The clinical symptoms of NSCLC are varied. Because of the highly aggressive nature of these cancers, two thirds of patients exhibit symptoms resulting from metastatic or systemic disease at the time of initial presentation. The presence of symptoms is usually related to either locally advanced or metastatic disease. Perhaps with lung cancer screening strategies, more asymptomatic patients with lung cancer will be detected in the future. Clinical symptoms can be categorized simply as pulmonary, extrapulmonary thoracic, and extrathoracic symptoms.
Pulmonary symptoms include cough, dyspnea, wheeze, stridor, hemoptysis, and pneumonic symptoms.
Although cough is the most frequent presenting symptom of NSCLC, occurring in 75% of patients, the finding is nonspecific and is frequently present in cigarette smokers. The finding of a new, unremitting cough, or change in cough, especially in patients older than 50 years of age who have a significant tobacco history, warrants further investigation.
Dyspnea is present in 50% to 60% of patients with NSCLC.The etiology of dyspnea is diverse, including obstructive and compressive factors. Central airway obstruction may result from endobronchial disease, usually related to squamous cell carcinoma. Extrinsic bronchial compression by a large central tumor or malignant mediastinal adenopathy can cause dyspnea, which is common with patients with small cell lung cancer (SCLC). In addition, parenchymal compression may be caused by a significant malignant pleural effusion. Dyspnea may be caused by large pericardial effusions or by vena caval obstruction. Rarely, a large or infiltrative tumor burden limits alveolar function (e.g., diffuse lepidic adenocarcinoma or lymphangitic carcinomatosis).
As with dyspnea, obstruction or compression of the bronchus by the tumor itself or by enlarged lymph nodes can cause wheezing if peripheral bronchi are affected or stridor if the trachea or main bronchi are involved.
Hemoptysis is usually not massive in NSCLC and may occur in 25% to 40% of patients. This presentation occurs more commonly with centrally located tumors such as squamous cell tumors and SCLC.
Almost all of the pulmonary symptoms of NSCLC can be mimicked by a postobstructive pneumonia. In addition, postobstructive atelectasis and pneumonia can cause fever, chills, sputum production, and pleuritic chest pain. Finally, lung abscess may be the end result of a chronic postobstruction pneumonia. In a patient with nonresolving pneumonia, bronchoscopy should be strongly considered to rule out endobronchial obstruction.
Direct invasion of chest wall or mediastinal structures by either the tumor or enlarged lymph nodes may lead to diagnostic symptoms.
Chest Wall Pain
Peripheral tumors may extend through the visceral pleura to invade the parietal pleura, intercostal muscles or nerves, or ribs. Pain may result from invasion of any of these structures. Visceral pleural invasion leads to pleuritic pain, and invasion of the chest wall structures leads to somatic gnawing pain. Invasion of a neurovascular bundle may lead to radicular pain. A special case of chest wall invasion is the superior sulcus (Pancoast) tumor, which involves the thoracic outlet. This invasion can cause any of the classic triad of symptoms, including shoulder pain from direct muscle or rib invasion, radicular arm pain from invasion of C8 and T1 nerve roots, and Horner’s syndrome (ipsilateral ptosis, miosis, and facial anhidrosis) from invasion of the stellate sympathetic ganglion.
Symptoms From Mediastinal Involvement
Mediastinal involvement of lung cancer causes specific symptoms based on the structure involved. Direct invasion of the phrenic nerve can cause either chronic hiccups or frank diaphragmatic paralysis. Also, because of its origin from C3-C5, phrenic nerve involvement can cause referred shoulder pain. Involvement of the recurrent laryngeal nerve occurs most commonly on the left due to the proximity of the nerve to lymph nodes in the aortopulmonary window. This may lead to subtle voice alteration or to hoarseness. Chronic recurrent laryngeal nerve paralysis contributes to lung dysfunction because of recurrent aspiration secondary to inability to adequately protect the airway. Extensive tumor involvement of right mediastinal lymph nodes may result in superior vena caval syndrome, which is characterized by a plethoric appearance, distention of the venous drainage of the arm and neck, and edema of the face, neck, and arms. Vena caval obstruction usually progresses gradually over time, allowing the development of collateral venous drainage that is detectable on physical examination with dilated subcutaneous veins. SCLC is more often the cause, rather than NSCLC. Pericardial involvement may lead to symptomatic effusion and even tamponade. Esophageal compression with dysphagia sometimes occurs because of compression by enlarged subcarinal mediastinal lymph nodes. Finally, vertebral bodies may be involved by posterior tumors, leading to back pain.
Paraneoplastic syndromes—symptoms or findings that are related to the primary tumor or its metastases by hormonal intermediates—may accompany lung cancer. Although paraneoplastic syndromes are unusual, the resulting symptoms may predate thoracic manifestations of a curable primary tumor. Systemic manifestations of NSCLC include cachexia, parathyroid-like hormone secretion with concomitant hypercalcemia, hypertrophic pulmonary osteoarthropathy, and various neurologic syndromes. Weight loss and anorexia occur in up to one third of patients. Even in noncachectic patients, however, increases in protein turnover, glucose production, and muscular catabolism may be demonstrated.
Paraneoplastic syndromes are frequently associated with SCLC, and they are more often present at the time of diagnosis in patients with SCLC than in those with NSCLC. In addition to weight loss, anorexia, and neuromyopathies, paraneoplastic syndromes may result from tumor elaboration of antidiuretic hormone, adrenocorticotropin (ACTH), calcitonin, or parathyroid hormone.
Skeletal syndromes include hypertrophic pulmonary osteoarthropathy (HPO) and clubbing. HPO is a proliferation of periosteum of the ends of long bones. Affecting primarily the tibia, fibula, and radius, the periostitis causes tenderness and swelling. Unlike many of the other paraneoplastic syndromes, HPO is more common in NSCLC than in SCLC. Alkaline phosphatase levels are often elevated, but hepatic enzymes are normal. HPO is always associated with clubbing of the digits, but the converse is untrue. Clubbing occurs in 35% of NSCLC patients .
The syndrome of inappropriate antidiuretic hormone secretion (SIADH) occurs in up to 46% of patients with SCLC,but also may occur, less frequently, in NSCLC. Symptoms include those of hyponatremia: anorexia, nausea, vomiting, confusion, lethargy, and seizures. These same signs and symptoms may be present with secretion of atrial natriuretic peptide (ANP), which may also be secreted by SCLC. The distinction between the two conditions may be made by measurement of serum ADH. In SCLC, the treatment is chemotherapy.
Hypercalcemia develops in 10% of patients with lung cancer. However, only 15% of these cases are caused by the production of parathyroid hormone or other humoral substances. For this reason, metastatic bone disease must be ruled out in these patients by bone scan or by positron emission tomography (PET). The most frequent lung tumor that produces parathyroid hormone is squamous cell cancer. These tumors are often resectable, and, after complete resection, the calcium level normalizes. However, tumor recurrence after resection is common and may manifest as recurrent hypercalcemia.
Ectopic production of an ACTH-like substance may lead to Cushing’s syndrome. This is much more common in SCLC than in NSCLC. The resulting cortisol production is not suppressible by dexamethasone. Because of the rapidity of the ACTH elevation, physical signs of Cushing’s syndrome are usually absent, and the symptoms that do appear are primarily caused by metabolic consequences. Among these are hypokalemia, metabolic alkalosis, and hyperglycemia.
Neurologic paraneoplastic syndromes are most commonly associated with SCLC and squamous cell cancer and are thought to be immune mediated. The cancer cells may express antigens that are normally expressed only by nervous system tissues. Unlike patients with the endocrine paraneoplastic syndrome, these patients tend to exhibit their symptoms later in the disease process. Symptoms resulting from this immune process range from sensory, sensorimotor, and autonomic peripheral neuropathies to the central neuropathies of cerebellar degeneration, dementia, brain stem encephalitis, and encephalomyelitis. Weight loss often accompanies these symptoms, making workup of metastatic disease essential. The peripheral neuropathies are the most common paraneoplastic syndromes associated with lung cancer, occurring in up to 16% of patients. Of these, 56% have SCLC, 22% have squamous cell cancers, 16% have large cell cancers, and 5% have adenocarcinomas.
Lambert-Eaton myasthenic syndrome is most frequently seen in SCLC and may lead to proximal muscle weakness and fatigability (particularly of the thighs), a waddling gait, and dry oral mucosa. This syndrome is produced by immunoglobulin G (IgG) antibodies, which target voltage-gated calcium channels that function in the release of acetylcholine from presynaptic sites at the motor endplate. These antibodies are generated through an immune response to similar channels present on tumor cells. These symptoms often occur before the onset of symptoms of the primary tumor and may precede radiologic evidence of the tumor by up to 4 years. As with most of the paraneoplastic syndromes, treatment of the primary tumor may lead to dramatic symptomatic improvement.
Lung cancer most frequently metastasizes to the brain, bones, liver, adrenal glands, lungs, and skin/soft tissues.
Central Nervous System Metastases
Central nervous system (CNS) metastases are present in approximately 10% of patients at diagnosis. Over time, 10% to 15% of the other patients will develop CNS lesions. Although these metastases are frequently asymptomatic, symptoms of increased intracranial pressure (headache, nausea, vomiting, altered level of consciousness) predominate. More focal symptoms, such as weakness/numbness or seizures, are less common.
Twenty-five percent of lung cancer patients develop bony metastases. Of these, 55% occur in the axial skeleton (spine, pelvis, sternum, ribs). These patients complain of pain at the affected area.
Hepatic and Adrenal Metastases
Hepatic metastases are usually asymptomatic and are found on computed tomographic (CT) or PET scans. Adrenal metastases are often found on a staging CT scan or PET scan and are also usually asymptomatic. Addison’s or Conn’s disease rarely is present, even in the presence of bilateral metastases.
Skin and Soft Tissue Metastases
Skin and soft tissue metastases are present as late-stage findings in 8% of lung cancer patients. These lesions are usually subcutaneous and painless. Occasionally, they erode through the skin and cause a chronic wound, which may require excision.
Anorexia, weight loss, fatigue, and malaise are poorly understood symptoms of metastatic lung cancer. All patients with lung cancer need to be queried regarding such symptoms, and their presence initiates a search for metastatic disease.
This subject is covered in more detail elsewhere; however, a brief review is pertinent to this discussion of diagnosis. The purpose of screening is to identify lung cancers at an early stage, before signs and symptoms develop. Achieving this goal would increase the number of patients who are eligible for curative surgical resection, improving overall survival and resource utilization.
In general most lung cancers are discovered when symptomatic and either locally advanced or metastatic. To reduce the mortality of lung cancer, it is apparent that the disease must be identified at an earlier stage when it is more likely to be curable. This is the premise of lung cancer screening. In the past chest x-ray (CXR) was used as a screening test but several studies failed to demonstrate a benefit of screening with CXR. With the development of CT scan, smaller lung nodules could be identified and early studies proved that low dose CT could identify early curable lung cancers.
In order for CT-based screening programs to improve on programs based on plain chest roentgenography (CXR), a significant relationship must be identified between smaller size of a tumor and lower stage and subsequently better survival. Analysis of the Surveillance, Epidemiology and End Results (SEER) registry focusing on 84,152 cases of lung cancer diagnosed before autopsy revealed a significant correlation between small size and low stage. Although this finding is encouraging, no assessment of symptoms was available, limiting its applicability to screening.
The Early Lung Cancer Action Project (ELCAP) study screened 1000 smokers (≥10 pack-years) who were asymptomatic and deemed fit for thoracotomy, with CXR and helical CT. Noncalcified nodules were identified in 233 patients by helical CT and in 68 patients by CXR. These 233 subjects then underwent conventional CT, and 28 were biopsied, with 27 cancers diagnosed. Of these, 23 were stage I and 26 (96.3%) were resectable. The incidence in this study was 31 per 1000 subjects. The Mayo Clinic screened 1520 smokers (>20 pack-years) older than 50 years of age with annual spiral CT for 5 years.This trial found noncalcified nodules in 74% of the patients. Of these, 66 were found to be lung cancer.
These two studies demonstrate the CT can identify early stage curable lung cancer, but also show that many people have lung nodules that are not malignant. A management algorithm for small asymptomatic nodules is essential to prevent unnecessary testing and surgery for such individuals. For lung cancer screening the LungRADS algorithm ( discussed in the chapter on screening) should be applied, whereas for incidentally identified nodules the Fleischner Society guidelines should be used.
Overall the approach to the solitary pulmonary nodule is the similar whether the nodule was discovered as an incidental finding on a CT scan performed for other reasons or whether it was discovered in the context of a lung cancer screening program. With the increasing use of CT scans many more incidental nodules will be identified. A review of CTs performed between 2006 and 2012 reported 1.5 million nodules in 4.8 million people ( ie 31%) of CTs will have at least one nodule. Most such incidental nodules are not cancer, whereas the likelihood of malignancy increases in nodules detected through a screening program which by definition includes only those with increased risk of lung cancer.
A solitary pulmonary nodule is defined as one that is less than 3 cm in diameter and is surrounded by pulmonary parenchyma. The differential diagnosis is broad (table 1); in follow-up, historically, these are found to be malignant up to 70% of the time. This proportion will undoubtedly change with increased use of CT and detection of smaller nodules. Of the benign lesions, 80% are infectious granulomas, and 10% are hamartomas.
In evaluating a patient with a pulmonary nodule many factors should be considered, recognizing that lung cancer is perhaps the most important and most common diagnosis. Factors suggestive of benign lesions include symptoms suggestive of respiratory tract infection, exposure to tuberculosis, rheumatoid arthritis or Granulomatosis with Polyangitis (GPA). Risk factors for lung cancer include age (lung cancer is rare before age 35), family history of lung cancer, race ( black men, Hawaiian men), cigarette smoking, second hand smoke exposure, exposure to asbestos, radon, and uranium as well as chronic lung disease such as emphysema and pulmonary fibrosis. Young age does not preclude lung cancer nor does an absence of smoking history. Male sex was previously considered to be a risk factor but more recently female sex was identified as a risk factor in the PanCan Trial .
In determining management of incidentally detected nodes, the Fleischner Society updated their guidelines in 2017. This update was driven by the recognition that with more frequent use of CT scans, many individuals will be found to have one or more incidental pulmonary nodules. Importantly the miniumum threshold size for followup was increased, the number of followup scans for stable solid nodules was decreased and the duration of followup required for subsolid nodules was increased to 5 years. See Table 1.
A. Solid Nodules*
< 6 mm (< 100 mm³)
6-8 mm (100-250 mm³)
>8 mm (>250 mm³)
No routine follow-up
CTat 6-12 months, then consider CT at 18-24 months
Consider CT at 3 months, PET/CT, or tissue sampling
Nodules < 6 mm do not require routine follow-up in low-risk patients (recommendation 1A).
Optional CT at 12 months
CT at 6-12 months, then CT at 18-24 months
Consider CT at 3 months, PET/CT, or tissue sampling
Certain patients at high risk with suspicious nodule morphology, upper lobe location, or both may warrant 12-month follow-up (recommendation 1A).
No routine follow-up
CT at 3-6 months, then consider CTat 18-24 months
CT at 3-6 months, then consider CT at 18-24 months
Use most suspicious nodule as guide to management. Follow-up intervals may vary according to size and risk (recommendation 2A).
Optional CT at 12 months
CT at 3-6 months, then at 18-24 months
CT at 3-6 months, then at 18-24 months
Use most suspicious nodule as guide to management. Follow-up intervals may vary according to size and risk (recommendation 2A).
B: Subsolid Nodules*
< 6 mm (< 100 mm³)
≥6 mm (>100 mm³)
No routine follow-up
CT at 6-12 months to confirm persistence, then CT every 2 years until 5 years
In certain suspicious nodules < 6 mm, consider follow-up at 2 and 4 years. If solid component(s) or growth develops, consider resection. (Recommendations 3A and 4A).
No routine follow-up
CT at 3-6 months to confirm persistence. If unchanged and solid component remains < 6 mm, annual CT should be performed for 5 years.
In practice, part-solid nodules cannot be defined as such until ≥6 mm, and nodules < 6 mm do not usually require follow-up. Persistent part-solid nodules with soSd components ≥6 mm should be considered highly suspicious (recommendations 4A-4C0.
CT at 3-6 months. If stable, consider CT at 2 and 4 years.
CT at 3-6 months. Subsequent management based on the most suspicious nodule(s).
Multiple < 6 mm pure ground-glass nodules are usually benign, but consider follow-up in selected patients at high risk at 2 and 4 years (recommendation 5A).
These do not apply to lung cancer screening, patients with immunosuppression or patients with no primary cancer.
Carcinoid, atypical carcinoid
Other rare primary pulmonary malignancies
Granuloma ( including TB)
Wegener’s granulomatosis nodule (GPA)
External object (nipple, mole)
Pseudotumor (fluid in fissure)
NSCLC, non–small cell lung cancer; SCLC, small cell lung cancer.
Characteristics of nodules on chest CT may help disclose malignancy. (See table 2). First among these is the border of the lesion. Margins of a lesion may be classified as smooth, lobulated, or irregular (spiculated). Smooth nodules are generally benign, although 21% of malignant nodules have smooth borders.A lobulated contour may signify the uneven growth of a malignancy, but it may also occur in up to 25% of benign nodules. Spiculated nodules have a corona radiata appearance and are highly likely to be malignant.,,,,
Internal characteristics and density of a nodule may also help to diagnose malignancy. Homogeneous attenuation is seen in 55% of benign lesions but also in 20% of malignant lesions. Cavitation may occur in both benign and malignant nodules. However, benign nodules tend to have thin walls, whereas malignant nodules tend to have thick and irregular walls. Most nodules with a wall thickness greater than 16 mm are malignant, and most of those with wall thickness less than 4 mm are benign. Certain characteristics may also help with specific diagnoses: pseudocavitation for lepidic type adenocarcinoma, and intranodular fat for hamartoma.Finally, calcification helps to diagnose benign nodules. Central, diffuse solid, and laminated calcifications are seen in loci of prior infections, whereas so-called popcorn calcification is seen in hamartoma. However, up to 63% of benign nodules are not calcified, and calcification may also occur in up to 6% of lung cancers. In malignant nodules, calcification is often eccentric.
In like manner, the solidity of a nodule may be important. A nonsolid nodule is defined as a density through which aerated lung is visible on the CT lung windows . This type of nodule is called a pure ground glass nodule (GGN) and has a low risk of malignancy ( < 6mm: < 1%); however, as size increases to greater than 15 mm, the risk of malignancy increases. A part-solid nodule contains a solid portion that obliterates aerated lung in addition to a nonsolid portion, and these nodules carry a much higher risk of malignancy (40%-50% even for nodules smaller than 15 mm). Again, the risk of cancer increases with the size of the solid component. Of small solid nodules, only approximately 15% are malignant. A progression from GGN to part-solid to solid nodules may occur and this change represents the development of an invasive malignancy.
Assessment of growth rate may also allow prediction of malignancy. Volume-doubling time for malignant nodules is between 30 and 400 days and correlates with a 26% increase in diameter. Because of this, stability of a nodule’s size over 2 years has generally been considered to be reliable for benignity. This rule is easier to apply and probably more accurate for larger lesions rather than small ones. With the increasing frequency of adenocarcinoma spectrum cancers, 2 years of followup may not be adequate as these tumors often grow slowly such that up to 5 year followup is now recommended. Conversely, rapidly growing nodules are often inflammatory.
Bayesian analysis allows precise determination of the probability of malignancy by calculating a likelihood ratio (the number of malignant nodules with a given feature divided by the number of benign nodules with the same feature) for each characteristic and then taking the product of these ratios., The probability of malignancy based on various clinical and radiologic characteristics is shown in Table 3.
Size >3 cm
Age < 70 years
Malignant growth rate
Upper lobe location
Size < 1 cm
Age 30-39 years
Age 20-29 years
Benign growth rate
ELCAP data provide some interesting perspectives on this problem. The circumstance of the CT scan is important. In the ELCAP series of 1000 screening CTs, 23.3% of scans showed one to six noncalcified nodules; of these, 12% contained cancer. However, on repeat screening 1 year later, only 2.5% of scans contained new nodules. Twelve of these nodules resolved 4 to 6 weeks later; of the 18 nodules that did not resolve, 39% had lung cancer.These data indicate that, although new nodules are more common on screening studies, false-positive results are much more common on baseline screening than on 1-year interval screening studies.
Although degree of smoking is clearly important in estimating lung cancer risk in a solitary nodule, some subtleties have emerged. Whereas cessation of smoking clearly limits lung cancer risk, the risk may level off rather than decline.The trend toward lower-tar, filtered cigarettes has corresponded with a shift from central lesions to peripheral adenocarcinomas as the predominant cancers found. CT may be the test of choice to identify these peripheral nodules.
Diagnosis and staging of lung cancer are often performed concurrently. Therefore, before discussion of the various diagnostic modalities, a description of the staging system is necessary. The purposes of any staging classification are the assessment of prognosis and assignment of therapy. As with other cancers, the staging system for lung cancer is based on a tumor-node-metastasis (TNM) system (Tables 4 and 5). Before operation, patients are assigned a clinical stage; after surgery, a pathologic stage is obtained. Because the pathologic stage allows more careful assessment of the primary tumor and nodes, survival rates based on pathologic stage are higher than those based on clinical stage.
In 1974, the American Joint Committee for Cancer (AJCC) introduced a lung cancer staging system based on the TNM system. Shortly thereafter, Naruke and colleaguesput forth an anatomic lymph node map. This schema was subsequently modified by the American Thoracic Societyand by Mountain and Dresler (Mountain and Dresler, 1997).The current regional lymph node map was developed by the IASLC is presented in Figure 1. The TNM staging system has been updated several times over the years based on large international datasets. The 8th edition of the AJCC/UICC staging manual descriptors are shown in table 3 and 4.
Primary Tumor (T)
Primary tumor cannot be assessed or visualized by imaging or bronchoscopy but proven by malignant cells in sputum or bronchial washings
No evidence of primary tumor
Carcinoma in situ
Minimally invasive adenocarcinoma
Tumor ≤ 1 cm in greatest dimension, surrounded by lung or visceral pleura, without bronchoscopic evidence of invasion more proximal than the lobar bronchus
Tumor > 1-2 cm
Tumor > 2-3 cm
Tumor >3-4 cm in greatest dimension;
or involving the main stem bronchus >2 cm from the carina 3-4 cm; or invading the visceral pleura; or associated with atelectasis or obstructive pneumonitis that extends to the hilar region but does not involve the entire lung
Tumor >4-5 cm
or involving main stem bronchus >2 cm from carina >4-5 cm
Tumour >5-7 cm
Tumor of any size that invades the chest wall, mediastinal pleura, or parietal pericardium; or tumor in the main stem bronchus < 2 cm from the carina without invading the carina; or associated atelectasis or obstructive pneumonitis of the entire lung
Tumor >7 cm
Tumor of any size that invades diaphragm, mediastinum, heart, great vessels, trachea, esophagus, vertebral body, or carina; or tumor with a malignant pleural or pericardial effusion; or tumor with satellite nodule(s) within the same lobe
Regional Lymph Nodes (N)
Regional nodes cannot be assessed
No regional lymph node metastasis
Metastasis to ipsilateral peribronchial or ipsilateral hilar lymph nodes and intrapulmonary nodes, including involved by direct extension of the primary tumor
Metastasis to ipsilateral mediastinal or subcarinal lymph nodes
Metastasis to contralateral mediastinal, contralateral hilar, or ipsilateral or contralateral scalene/supraclavicular nodes
Distant Metastases (M)
No distant metastasis detected
Separate tumor nodule in a contralateral lobe
Tumor wiht pleural or pericardial nodules or malignant pericardial or pleural effusion
Single lesion ( includes involvement of a single distant ( nonregional) lymph node
5-Year Survival (%)
TX N0 M0
Tis N0 M0
T1a( mi) N0 M0
T1b N0 M0
T1c N0 M0
T2a N0 M0
T2b N0 M0
T1a,b,c N1 M0
T2a,b N1 M0
T3 N0 M0
T3 N1 M0
T1-2 N2 M0
T4 N0-1 M0
T1-2 N3 M0
T3-4 N2 M0
T3-4 N3 M0
T N M1a,b
T N M1c
Staging is usually carried out in a centripetal manner. Distant disease is ruled out first, so that patients who are not surgical candidates may be spared invasive procedures. Next, noninvasive methods to stage mediastinal nodes are performed before invasive tests are undertaken.
The CXR has been an integral tool since its inception and is believed by many to be the most important test in the diagnosis of lung cancer. If lung cancer is suspected based on clinical criteria, a posterior-anterior and lateral CXR has been the first test ordered (Figure 2). Lung masses are also often seen on CXR performed for other reasons. For a lesion to be visible on CXR, it must be at least 7 to 10 mm in diameter. The posterior-anterior and lateral CXR also allows assessment of peripheral versus central location of the tumor, presence of atelectasis due to bronchial obstruction, presence of effusion from malignant spread or from exudate, and presence of hemidiaphragm elevation due to phrenic nerve involvement. On occasion, a large degree of hilar lymphadenopathy is visible on the CXR. If a lesion is seen on CXR, review of a previous CXR should be undertaken if available to determine if the lesion is new or pre-existing.
If an abnormal nodule is suspected on CXR, the next step in diagnosis is usually chest CT. Ideally the CT scan should be performed with intravenous contrast. The CT scan confirms the presence, size, and location of the primary nodule, and it may also provide important information regarding the risk of malignancy of the primary tumor (Figure 3). In addition, the CT scan assesses the contiguous structures, hilar lymph nodes and mediastinal lymph nodes, the status of the remaining pulmonary parenchyma and pleura, and the presence of distant metastases (particularly in the liver and adrenal glands). As noted in the discussion of the solitary lung nodule, the margin of the lesion, the internal characteristics of the lesion, the rate of growth, the circumstance of the CT, the presence or absence of calcification, and the smoking history of the patient are all extremely important in assessing the risk of cancer in a lung nodule.
Assessment of mediastinal lymph nodes is also aided by chest CT. Size greater than 1 centimeter in the shorter axis is the usual criterion for defining adenopathy. In a 143-patient prospective study in which CT findings were tested by surgical pathology, the sensitivity of chest CT for mediastinal lymph node positivity was found to be 64%, and the specificity was 62%. Meta-analysis of 3438 patients in 20 studies revealed a lower sensitivity of 57% but a higher specificity of 82% ., Clearly, Lymph nodes that are suspicious in the mediastinum by chest CT require histologic confirmation.
Chest CT may also diagnose non-nodal metastatic disease. It provides detailed information regarding synchronous second nodules as well as satellite nodules. In addition, because routine chest CT includes the upper abdomen to the level of the adrenal glands, unsuspected metastases may be documented in the liver and in the adrenal glands. Rates of positive findings are 3% to 6% in the liver and 3% to 7% in the adrenals. These patients are then spared noncurative surgery.
Positron Emission Tomography
PET uses fluorodeoxyglucose (FDG), a D-glucose analogue that is labeled with positron-emitting fluorine 18. This agent is taken up by cancer cells and phosphorylated but is not metabolized further. Because malignant cells are more metabolically active than normal cells, the F-labeled agent preferentially accumulates in these cells. Whole-body imaging allows visualization of malignant cells at primary, nodal, and distant sites (Figure 4).
For diagnosis of cancer in a solitary pulmonary nodule, FDG-PET is accurate. Multiple studies have revealed a sensitivity of 96%, specificity of 88%, and accuracy of 94% in these nodules.,,,, False-positive results may occur with infectious and inflammatory processes. False-negative results may occur with small tumors or with less aggressive pulmonary malignancies such as carcinoid or lepidic predominant adenocarcinoma. Also, the results of PET scans in patients with lesions smaller than 1cm in diameter on chest CT may not be reliable because few such patients were included in the PET studies described earleir. Although FDG-PET is expensive, it has been shown to be cost-effective in a strategy combined with chest CT if the pretest probability of malignancy is 12% to 69%. With probabilities lower than 12%, observation is more cost-effective, and above 69%, resection is more cost-effective.
The success of PET in the diagnosis of malignancy in a primary nodule led to much enthusiasm in the assessment of mediastinal nodes. In one of the most quoted studies evaluating PET, Pieterman and colleagues prospectively compared PET versus conventional staging in 102 patients who underwent staging of the mediastinum. PET improved sensitivity from 75% to 91% compared to CT, and it improved specificity from 66% to 86%. The combination of the two approaches had a sensitivity of 94% and a specificity of 86%. Initial studies documented improved accuracy for FDG-PET compared to chest CT, with an increase or decrease in TNM stage in 21% to 40% of patients., In addition, the combined use of CT and PET was found to be cost-effective. However, much as in the assessment of malignancy in a primary nodule, false-positive results may occur. In the mediastinum, this occurs most often with infection, inflammation, hyperplasia, sarcoidosis, and anthracotic nodes.,, Because the identification of mediastinal nodal disease changes treatment recommendations, and the positive predictive value for PET positive mediastinal nodes is only 64% (95% CI: 43-80% ) based on a trial of early stage biopsy proven NSCLC (ELPET), PET positive mediastinal nodes must be biopsied. In contrast, in the same trial the negative predictive value in patients with early stage NSCLC and PET negative mediastinal nodes was 95% (95% CI: 90-98%) suggesting that such patients do not require pre-resection invasive mediastinal staging. A European prospective randomized trial assessing the role of PET in staging was the PET in Lung Cancer Staging (PLUS) trial, which randomized 188 patients to either PET or conventional staging. The end point of the trial was futile thoracotomy (defined as thoracotomy without resection, thoracotomy with mediastinal nodal involvement, or recurrence within 1 year after resection). Although the PET group had a significantly lower rate of futile thoracotomy compared to the conventional staging group (21% versus 41%), criticisms of this study have focused on inadequacy of the conventional staging (few studies despite 24% of patients having stage III disease) and inadequate evaluation (15% with weight loss and 9% with poor performance status) . Patients in this trial had clinical suspicion of NSCLC so some of the benefit of PET was attributable to PET characterization of the primary lung lesion. Toloza and colleagues, in their recent meta-analysis, found a false-positive rate of 22% for PET in assessment of mediastinal involvement. Another recent meta-analysis demonstrated a 30% rate of malignancy in CT-documented enlarged mediastinal lymph nodes when the PET was negative.Other studies also call into question the accuracy of PET in mediastinal nodal assessment. A recent retrospective review from Duke University assessed 202 patients who underwent PET followed by mediastinoscopy; PET was found to have 64% sensitivity, 77% specificity, and 74% accuracy for mediastinal disease. In this study, the false-positive rate was 32% and the false-negative rate was 12%. The American College of Surgeons Oncology Group (ACOSOG) published the results of their Z0050 trial, which sought to assess the utility of PET in staging of operable patients. Of the 303 patients randomized, 84% had clinical stage I or II disease, and 16% had clinical stage IIIA lung cancer. The sensitivity of PET for mediastinal nodal involvement was found to be 61%, with a specificity of 84%, a negative predictive value of 87%, and a positive predictive value of 56%. Taken together, these data demonstrate that PET lacks the accuracy to replace invasive staging for assessment of mediastinal nodal disease in NSCLC. A negative PET scan in the mediastinum does not preclude N2 or N3 disease, and a positive PET result needs to be verified pathologically. In particular, nodes that are suspicious on CT should be assessed by invasive staging, even if PET is negative.
PET has been studied in the assessment of unsuspected distant metastatic disease. In patients who have symptoms of systemic metastases (e.g., fatigue, weight loss, anorexia), PET allows total-body imaging with one study. Pieterman and associates found undiagnosed metastases in 11% of patients. In the ELPET trial, patients with proven NSCLC were randomized to conventional staging versus CT chest plus PET. Identification of metastastic disease in the PET arm was 13.8% vs 6.8% in the conventional staging arm. However, PET incorrectly upstaged 4.8% compared to 0.6% in the conventional arm, underscoring the need to confirm positive PET findings. In the ACOSOG Z0050 study, the sensitivity of PET for metastatic disease was 83%, and the specificity was 90%; the negative predictive value was 99%, but the positive predictive value was only 36%. Therefore, a negative result in an asymptomatic patient obviates further attempts to stage patients as M1 before resection. However, a positive result requires histologic confirmation.
In excluding metastatic disease, PET has proved more accurate in some areas than in others. Because of high FDG uptake in normal gray matter, PET has not been useful for identification of brain metastases. For adrenal metastases, PET was found to be 100% sensitive in three studies.,, Also PET may replace bone scans based on it high sensitivity and specificity.
Analysis of sputum cytology has previously been used for diagnosing lung cancer. Its accuracy depends on several factors. Specimens may be induced by saline or collected as a 3-day sample of morning sputum that is preserved in 50% ethanol/2% polyethylene glycol or 70% ethanol. In a recent meta-analysis of 16 studies of unselected patients, the overall sensitivity was 66%, and the overall specificity was 99%.However, an earlier study of patients presumed clinically to have lung cancer revealed a sensitivity of 87% and a specificity of 90%. Another factor of importance in optimizing accuracy is the number of samples obtained. Several authors have shown optimal accuracy with three samples per patient.,, Sputum cytology is also more sensitive for central versus peripheral lesions. Finally, larger tumors, tumors associated with atelectasis/obstruction, and lower lobe tumors may have higher yield. Because false-positive results may occur, albeit infrequently, positive results require clinical correlation. Wtith the increasing frequency of adenocarcinomas which are more likely to be peripheral, sputum cytology has become less useful and is no longer recommended. F.urthermore, attempts to collect adequate sputum samples may lead to delay in diagnosis, or false reassurance if samples are negative.
The availability and ease of performance of flexible bronchoscopy have made this an indispensable test in the diagnosis of lung cancer. It may be performed transnasally, transorally, or via laryngeal mask airway or endotracheal tube. Methods of obtaining cells or tissue include direct biopsy, brushing, bronchoalveolar lavage, and transbronchial biopsy including using a radial ultrasound miniprobe for localization. If endobronchial tumor can be visualized, direct biopsy/brushing has been shown to have a sensitivity of 80% to 100%. As with sputum cytology, bronchoscopy is much more likely to be successful in diagnosis of central lesions compared with peripheral lesions. The combination of bronchoscopy with CT abnormalities may also increase diagnostic accuracy. The presence of an endobronchial lesion on CT is strongly associated with a positive tissue diagnosis on bronchoscopy. In addition, the presence of a bronchus within or leading to a lesion on CT increases the yield of bronchoscopy particularly when combined with radial miniprobe.
Transbronchial needle aspiration was originated in 1958 by Schiepatti. In the early 1980s, Wang and Terry used a 20- to 22-gauge rigid needle to perform transbronchial needle aspiration. Experience with this technique has led to a sensitivity of 50% and a specificity of 96%. Addition of fluoroscopy may lead to higher diagnostic yield. Although this technique is useful in establishing a diagnosis, it was limited in mediastinal staging because of the inability to determine whether a node or bronchial lesion or both was included in the specimen.These blind biopsy techniques have largely been replaced by Endobronchial Ultrasound Transbronchial Fine Needle Aspiration ( EBUS-TBNA)( covered in detail elsewhere).
Advanced bronchoscopic techniques have been developed including: virtual bronchoscopy, radial probe endobronchial ultrasound, electromagnetic navigational bronchoscopy and robotic bronchoscopy. These techniques are described in detail elsewhere, but share a common goal of enabling the operator to identify peripheral lung nodules beyond the reach of conventional bronchoscopy either for purposes of transbronchial biopsy or to facilate localization for resection. At this point in time, the ability to diagnose peripheral lung cancer by advanced bronchoscopic techniques is inferior to CT guided biopsy but this will likely change with further technical improvements.
The discussion so far has applied to bronchoscopy performed with a standard white light. More recently, green or blue light has been used in a technique called autofluorescence bronchoscopy. Because normal, dysplastic, and malignant cells show differences in autofluorescence when exposed to green or blue light, this technique may allow earlier detection of premalignant and malignant lesions. The best known system is the Light Induced Fluorescence Endoscopy (LIFE) device, which uses a helium-cadmium laser to produce 442-nm light. Use of this procedure enhances the ability to detect preinvasive lesions and carcinoma in situ 1.5- to 6.3-fold.,,,Limitations to the procedure include applicability only to squamous cell cancer, lack of evidence that eradicating premalignant lesions produces a survival advantage, and cost. Despite these factors, the technique may be useful as a screening study in high-risk patients.
Transthoracic Needle Aspiration
Percutaneous transthoracic fine-needle aspiration is used to make a tissue diagnosis in patients. Both fluoroscopically guided and CT-guided approaches are available, but CT has become preferred because of superior anatomic precision and improved diagnostic yield. In addition, CT may allow biopsy of non-necrotic portions of masses. Transthoracic biopsy is 90% sensitive, with a 98% specificity. The size of the lesion may be important in sensitivity, with a trend toward lower sensitivity for smaller lesions (as low as 78% for lesions < 1.5 cm in diameter). The type of needle is also important, in that use of a cutting needle, as opposed to a fine needle, allows more specific diagnosis of nonmalignant lesions., Core biopsies can now be obtained and these allow molecular testing for EGFR mutations, alk fusion protein, PDL-1 expression etc. Molecular subtyping of lung cancer, particularly adenocarcinoma, is increasing in importance in guiding treatment.
The role of needle biopsy is controversial. A positive biopsy may faciliate informed discussions with the patient and in planning staging (specifically the need for invasive mediastinal staging) but a negative result does not rule out cancer. This raises the question of the role of biopsy in an operable individual. Although a positive test allows an earlier treatment plan, it does not lead to avoidance of surgery. However if a clearly benign lesion is diagnosed then surgery may be avoided. More often a noncancer result is reported as nondiagnostic, or can not rule out cancer. In such cases, the surgeon is faced with proceeding with surgery for diagnosis and possible treatment if malignancy is confirmed on frozen section. The alternative to surgical resection is ongoing CT surveillance and resection if the lesion grows, or repeat CT guided biopsy. A PET scan may be helpful but as noted above maybe falsely positive or negative.
In a larger lesion that has a higher probabliity of malignancy, a CT guided needle biopsy may be helpful in patients who are not operable because of comorbidities. Additionally larger tumors are more likely to have metastatic disease and thus should have invasive mediastinal staging.
The main complications of this procedure are pneumothorax and hemoptysis. Pneumothorax requiring intervention occurs approximately 1.6% to 17% of the time, and chronic obstructive pulmonary disease is an important risk factor.,, Most pneumothoraces are apparent 1 hour after the procedure, and fewer than 20% of patients require chest tube placement., Hemoptysis is more rare, occurring in 5% to 10% of cases. Because of the small-gauge needles used, massive hemoptysis is extremely rare.,
Acquisition of mediastinal lymph nodal tissue is useful for both diagnosis and staging of lung cancer. Cervical mediastinoscopy is a technique that needs to be in the armamentarium of all practicing thoracic surgeons. First described by Carlens in 1959, the technique involves placing a rigid lighted scope through a transverse suprasternal incision into the avascular pretracheal space. This allows forceps biopsy of nodal stations 2R, 2L, 4R, 4L, and 7. Great care must be taken to avoid the azygos vein when assessing 4R, the innominate artery when assessing 2R and 4R, and the right pulmonary artery when assessing 4R and 7. If injury to a large vessel occurs, the area is packed, and immediate median sternotomy or right thoracotomy is performed. Despite these high-risk areas, cervical mediastinoscopy has proved to be extremely safe. A review of 2137 patients revealed a morbidity rate of 0.6%, with 0.05% mortality.
Cervical mediastinoscopy has been the gold standard for invasive staging of the mediastinal lymph nodes. A recent meta-analysis of 5687 patients with lung cancer demonstrated a sensitivity of 81% and a specificity of 100%. In addition, negative mediastinoscopy predicts a 93% rate of complete resection. However, EBUS-TBNA of mediastinal lymph nodes has become more common and appears equivalent to mediastinoscopy in accuracy (see chapter on EBUS).
Although mediastinoscopy is both safe and effective, many surgeons do not use it routinely. Mediastinoscopy is clearly indicated for lymphadenopathy greater than 1 cm on CT and for positive mediastinal nodes on PET. Other indications include T3 or T4 tumor, central tumors, evidence of N1 disease on CT, adenocarcinoma histology, and large-cell histology. T2 tumors are considered a relative indication by many surgeons although others would routinely do invasive staging for T2 tumors.
Left anterior mediastinotomy was popularized by McNeill and Chamberlain to biopsy enlarged lymph nodes at either level 5 or level 6 in left upper lobe tumors. This technique involves a transverse incision over the 2nd rib. After removal of the 2nd costal cartilage, blunt dissection of the pleura reveals the para-aortic space, and nodal tissue may be biopsied. Many surgeons do not remove the costal cartilage, instead using a mediastinoscope to visualize the area for forceps biopsy. Review of the literature on this technique reveals it to be effective and safe, with low morbidity (8%) and no mortality.
Another option for assessment of level 5 and level 6 nodes in left upper lobe tumors is extended cervical mediastinoscopy, in which blunt dissection creates a space posterior to the innominate vein and anterior to the aorta between the left carotid and innominate arteries. The mediastinoscope is passed into this space, and biopsy may be performed. This procedure is avoided in patients with calcified or postoperative ascending aortas and aortic arches. Catastrophic bleeding has been reported and this technique is rarely used.
Endobronchial and Endoscopic Ultrasound Guided Fine Needle Aspiration
Endobronchial Ultrasound Transbroncial Fine Needle Aspiration
Endobronchial Ultrasound Transbronchial Fine Needle Aspiration ( EBUS-TBNA) is an important technique which should be in the armamentarium of the thoracic surgeon. It has largely replaced mediastinoscopy in many centers. For more detail see the chapter elsewhere in this book. EBUS-TBNA can access the same stations as mediastinoscopy as well as hilar lymph nodes and has been proven to have similar accuracy. It is safe with low morbidity and no reported mortality. It can be performed in with conscious sedation, obviating the need for general anesthesia. In the event of a negative EBUS-TBNA in a patient with a suspicious node on either CT or PET, mediastinoscopy is recommended.
Endoscopic Ultrasonography/Fine-Needle Aspiration
Another diagnostic technique that has gained popularity is fine-needle aspiration through an esophagogastroscope using endoscopic ultrasound (EUS) guidance. Because EUS can easily visualize lymph nodes that are in proximity to the esophagus, this technique allows sampling of nodes in the posterior mediastinum, which may not be accessible via mediastinoscopy. Several prospective studies demonstrate a sensitivity of 88% to 96% and a specificity of 80% to 100% for detection of posterior mediastinal metastases.,,,,,,A recent study demonstrated the utility of EUS-guided fine-needle aspiration to verify PET hot spots in the lower mediastinum and retroperitoneum. In this study, 62% of patients were spared mediastinoscopy and thoracoscopy/thoracotomy.
Video-Assisted Thoracoscopic Surgery
Video-assisted thoracoscopic surgery (VATS) has become an invaluable tool in the diagnosis of lung cancer. It allows high-quality visualization of the thorax and biopsy of tissue through small incisions without painful rib spreading. Development of fiberoptics and laparoscopic equipment that may also be used in the chest allowed expansion of this approach. This approach requires general anesthesia and a dual-lumen endotracheal tube.
Many studies have demonstrated the ability of VATS lung biopsy to effectively diagnose indeterminate lung nodules with a sensitivity and specificity nearing 100%.,, Failure has occurred mainly with small nodules situated deep in the lung parenchyma. However, with experience in correlating CT images with thoracoscopic anatomy as well as newer localization techniques, these nodules can almost always be excised thoracoscopically. In addition to resection of indeterminate nodules, mediastinal lymph nodes that are inaccessible by other methods may be sampled thoracoscopically.,
Organ-Specific Scanning for Diagnosis of Distant Metastatic Disease
Several studies have demonstrated low yield in routine multiorgan scanning for distant metastatic disease in asymptomatic patients., The frequency of such findings was 2.7% to 15%. A large meta-analysis revealed a high negative predictive value (89%-95%) for the clinical evaluation using head CT, abdominal CT, and bone scan as gold standards. Many clinicians believe that routine multiorgan scanning is indicated in the presence of significant comorbidities that would make resection hazardous, T2 - T4 tumors, or stage IIIA disease.
Brain CT produces a false-positive result in 11% of patients. Routine imaging of the brain in asymptomatic patients with early-stage lung cancer is not a cost-effective strategy; the risk of brain metastasis is low, and the sensitivity of head CT is also low. However, all patients with suspicious neurologic symptoms or with locally advanced or distant metastatic disease should have imaging of the brain. A magnetic resonance imaging (MRI) of the brain is recommended because of its superior senstivity. Brain imaging is not recommended for clinical stage I NSCLC however should be performed for all SCLC.
Bone scanning has been useful in the past because of high sensitivity, but these scans are notorious for lower specificity. One study noted a false-positive rate of 40% and another demonstrated that only 50% of solitary foci of uptake represented metastases. Routine bone scans in asymptomatic patients with early-stage lung cancer is not a cost-effective strategy. All patients who have suspicious symptoms of pain, locally or distantly advanced disease, or elevated serum calcium or alkaline phosphatase concentrations need to be investigated but bone scanning has been essentially replaced by PET scans.
Abdominal CT scanning may identify the adrenal adenomas that are present in 2% to 8% of the population. PET scanning is very accurate in differentiating benign adenomas or hyperplasia from metastases.
The power of a staging system, based on large databases, in predicting prognosis is self-evident. Nevertheless, there is an inherent inaccuracy in this staging process. According to the TNM system, the predicted 5-year survival rate after complete resection for T1 N0 M0 NSCLC (stage IA) is only 67%. Therefore, 33% of patients with stage IA NSCLC are incorrectly staged at presentation. Even with optimal therapy, these patients will die of their disease, predominantly from the development of metastatic disease not detected at the time of diagnosis and initial therapy, despite the use of standard staging procedures. Similarly, a significant fraction of all patients with stage IB or II disease are incorrectly staged, resulting in inaccurate assessment of extent of disease, prognostic stratification, and selection of therapy.
Molecular biologic substaging—the use of molecular markers as a strategy for risk stratification—has been validated in retrospective studies ,,,,,,, and is under evaluation prospectively. Assessment of the primary tumor with molecular techniques may improve the prognostic stratification of patients with NSCLC by predicting which patients are most likely to experience recurrence after surgical resection. In addition, the profile of the primary tumor may be used to assess its sensitivity to selected adjuvant therapy.
Molecular biologic staging in patients with stage I NSCLC may have the potential to alter therapy, in addition to improving risk stratification. The ability of molecular biologic markers to predict results of chemotherapy would enable the clinician to design therapy based on the individual tumor. In addition, identifying and understanding the mechanisms of treatment resistance offers another pathway to intervene, by blocking or reversing the mechanism of resistance. Furthermore, understanding of the molecular mechanism of receptor activity and DNA repair enables the study of pharmacologic targeting with chemotherapy or biologic agents. The ultimate power of molecular biologic staging depends on the ability to alter therapy and improve outcome, which has not yet been demonstrated.
In a patient who presents with a presumed lung cancer based on CXR or CT, the next step depends on symptoms or signs. Symptoms or signs are usually related to locally advanced or metastatic disease.The goal is to accurately stage the patient with the least number of tests. If a patient has symptoms suggestive of metastatic disease, this is investigated first and if stage IV disease is confirmed, no further tests are required. For example a liver lesion may be biopsied and the diagnosis pathologically confirmed and at the same time stage IV disease is confirmed. This is in contrast to doing a biopsy of the lung first and then a biopsy of the liver. If the patient is asymptomatic, a CT chest is generally followed by a PET scan to look for extrathoracic disease and evaluate the mediastinum. As noted above, if the PET suggests extrathoracic metastasis, confirmatory testing either by biopsy or further imaging is required. If there is no evidence of extrathoracic metastasis, the mediastinum is assessed. Enlarged or FDG avid nodes are biopsied. If negative, consideration is given to mediastinoscopy or proceeding to surgery. Prior to undertaking surgery, physiologic assessment is undertaken including cardiopulmonary assessment as appropriate, assessment of frailty, performance status and sarcopenia or nutritional status. If the patient is not an operative candidate for reasons of comorbidites or patient preference, alternatives to surgery are considered. In most cases, a definitive cancer diagnosis is required for nonsurgical treatment. The role of lung biopsy is controversial in a surgical patient. A nondiagnostic CT guided or transbronchial biopsy does not rule out cancer and surgical resection for diagnosis should be performed if the patient is candidate for surgery.
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES
Section: KEY REFERENCES