEMS), which occurs in approximately 2% to 4% of patients with SCLC. LEMS is characterized by proximal muscle weakness, hyporeflexia, dysarthria, blurred vision, and autonomic dysfunction. In contrast with myasthenia gravis, muscle strength typically does not improve with the administration of anticholinesterases, but some recovery of strength may be obtained with treatment of the underlying malignancy. LEMS is associated with an antibody that inhibits acetylcholine release by binding to calcium channels in peripheral cholinergic nerve terminals. The diagnosis usually is based on electromyography demonstrating small amplitude of the resting muscle action potential, which increases with repeated nerve stimulation or exercise. The syndrome of encephalomyelitis–subacute sensory neuropathy is associated with antineuronal nuclear-antibody types 1 and 2 (ANNA-1 and ANNA-2), also called anti-Hu and anti-Ri antibodies, respectively, which react with SCLC tumor cell surface proteins and with neuronal nuclear antigens. Anti-Hu antibodies also are seen in patients with SCLC and cerebellar degeneration as well as opsoclonus-myoclonus. Other paraneoplastic neuropathic syndromes include encephalomyelitis, autonomic neuropathy, and cancer-associated retinopathy.

Asymptomatic Presentation

Approximately 25% of patients with lung cancer present at an early stage of their disease, so only a minority are asymptomatic at the time of diagnosis. The poor overall 5-year survival for patients with lung cancer reflects the disproportionate number of patients who are diagnosed at an advanced stage. Interest in establishing screening for early disease detection in asymptomatic patients has understandably been the focus of many studies. The reasons to develop a reliable lung cancer screening tool are quite evident. Most lung cancers are discovered because of symptom-driven evaluation; a diagnosis of advanced-stage disease unfortunately portends a limited prognosis. By contrast, asymptomatic patients with early-stage cancers are more likely to experience long-term survival with treatment.

Large studies performed since the 1970s evaluating chest radiography, sputum cytologic analysis, and low-radiation-dose chest computed tomography (CT) scanning for lung cancer screening have consistently demonstrated an increase in the number of lung cancers diagnosed and a shift toward identifying earlier-stage disease accompanied by an improvement in survival, but notably without a decrease in lung cancer mortality rates. The National Lung Screening Trial (NLST) is the first large-scale clinical trial to demonstrate a benefit in lung cancer mortality related to screening with low-dose chest CT scanning. The NLST enrolled 53,000 subjects aged 55 to 74 years who were current or former smokers with at least a 30 pack-year history, and who had no previous history of lung cancer. The study compared low-dose chest CT scanning and chest radiography as screening tools, with participants randomized to receiving three annual screens with one or the other modality, with follow-up over another 5 years. The trial was closed in late 2010 when initial results demonstrated that subjects assigned to the CT screening group had 20% less lung cancer deaths than those assigned to chest radiography screening. The disadvantages of screening include a high rate of false-positive findings, which incur additional diagnostic evaluation with the potential for risk or harm related to such evaluation; the anticipated high cost of screening and subsequent evaluation; and the potential consequences of cumulative exposure to radiation from multiple scans. The optimal approach to lung cancer screening inevitably will need to evolve over the coming years, as the results of the NLST are applied to clinical practice, and outcomes related to both benefit and disadvantages can be longitudinally examined.

The Solitary Pulmonary Nodule

The typical presentation of an asymptomatic early stage lung cancer is as a solitary pulmonary nodule (SPN), defined as a solitary lesion 3 cm or less in diameter, surrounded by normal lung and not associated with other thoracic abnormalities such as hilar or mediastinal adenopathy, atelectasis, or pleural effusion. SPNs are common radiographic lesions, frequently identified as incidental findings on chest imaging studies done for issues unrelated to lung cancer. Typical scenarios include routine preoperative evaluation in which a lung nodule or mass is found on the plain chest radiograph, or is identified on a chest CT scan performed for screening purposes or for other reasons unrelated to any chest symptoms, and investigation featuring a CT scan of the abdomen, heart, or spine, in which a portion of the lungs is almost inevitably included.

In evaluating SPNs, it is useful to distinguish small (8 mm or less in diameter) from larger nodules. Small SPNs are very common findings in patients who have participated in CT screening studies for lung cancer, as well as in those undergoing CT scanning for non-lung-related reasons, as noted earlier. In the feasibility study for the NLST performed within the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial, 21% of subjects had abnormalities identified on the baseline screening CT scan, most of which were small SPNs. Similarly, in the Mayo Clinic lung cancer screening study, 51% of subjects had abnormalities found on the baseline screening study, with an increase to 69% by the second annual screen, the vast majority of which were small SPNs. The enormous number of small SPNs generated by an increased volume of CT scanning prompted a position statement from the Fleischner Society proposing guidelines for the management of small (8 mm or less) SPNs detected on CT scans. The recommendations are outlined in Table 66-2. Appropriately applied, the guidelines outline algorithms for patients based on their risk of lung cancer, providing timelines for follow-up of incidentally discovered nodules that minimize the number of CT scans necessary in the course of evaluation. These practical recommendations will become even more relevant if CT screening for lung cancer becomes the standard of practice. Of note, these recommendations apply only to patients in whom SPNs are discovered incidentally, unrelated to any known underlying disease, and only for solid SPNs 8 mm or less in diameter. Specifically, the guidelines are not intended to apply to patients known to have or suspected of having malignant disease, patients younger than 35 years of age, or patients with unexplained fever. Furthermore, the recommendations are not fully applicable to persons with nonsolid (ground glass appearance) or partially solid nodules, who may require longer follow-up to permit exclusion of biologically indolent cancers with greater confidence.

Table 66-2 Fleischner Society Guidelines for Management of Small Pulmonary Nodules Detected on Computed Tomography (CT) Scans

Nodule Size*	Low-Risk Patient†	High-Risk Patient‡
≤4 mm	No follow-up needed§	Follow-up CT at 12 months; if unchanged, no further follow-up¶
>4-6 mm	Follow-up CT at 12 months; if unchanged, no further follow-up¶	Initial follow-up CT at 6-12 months, then at 18-24 months if no change¶
>6-8 mm	Initial follow-up CT at 6-12 months, then at 18-24 months if no change	Initial follow-up CT at 3-6 months, then at 9-12 months and 24 months if no change
>8 mm	Follow-up CT at around 3, 9, and 24 months, then dynamic contrast-enhanced CT, PET, and/or biopsy	Same as for low-risk patient

NOTE: Guidelines refer to newly detected indeterminate nodules in persons 35 years of age or older.

PET, positron emission tomography.

* Average of length and width.

† Minimal or absent history of smoking and of other known risk factors.

‡ History of smoking or of other known risk factors.

§ The risk of malignancy in this category (<1%) is substantially less than that determined with a baseline CT scan in an asymptomatic smoker.

¶ Nonsolid (ground glass) or partly solid nodules may require longer follow-up period to exclude indolent adenocarcinoma.

Modified from MacMahon H, Austin JH, Gamsu G, et al: Guidelines for management of small pulmonary nodules detected on CT scans: a statement from the Fleischner Society, Radiology 237:395–400, 2005.

Larger SPNs (8 to 30 mm) require further investigation. A review of any previous imaging studies should be the first step, because an established baseline can provide information important to the evaluation. A nodule that is identified as new or growing warrants clinical concern, whereas radiographic stability of a solid SPN over a 2-year period generally is considered to constitute reasonable evidence of benignity. The exception to this general rule is that it is increasingly evident that a 2-year period of stability is insufficient to allow confident designation of benignity for nodules that are purely of ground glass density. Other CT features also may reliably exclude malignancy, including the presence of fat identifying a hamartoma, complete or laminated calcification of the nodule, the presence of a feeding artery and a draining vein identifying a pulmonary arteriovenous malformation, the “comet tail” sign of swirling bronchovascular structures pointing to the hilum characteristic of rounded atelectasis, and the movement of a nodule within a cavity with positional changes characteristic of a mycetoma.

The larger the nodule, the more likely it is to be a cancer. A majority of asymptomatic pulmonary masses greater than 30 mm in diameter are malignant. Figure 66-2 outlines a management strategy for SPNs 8 to 30 mm in diameter recommended by the American College of Chest Physicians evidence-based clinical practice guidelines for lung cancer. The algorithm is appealing in its simplicity. The key step is the clinical definition of a pretest probability of malignancy for every patient with an SPN, because this will inform decisions about the extent of evaluation. SPN factors associated with malignancy include larger size, positive smoking history, older age, a previous history of cancer, the location within the lung (higher risk with upper lobe location), and nodule edge characteristics (higher risk with spiculated than with smooth borders) (Figure 66-3). Mathematical prediction models incorporating these factors typically perform similarly to the clinical acumen of experienced physicians. Patients with a low probability (less than 5%) of malignancy can be monitored by serial CT imaging. Patients with a high probability (more than 60%) of malignancy should proceed to definitive surgical biopsy or resection. For the large group of patients with an intermediate probability (5% to 60%) of malignancy, the use of a variety of noninvasive and invasive modalities should be considered in deciding whether the SPN should be observed, biopsied, or removed.

Figure 66-2 Management algorithm for patients with solitary pulmonary nodules (SPNs) 8 to 30 mm in diameter. CT, computed tomography; CXR, chest radiograph; PET, positron emission tomography; XRT, radiotherapy.

(From Gould MK, Fletcher J, Iannettoni MD, et al: Evaluation of patients with pulmonary nodules: when is it lung cancer? Chest 132:108S–130S, 2007.)

Figure 66-3 Asymptomatic solitary pulmonary nodules visualized on computed tomography. A, A 12-mm left upper lobe nodule with spiculated borders in a 74-year-old man with a 50 pack-year smoking history. Surgical resection revealed adenocarcinoma, pT1aN0M0, stage Ia. B, A 14-mm left lower lobe nodule with smooth border in a 64-year-old man who was a lifelong nonsmoker. Histopathologic/cytologic evaluation revealed a hamartoma.

Noninvasive imaging modalities commonly used to evaluate SPNs include CT and PET imaging. As already described, certain CT features may establish a benign diagnosis or support the rationale for a period of watchful waiting. By contrast, other findings may more strongly suggest malignancy, including spiculated nodule borders, a dilated bronchus leading into the nodule, and cavitation associated with a thick and irregular wall. CT with dynamic contrast enhancement, in which the demonstration of an increase in Hounsfield units with the administration of contrast is associated with a higher likelihood of malignancy, may be useful in institutions with proficiency in this technique. PET scanning with 18-fluorodeoxyglucose (FDG) is increasingly utilized in the characterization of SPNs. Increased glycolysis is a well-described metabolic characteristic of malignant cells, resulting in enhanced uptake of glucose and FDG. FDG accumulates in these cells because it cannot be completely metabolized.

In SPNs greater than 10 mm in diameter, the sensitivity of PET for identifying malignancy is high (80% to 100%), although specificity is less robust. PET is less sensitive for nodules measuring less than 10 mm in diameter, and its use for evaluating the likelihood of malignancy in small nodules should be discouraged. This modality may yield false-positive results in the setting of inflammation or infection, including tuberculosis, fungal infections, rheumatoid nodules, and sarcoidosis, all of which may mimic lung cancer radiographically. False-negative findings typically are described for patients with well-differentiated tumors, including adenocarcinoma in situ (bronchioloalveolar carcinoma) and carcinoid tumors, for which the tumor metabolic rate presumably is low. PET may be most useful in a situation in which the clinical suspicion for malignancy is low despite suggestive CT findings; in this case, a positive result on PET scanning would lower the threshold to pursue some type of invasive evaluation, whereas a negative result would add reassurance regarding a period of observation.

Invasive modalities of assessing SPNs include surgical removal, bronchoscopic biopsy or aspiration, and percutaneous transthoracic needle aspiration (TTNA). For asymptomatic patients in whom the clinical pretest probability of malignancy is very high and for whom the risk of surgery is acceptable, a strong argument can be made to proceed directly to surgical resection, because this strategy would offer the most efficient, albeit most invasive, approach to establishing a diagnosis and stage, while also potentially providing definitive therapy. In many situations, however, a limited nonsurgical biopsy may be warranted—for example, in patients in whom a benign diagnosis is a strong consideration and a nonsurgical biopsy may potentially be definitive (e.g., tuberculosis or hamartoma), patients in whom surgical resection carries a high risk for complications, and patients who are unwilling to accept a surgical approach without at least an initial attempt to establish a definitive diagnosis. Before bronchoscopy or TTNA is undertaken, a discussion should be held with the patient regarding the possible outcomes, and in particular to consider the approach to be taken should the biopsy be nondiagnostic. This discussion is of particular importance, because a definitive benign diagnosis often is difficult to establish with these procedures, which frequently provide limited biopsy material. Bronchoscopic biopsy has a reasonable diagnostic yield (55%) in nodules greater than 20 mm in diameter, with the yield increasing as nodule size increases and if radiographic evidence of a bronchus leading into the nodule (“bronchus sign”) is present. For peripheral nodules less than 20 mm, the diagnostic yield with bronchoscopy is considerably lower, in the range of 10% to 50%, which may be increased by incorporating radial ultrasound or electromagnetic navigational technology to enhance accurate localization.

Percutaneous TTNA commonly is performed under CT or fluoroscopic guidance, with improved diagnostic capability when a core needle biopsy is done in addition to aspiration. TTNA typically is considered for SPN evaluation when bronchoscopy is unlikely to render a diagnosis, usually for SPNs that are relatively small (less than 20 mm) and peripherally located. The diagnostic yield of TTNA for malignant lesions is higher (up to 85%) in studies in which the prevalence of malignancy is high and nodule size is 20 to 40 mm. For smaller nodules, the yield falls, with nondiagnostic biopsy samples obtained in up to 41% of patients with SPNs less than 40 mm in diameter, and in particular when the lesion is benign. The major drawback of TTNA is a high rate of pneumothorax (10% to 35%). Approximately 5% of procedures are complicated by pneumothorax of enough clinical significance to warrant placement of a chest tube. TTNA may pose a higher risk in situations in which the lung traversed by the needle en route to the lesion is emphysematous or bullous, or in patients who are unable to hold their breath as directed during the procedure.

Lung Cancer Staging

Staging is an essential and critical part of the management of every patient suspected of having or diagnosed with lung cancer. The staging classifications of the American Joint Committee on Cancer for non–small cell lung cancer (NSCLC) and SCLC now both follow the tumor-node-metastasis (TNM) system. The current seventh edition, published in 2010, was developed by the International Association for the Study of Lung Cancer (IASLC) using an international database of over 80,000 lung cancer cases. The distribution of database cases informing the staging classification spanned four continents: Europe 58%, North America 21%, Asia 14%, and Australia 7%. Although the database still included a majority of white patients and thus may be limited in application to nonwhite patient populations, the magnitude and diversity of the IASLC database constituted a remarkable improvement over the previous lung cancer staging system, which had been developed in a largely single-institution North American patient cohort.

SCLC previously was staged according to the Veterans Administration Lung Cancer Study Group (VALSG) two-stage system. In this system, “limited-stage” disease refers to cancer confined to one hemithorax, which may include the ipsilateral lymph nodes. From a practical perspective, a limited-stage SCLC is confined to an area that can be encompassed in a single radiation treatment field. “Extensive-stage” SCLC includes any tumor extending beyond the confines of limited-stage disease, including tumor confined to one hemithorax but involving the pleura. Although this two-stage system is no longer the official staging system for SCLC, it is still clinically appealing in its relatively easy application to treatment planning.

Different types of staging may be used to describe a malignancy during a course of evaluation (Table 66-3). The most common staging assessments are the clinical stage (indicated by the prefix c), defined as the stage before treatment as determined by all available imaging and biopsy data, and the pathologic stage (indicated by the prefix p), defined as the stage determined on the basis of the available imaging and biopsy data, as well as analysis of tissue samples obtained after a surgical resection performed with intent to cure. The completeness of resection is determined histopathologically after surgery (Table 66-4). Clinical staging typically is associated with poorer outcomes than those achieved with pathologic staging, the latter being inherently more accurate. All patients will undergo clinical staging; only in a select number will their disease be able to be staged pathologically. Restaging may take place after treatment, at the time of a recurrence, or at autopsy, identified by the prefix y, r, or a, respectively, as outlined in Table 66-3.

Table 66-3 Designations for Types of Staging Assessments

Prefix	Name	Definition
c	Clinical	Before initiation of any treatment, using any and all information available (e.g., including mediastinoscopy)
p	Pathologic	After resection, based on pathologic assessment
y	Restaging	After part or all of the treatment has been given
r	Recurrence	Stage at time of a recurrence
a	Autopsy	Stage as determined at autopsy

From Detterbeck FC, Boffa DJ, Tanoue LT: The new lung cancer staging system, Chest 136:260–271, 2009.

Table 66-4 Staging Designation for Completeness of Surgical Resection

Symbol	Name	Definition
R0	No residual	No identifiable tumor remaining; negative surgical margins
R1	Microscopic residual	Microscopically positive margins but no visible tumor remaining
R2	Gross residual	Gross (visible or palpable) tumor remaining

From Detterbeck FC, Boffa DJ, Tanoue LT: The new lung cancer staging system, Chest 136:260–271, 2009.

The TNM paradigm is based solely on anatomy (Table 66-5). The T descriptor characterizes the primary tumor, the N descriptor, involvement of the regional lymph nodes; and the M descriptor, the presence or absence of distant metastases. Combinations of these descriptors are grouped into stages. Even though limitations to the generalized application of the TNM system are increasingly appreciated, it serves as a consistent structure to describe extent of disease and to provide prognosis. Furthermore, staging provides a platform for comparisons between reasonably homogeneous populations in therapeutic clinical trials. It may be tempting to use the clinical stage as the major determinant of therapy, but it is important to recognize that the staging classification system is not intended to be an algorithm for assigning treatment. The designation of stage inherently portends prognosis, because the separation of TNM combinations into stages is fundamentally determined by survival. Two of the limitations of the TNM system become immediately evident. First, the anatomic descriptors yield no information relating to biologic behavior of the cancer. With a burgeoning recognition that the molecular characteristics of an individual tumor may drive outcomes and more specifically render some tumors very responsive to targeted treatments, the prognostic ability of the TNM system becomes less reliable. Second, patients who share similar survival outcomes may have clinically dissimilar cancers; this divergence becomes clear on examining the stage groupings outlined in Table 66-6. Within a given stage of disease, the patients often are not clinically homogeneous, even though they may have similar survival rates. Despite the large size of the IASLC database, outcomes for specific TNM groups with small numbers of patients could not be adequately evaluated, with an inevitable “lumping” rather than “splitting” of these groups. Other issues remain unresolved with use of the current system, including how to incorporate specific radiographic features such as the density of pulmonary nodules, or how to properly prognosticate in clinical situations that fall out of the staging paradigm, for example, the presence of multiple synchronous primary tumors, but the most important limitation is the absence of components reflecting tumor biology. As current understanding of molecular pathways important to neoplasia expands, it seems inevitable that the staging system must incorporate such information in the future.

Table 66-5 Definitions for Tumor-Node-Metastasis Descriptors

Table 66-6 Lung Cancer Stage Groupings by Tumor-Node-Metastasis (TNM) Descriptors *

Stage	TNM Grouping
0	TisN0M0

IIA

IIB

IIIA

IIIB

* As defined in the seventh edition of the American Joint Commission on Cancer and the Union Internationale Contre le Cancer classification.

From Edge SB, editor: The American Joint Commission on Cancer cancer staging manual, ed 7. Chicago, Springer, 2010.

Ultimately, even with its limitations, the current staging classification represents a major advance in lung cancer care, providing a reliable language to facilitate consistent description and communication, and to give patients, their families, and their physicians insight regarding prognosis. The tools and strategies applied most commonly in the process of clinical evaluation are discussed in the following sections within the context of their use in defining the separate clinical stages.

Stage I

In the IASLC database, 28% of patients with NSCLC presented with stage I disease. Patients with stage IA lung cancer (T1aN0M0, T1bN0M0) are those with malignant solitary pulmonary nodules 3 cm or less in diameter. These patients typically are asymptomatic. In general, the smaller the SPN, the less likely spread will have occurred beyond the primary site. CT scan should be performed in all patients with known or suspected lung cancer, with inclusion of the upper abdomen for evaluation of the liver and adrenal glands, because these are common sites of metastasis. By definition, CT scanning in stage I disease will not demonstrate enlarged hilar or mediastinal nodes.

In a systematic review of the medical literature performed to inform the American College of Chest Physicians evidence-based guidelines for the diagnosis and management of lung cancer (5111 total evaluable patients with a median prevalence of mediastinal metastasis of 28%), the pooled sensitivity and specificity of CT scanning for identifying mediastinal lymph node metastasis were 51% and 86%, respectively. In the specific clinical situation combining an SPN 3 cm or less and a normal-appearing mediastinum on chest CT, the likelihood of finding malignant mediastinal lymph node involvement by surgical lymph node sampling is low, approximately 6%. Whether this small percentage justifies further noninvasive staging beyond CT scanning is controversial. Overall, PET gives false-negative results in approximately 20% of patients with lung cancer associated with normal-sized but malignant mediastinal nodes. The decision of whether or not to pursue PET imaging must take this into consideration, as well as the possibility that identifying a false-positive abnormality may trigger unnecessary evaluation. Similarly, brain magnetic resonance imaging (MRI) and CT scanning are of low yield in general in patients with lung cancer in whom findings on clinical examination are normal (0 to 10%), and particularly so in patients with clinical stage I NSCLC. Conversely, patients with symptoms suggesting distant disease should undergo further noninvasive imaging, even though the primary tumor may be small.

Patients with stage IB lung cancer have larger solitary tumors (ranging from less than 3 cm to more than 5 cm in diameter). In such cases, the primary tumor may be surrounded by normal lung but also may invade the visceral pleura, involve the main bronchus 2 cm or more distal to the carina, or be associated with atelectasis or obstructive pneumonia not involving the entire lung. The likelihood of finding regional nodal or distant metastasis in the setting of a larger primary lesion is higher than in stage IA, so these patients should undergo PET scanning and brain imaging (MRI or CT with contrast) as part of their staging evaluation (Figure 66-4). With the recognition that false-positive results on PET are frequent with nonmalignant inflammation or infection, tissue confirmation should be obtained whenever possible to establish the presence or absence of regional or distant metastasis, because patients should not be denied potentially curative treatments on the basis of imaging alone. Conversely, controversy continues regarding whether a negative result on PET scan in a patient with a T2 or T3 solitary pulmonary lesion and normal mediastinal nodes on CT scan should be followed by tissue biopsy to prove that the mediastinum is truly free of disease. In a metaanalysis evaluating this question, the probability of finding malignant lymph node involvement when both CT and PET findings were negative was approximately 9% (of note, the pretest probability of mediastinal lymph node metastasis was estimated at 35%, and patients were not differentiated by the size of the primary tumor). Clinicians ultimately must decide whether the overall 9% probability of mediastinal lymph node metastasis in patients with clinical stage IB disease and negative results on CT and PET imaging of the mediastinum warrants further invasive evaluation. This decision may be influenced by the recognition that certain subgroups, including patients with central or larger tumors, have an increased likelihood of mediastinal nodal involvement.

Figure 66-4 The patient was a 69-year-old man with a 45 pack-year smoking history who presented with cough. A, Chest computed tomography (CT) scan demonstrates a 4.8-cm right upper lobe mass. The hila and mediastinum appeared normal. B, Fused transaxial positron emission tomography (PET) image at the same level as in A demonstrates intense fluorodeoxyglucose (FDG) uptake in the right upper lobe mass. C, Coronal PET image demonstrates intense FDG uptake in the right upper lobe mass as well as in the right hilum and the mediastinum. Brain magnetic resonance imaging (MRI) was negative. Mediastinoscopy demonstrated adenocarcinoma. Clinical stage was T2aN2M0, stage IIIa.

Stage II

Stage II encompasses a variety of TNM groupings, including those comprising patients who have large lung masses that may invade local structures, patients with small primary tumors and hilar nodal involvement, and patients whose tumors have an endobronchial component within 2 cm of but not involving the main carina. The range of T descriptors in stage II emphasizes that the stage designations are defined by survival outcome, not by anatomy alone. Clinical staging for patients with stage II disease parallels that for stage IB, because the evaluative process typically must include a decision about whether to pursue an invasive mediastinal examination in the setting of negative results of mediastinal imaging on chest CT. Patients with stage II disease are more likely to have occult mediastinal lymph node (N2) involvement, so PET scanning should be routinely included in their staging assessment. The hilar (N1) nodes are difficult to biopsy, except perhaps by a bronchoscopic approach under endobronchial ultrasound (EBUS) guidance. If the ipsilateral N1 nodes are enlarged on CT scan or demonstrate FDG uptake on PET imaging, it usually is not necessary to sample them before surgery, because they will be removed at the time of resection. However, N1 involvement increases the likelihood that N2 nodes also are involved, even if CT and PET results are negative, and in such cases, invasive evaluation of the mediastinum should be performed to evaluate for the possibility of higher-stage disease. Because by definition the mediastinal nodes in this case are not enlarged, a focused evaluation will not be possible. Accordingly, the most complete mediastinal evaluation possible should be pursued, which may vary institutionally, with preferential use of mediastinoscopy, EBUS, or EBUS with endoscopic ultrasound (EUS).

MRI is not frequently indicated in the evaluation of patients with lung cancer, with the notable exception of patients in whom a more accurate evaluation relating to the bones of the chest cage, the soft tissues and vasculature, or the brachial plexus (in the setting of a superior sulcus [Pancoast] tumor) is necessary. Patients with stage II lung cancer defined by large primary tumors invading adjacent structures often fall into these categories. In these challenging clinical situations, MRI may be superior to CT scanning in distinguishing neoplastic from normal tissues.

Stage III

Approximately one third of patients with lung cancer present with clinical stage III disease. All patients with evidence of stage III disease on chest CT scan should be evaluated for the presence of distant disease; this may be accomplished by a combination of (1) brain MRI or head CT and (2) PET imaging or abdominal CT scanning with bone scan. The presence of metastatic disease immediately changes the clinical stage to IV. As with stage II, stage III comprises a wide variety of TNM groupings (see Table 66-6). However, a majority and perhaps the most clinically challenging of these patients with such disease will be defined by involvement of the N2 nodes; careful evaluation of the mediastinum is thus vitally important. The IASLC lymph node map defining the lymph node stations is shown in Figure 66-5.

Figure 66-5 The International Association for the Study of Lung Cancer lymph node map.

(From Rusch VW, Asamura H, Watanabe H, et al: The IASLC Lung Cancer Staging Project: a proposal for a new international lymph node map in the forthcoming 7th edition of the TNM classification for lung cancer, J Thorac Oncol 4:568–577, 2009.)

Mediastinal Evaluation: Noninvasive Modalities

Chest CT remains the most accurate anatomic modality for evaluation of the mediastinum. Pathologic enlargement of the mediastinal nodes is defined as a short-axis diameter of 1 cm or greater. CT in the setting of lung cancer is a relatively inaccurate tool for identification of malignant involvement in the mediastinum. Many studies evaluating CT accuracy in this setting have demonstrated that approximately 20% of mediastinal nodes that are not enlarged are actually malignant; conversely, approximately 40% of enlarged mediastinal nodes are actually benign. For all patients with enlarged discrete mediastinal lymph nodes on CT, further evaluation of the mediastinum should be performed before a definitive treatment decision is reached, in view of the high false-positive rate. The exception to this approach would be in patients with bulky mediastinal disease, with lymph node infiltration around vessels and airways so extensive that discrete lymph node measurements are meaningless (Figure 66-6). In such cases, malignant involvement of the nodes would be assumed, and the establishment of cell type could be pursued by sampling the most accessible site, either the mediastinal nodes or the primary tumor.

Figure 66-6 The patient was a 62-year-old woman with a 30 pack-year smoking history who presented with cough. Chest computed tomography (CT) reveals a right upper lobe mass with extension of tumor into the mediastinum and precarinal mediastinal adenopathy. Bronchoscopy with endobronchial ultrasound imaging demonstrated endobronchial invasion of tumor at the level of the main carina; biopsy specimens confirmed adenocarcinoma. There was no evidence of distant metastasis. Clinical stage was T4N2M0, stage IIIb.

PET scanning is recommended for further evaluation of all patients with clinical stage III lung cancer because of the high likelihood of distant disease. PET scanning is now widely available, with utilization of this modality markedly increasing over the past decade. Abdominal CT with bone scan is a reasonable substitute if PET is unavailable. PET is superior to CT with respect to evaluation of the mediastinum. In a large systematic review including 2865 evaluable patients with lung cancer, the pooled sensitivity and specificity of PET for identifying mediastinal lymph node metastasis were 74% and 85%, respectively. In a separate metaanalysis of patients with enlarged mediastinal lymph nodes, the median sensitivity and specificity of PET were reported at 100% and 78%, respectively. Thus, PET is superior to CT in identifying malignant mediastinal lymph nodes, but approximately 20% of nodes identified as abnormal by PET imaging are actually benign, reflecting the fact that tissues other than cancer can exhibit increased cellular glucose uptake. Because malignant N2 nodal involvement may be the critical factor determining whether a patient is surgically resectable, it is important that tissue confirmation be obtained whenever possible before making a final determination of clinical stage.

Mediastinal Evaluation: Invasive Modalities

A variety of minimally invasive modalities are available for the evaluation of the mediastinal lymph nodes (Table 66-7). TTNA, EUS, extended cervical mediastinoscopy, and EBUS typically are utilized for the purpose of directed biopsy in patients with discrete abnormalities on CT or PET scanning. The choice between these procedures will be decided by the anatomic location of the abnormal nodes and by institutional expertise. The sensitivities of all of the directed minimally invasive techniques are very high, in part because they typically are done in situations in which the prevalence of malignancy in the abnormal nodes also is very high. In theory, all mediastinal lymph nodes are amenable to TTNA, although the individual anatomy will determine whether biopsy is feasible. Ultrasound-guided biopsy techniques using endoscopy or bronchoscopy are increasingly available and have high diagnostic yield.

Table 66-7 Mediastinal Lymph Node Stations: Accessibility to Minimally Invasive Modalities

Procedure	Accessible Mediastinal Lymph Node Stations
Mediastinoscopy	Stations 2, 3, 4, 7 (anterior only)
Extended cervical mediastinoscopy	Station 6
Video-assisted thorascopic surgery (VATS)	Ipsilateral stations 2, 3, 4, 7; station 6 with left-sided procedure
Parasternal mediastinotomy (Chamberlain procedure)	Station 5 (aortopulmonary window)
Endobronchial ultrasound (EBUS)	Stations 2, 3, 4, 7
Endoscopic ultrasound (EUS)	Stations 4, 5, 7, 8, 9
Transthoracic needle aspiration (TTNA)	All stations