Lung cancer is the leading cause of cancer-related deaths in men and women worldwide. The incidence of lung cancer in women has increased even faster than the overall incidence, reflecting their increased use of tobacco in the past 30 years. Most patients present late with advanced disease, which results in a poor prognosis.

The most important prognostic indicator in lung cancer is the extent of disease. The Union Internationale Contre le Cancer (UICC) and the American Joint Committee for Cancer Staging (AJCC) have developed the tumor, node, and metastases (TNM) staging system, which takes into account the degree of spread of the primary tumor, the extent of regional lymph node involvement, and the presence or absence of distant metastases.

Non–small cell lung cancer (NSCLC) accounts for approximately 75% of all lung cancers. NSCLC is subdivided into adenocarcinoma, squamous cell carcinoma, and large cell carcinoma. Despite their histologic and clinical differences, they share a similar prognosis and management and are staged by using the same TNM system.

Surgery is the treatment of choice for NSCLC if the primary tumor is resectable and if metastatic disease is absent. Chemotherapy and radiation therapy are used to treat tumors that are unresectable because of intrathoracic spread or distant metastases. Small cell lung cancer (SCLC), which metastasizes early and has a worse outcome than NSCLC, has a separate staging system (see Staging of SCLC).

TNM STAGING

The stages represent the nature and extent of spread of a neoplasm and, thus, the therapeutic options and prognosis in individual patients. Stages also provide a standard by which various therapies can be compared. A combination of clinical, laboratory, radiologic, and pathologic investigations are used to stage various neoplasms.

Different staging systems have been developed. The most widely used scheme for staging NSCLC is the TNM classification. The UICC and the AJCC initially introduced this system in 1972 with staging and end-results reporting. This scheme has been modified and refined over the years. The UICC and AJCC approved and published the last revision in 1997. Regional lymph node classification was also refined.

The TNM staging system takes into account the degree of spread of the primary tumor, represented by T; the extent of regional lymph node involvement, represented by N; and the presence or absence of distant metastases, represented by M. The TNM system is used for all lung carcinomas except SCLCs, which are staged separately (see Staging of SCLC).

In the TNM systems, 4 stages are further subdivided into I-III and A or B subtypes. These stages have important therapeutic and prognostic implications. The stages are as follows:

• Primary tumor

   

  • Tis - Carcinoma in situ

     

  • TX - Positive malignant cytologic findings, no lesion observed

     

  • T1 - Diameter of 3 cm or smaller and surrounded by lung or visceral pleura (see Image 1) or endobronchial tumor distal to the lobar bronchus

     

  • T2 -Diameter greater than 3 cm (see Images 2-3); extension to the visceral pleura, atelectasis, or obstructive pneumopathy involving less than 1 lung; lobar endobronchial tumor; or tumor of a main bronchus more than 2 cm from the carina

     

  • T3 - Tumor at the apex (see Image 5); total atelectasis of 1 lung; endobronchial tumor of main bronchus within 2 cm of the carina but not invading it; or tumor of any size with direct extension to the adjacent structures such as the chest wall mediastinal pleura (see Image 8), diaphragm, pericardium parietal layer, or mediastinal fat of the phrenic nerve

     

  • T4 - Invasion of the mediastinal organs, including the esophagus trachea, carina (see Image 11), great vessels (see Image 13), and/or heart; obstruction of the superior vena cava; involvement of a vertebral body; recurrent nerve involvement; malignant pleural or pericardial effusion; or satellite pulmonary nodules within the same lobe as the primary tumor

   

• Regional lymph node involvement

   

  • N0 - No lymph nodes involved

     

  • N1 - Ipsilateral bronchopulmonary or hilar nodes involved

     

  • N2 - Ipsilateral mediastinal nodes or ligament involved

     

    • Upper paratracheal lower paratracheal nodes
    • Pretracheal (see Image 4, Image 7, Image 10) and retrotracheal nodes
      
    • Aortic and aortic window nodes
      
    • Para-aortic nodes
      
    • Para-esophageal nodes
      
    • Pulmonary ligament
      
    • Subcarinal nodes (see Images 12-17)

     

  • N3 - contralateral mediastinal or hilar nodes involved (see Image 19) or any scalene or supraclavicular nodes involved

   

• Metastatic involvement

   

  • M0 - No metastases

     

  • M1 - Metastases present (see Images 20-27)

Stage groupings are as follows:

• IA - T1N0M0

   

• IB - T2N0M0

   

• IIA - T1N1M0

   

• IIB - T2N1M0 or T3N0M0

   

• IIIA - T1-3N2M0 or T3N1M0

   

• IIIB - Any T4 or any N3M0

   

• IV - Any M1

For each stage, the prognoses or estimated 5-year survival rates, in the United States are as follows:

• Stage IA - 75%

   

• Stage IB - 55%

   

• Stage IIA - 50%

   

• Stage IIB - 40%

   

• Stage IIIA - 10-35% (Stage IIIA lesions have a poor prognosis, but they are technically resectable.)

   

• Stage IIIB - 5% (Stage IIIB lesions are nonresectable.)

   

• Stage IV - Less than 5%

For each stage, the prognoses, or estimated 5-year survival rates, in Europe are as follows:

• Stage IA - 60%

   

• Stage IB - 38%

   

• Stage IIA - 34%

   

• Stage IIB - 24%

   

• Stage IIIA - 13% (Stage IIIA lesions have a poor prognosis, but they are technically resectable.)

   

• Stage IIIB - 5% (Stage IIIB lesions are nonresectable.)

   

• Stage IV - Less than 1%

METHODS OF STAGING

Staging of the primary tumor

Conventional chest radiography

Conventional chest radiographs (CXRs) usually demonstrate the size of the tumor, especially in peripheral lesions. Central tumors may be associated with atelectasis or obstructive pneumonitis. The proximal extent of central tumors is determined with bronchoscopy.

CXRs may also show a pleural effusion, direct extension into the chest wall with destruction of the ribs or vertebrae, phrenic nerve involvement with elevation of a hemidiaphragm, or mediastinal widening due to lymphadenopathy. In the absence of these signs, CXRs are unreliable in detecting invasion of the chest wall, diaphragm, or mediastinum, and CT or MRI is required to assess these conditions.

Computed tomography

Contrast-enhanced helical CT of the thorax and abdomen that includes the liver and adrenal glands is the standard radiologic investigation for staging lung cancers. The primary tumor should be measured by using lung window settings in 2 dimensions: the maximum long axis and the largest diameter perpendicular to the long axis.

CT reliably depicts mediastinal invasion, provided that the tumor surrounds the major mediastinal vessels or bronchi. A tumor that abuts the mediastinum cannot be considered invasive, even if the fat plane between the mediastinum and mass is obliterated.

CT criteria for resectability include the following:

• Contact between mass and mediastinum of less than 3 cm

   

• Circumferential contact between the mass and aorta of less than 90°

   

• Presence of a fat plane between the mass and mediastinum

Criteria for nonresectability include the following:

• Involvement of the carina

   

• Tumor surrounding, encasing, or abutting the aorta. Main or proximal portions of the right or left pulmonary arteries, or esophagus by more than 180°

Tumors with more than 3 cm of contact and no obvious invasion may be difficult to stage. CT and MRI cannot be used to distinguish tumor invasion of mediastinal fat from inflammatory changes.

Magnetic resonance imaging

MRI is superior to CT in assessing the pericardium, heart, and great vessels. It does not require intravenous contrast enhancement. Coronal images are useful in demonstrating the extent of tumor in the subcarinal region, aortopulmonary window, and superior vena cava. However, it is limited by poorer spatial resolution compared with that of CT and by cardiac and respiratory motion artifacts.

The overall difference in accuracy between MRI and CT is not significant. The sensitivity of CT is 63%, and that of MRI is 56%. In the distinction of T3 and T4 tumors from less extensive tumors, the specificity of CT is 84%, and that of MRI is 80%.

Positron emission tomography

Positron emission tomography (PET) has a limited role in the diagnosis of lung cancer. PET may be used in the workup of the indeterminate pulmonary nodule in a new patient or in a patient with known cancer. Benign nodules such as histoplasmosis can usually be differentiated from a malignant nodule.

Staging of mediastinal lymph nodes

Hilar (N1), ipsilateral (N2), or contralateral (N3) mediastinal lymph node metastases are often present at the time of presentation. The N stage is important because it determines both the prognosis and the suitability of curative surgery. On CXRs, CT scans, and MRIs, nodal enlargement may indicate nodal involvement; however, this finding may be inaccurate. Normal-sized nodes may contain metastases, and nodes may be enlarged due to inflammatory causes although they contain no malignant cells. Furthermore, the classification of nodes and the normal range of the size of mediastinal nodes have been disputed.

On the accepted lymph node map described by the AJCC and UICC in 1997, nodes in position 10 are designated as hilar nodes. The short-axis diameter is the most reliable measurement of lymph node size on CT scans. A short-axis diameter of greater than 10 mm is abnormal regardless of the nodal station.

Conventional chest radiography

Chest radiography is inferior to CT in the detection of mediastinal lymph node metastases. It has a sensitivity of only 10-30%, but its specificity of 90% is higher than that of CT.

Computed tomography

The visualization of mediastinal nodes is facilitated by the use of spiral or multisection CT, thin (5-mm) sections, the presence of a mediastinal fat, and intravenously administered contrast material.

The reported sensitivity and specificity of CT in the detection of mediastinal nodes vary considerably, with ranges of 40-84% and 52-80%, respectively. This variability reflects interobserver variability and differences in the size criteria for abnormal lymph nodes, in patient populations, and in the diagnostic criterion standard. CT is more specific in populations in Europe with a low incidence of granulomatous disease than in populations in the United States, which have a high incidence of histoplasmosis; rates are 80-90% and 50-70%, respectively.

The negative predictive value of CT is about 85%; as a result, patients with normal mediastinal appearances undergo thoracotomy. Mediastinoscopy or thoracoscopy is required during biopsy of enlarged noncalcified lymph nodes before surgery is ruled out.

Magnetic resonance imaging

Like CT, MRI size criteria are used to identify nodal involvement, and these are comparable to those used at CT. However, MRI can be used to distinguish nodes from vessels without intravenous contrast enhancement. Also, direct imaging in the sagittal and coronal planes is possible; this is helpful in the assessment of the subcarinal and aortopulmonary regions.

Positron emission tomography

Unlike MRI and CT, PET does not rely on the anatomic assessment of nodes. It is primarily a metabolic imaging technique that relies on a biochemical difference between normal and neoplastic cells. Mediastinal nodes containing tumor have an increased uptake of 2-[fluorine 18]-fluoro-2-deoxy-D-glucose (FDG), a glucose analog labeled with ¹⁸F, a positron emitter. Its high cost and limited availability have reduced the impact of PET in Western countries. Recent changes in regulations and reimbursement in the United States have increased the use of PET and access to PET centers.

PET is superior to CT in the assessment of mediastinal nodal metastases. In patients with N2 disease, PET had a sensitivity of 83% and a specificity of 94% compared with a sensitivity of 63% and a specificity of 73% with CT. The superiority of PET is even more marked in the assessment of hilar nodes, for which it has a 73% sensitivity and a 76% specificity compared with a 18% sensitivity and a 86% specificity with CT.

FDG PET enables accurate staging of regional lymph node disease in patients with stage I NSCLC. A negative PET scan in these patients suggests that mediastinoscopy is unnecessary and that thoracotomy may be performed.

About 35% of cases first staged with CT, the disease is upstaged after subsequent PET, with resultant changes in management. Nevertheless, PET can produce some false-negative results. In one study, 5 of 39 patients with negative PET scans had a carcinoma, as revealed with pathological staging. In 4 of these patients, correlation with CT findings led to the correct conclusion. Thus, PET and CT are complementary because the visual information on CT enables anatomic discrimination between hilar and mediastinal nodes and more accurate localization of the hot spots. The sensitivity of PET combined with CT was 93%, and the specificity was 97%.

Staging of distant metastases

The detection of distant metastases is of crucial importance because it usually implies that curative surgical resection of the primary tumor is contraindicated. Metastases occur in about 50% of patients with NSCLC. In patients with clinical or biochemical evidence of disease elsewhere, targeted imaging of those sites is performed. These sites include the brain, which can be examined with CT or MRI, and the skeleton, which can be examined with scintigraphy. Usually, these sites are not imaged in asymptomatic patients with NSCLC.

The probability of metastases is highest for SCLC, which is 60-80% on presentation, and lowest for squamous cell cancers; the incidence increases with advancing stage. No such trend exists for cancer involving the other cell types. Adenocarcinoma tends to metastasize to the brain and adrenals early in its course.

Liver and adrenal metastases

The staging CT scan of the thorax is usually extended to include the liver (see Image 20) and adrenal glands (see Images 21-22). CT has a sensitivity of about 85% in the detection of liver metastases. Similar rates may be obtained with MRI and ultrasonography performed by experienced imagers. Ultrasonography is superior to CT in discriminating metastases from liver cysts, which account for most of the benign lesions seen on CT scans.

Adrenal metastases are common and often solitary. They must be differentiated from adrenal adenomas, which occur in 1% of the adult population. Lesions smaller than 1 cm are usually benign. Metastases are usually larger than 3 cm; on nonenhanced CT scans, they have an attenuation coefficient of 10 HU or higher. Adenomas and metastases can also be distinguished by using MRI and PET.

Brain metastases

SCLC and adenocarcinoma are the most common sources of cerebral metastases (see Images 23-24). MRI is superior to CT, especially in the depiction of the posterior fossa and the area adjacent to the skull base. However, the brain is not routinely imaged in asymptomatic patients with NSCLC because the incidence of silent cerebral metastases is only 2-4%.

Bone metastases

Technetium-99m radionuclide bone scanning (see Images 25-26) is indicated in patients with bone pain or local tenderness. The test has a 95% sensitivity for the detection of metastases but a high false-positive because of degenerative disease and trauma the assessment of these require comparison of the bone scans with plain radiographs. Spinal metastases may cause spinal cord compression. Because only about 5% of bony metastases detected with radionuclide scans are asymptomatic, routine preoperative bone scanning is not usually performed.

Lung metastases

Pulmonary metastases (see Image 27) from a primary NSCLC are uncommon on presentation, but they are present at autopsy in 20% of cases. Accurate preoperative diagnosis of small lung nodules depicted on CT scans is often difficult because they may be indistinguishable from granulomata and fibrotic nodules.

Positive emission tomography

In a recent study using PET, FDG uptake was increased in 92% of proven adrenal metastases (23 of 25 patients) and normal in the 8 benign lesions. Whole-body PET will probably decrease the number of adrenal biopsies performed for indeterminate adrenal lesions. In another study, unsuspected distant metastases were found in 11% of patients with primary lung carcinoma.

Whole-body PET is more accurate than thoracic CT, bone scintigraphy, or brain CT or MRI in staging bronchogenic carcinoma. FDG PET results have prognostic value and are strongly correlated with survival rates in patients who undergo treatment for lung cancer. Patients with positive FDG PET results have a significantly worse prognosis than patients with negative results. Additionally, FDG PET may be helpful in guiding treatment.

The recent introduction of combined PET/CT machines (see Image 28) will affect the future workup and treatment of patients with cancer. These machines will also be used in radiation treatment planning.