----
-
accp section report
~III.~_----------
Clinical Staging of Patients with Non-small Cell Lung Cancer* Alex G. Little, M.D., F.C.C.R;t and Frederick R Stitik, M.D.+
This article is the first in a planned series from the Section on Lung Cancer of the ACCP addressing the important and clinically relevant aspects of what is now the most common malignancy in the world, lung cancer. This initial report addresses the problem of staging of lung cancer. Staging, or identifying the anatomic extent of disease according to the AJCC TNM classification scheme, is the first clinical activity in caring for a patient with known or presumed
of this report is not to simply present T hethepurpose recently modified AJCC TNM staging definitions for lung cancer, which are amply described in the following tabulation and Table 1:1
lung cancer because the results determine appropriate types of therapy. This is, therefore, a critically important aspect of the patient's care which forms the foundation for subsequent treatment. In addition, consistent use of this system, based on appropriate clinical and pathologic staging, in stratifying patients in clinical reports is mandatory; otherwise, meaningful comparisons and conclusions are impossible. (Chest 1990; 97:1431-38)
T4
PRIMARY TUMOR (T) TX
TO TIS TI
T2
T3
Tumor proven by the presence of malignant cells in bronchopulmonary secretions but not visualized roentgenographically or bronchoscopically, or any tumor that cannot be assessed, as in a retreatment staging. No evidence of primary tumor. Carcinoma in situ. A tumor that is 3.0 cm or less in greatest dimension, surrounded by lung or visceral pleura, and without evidence of invasion proximal to a lobar bronchus at bronchoscopy, A tumor more than 3.0 ern in greatest dimension, or a tumor of any size that either invades the visceral pleura or has associated atelectasis or obstructive pneumonitis extending to the hilar region. At bronchoseopy the proximal extent of demonstrable tumor must be within a lobar hronchus or at least 2.0 cm distal to the carina. Any associated atelectasis or obstructive pneumonitis must involve less than an entire lung. A tumor of any size with direct extension into the chest wall (including superior sulcus tumors), diaphragm, or the mediastinal pleura or pericardium without involving the heart, great vessels, trachea, esophagus, or vertebral body, or a tumor in the main
*From the Department of Surgery, University of Nevada School of Medicine, Las Vegas; and the Department of Radiology; Eastern Virginia Medical Center, Norfolk, Va. tProfessor and Chairman, Department of Surgery. :l:Professor of Radiology, EVMS. Reprint requests: Dr. Little, 2040 West Charleston, Las ~gas, Nevada 89102
bronchus within 2 em of the carina without involving the carina. A tumor of any size with invasion of the mediastinum or involving heart, great vessels, trachea, esophagus, vertebral body, or carina or presence of malignant pleural effusion.
NODAL INVOLVEMENT (N) NO NI
N2 N3
No demonstrable metastasis to regional lymph nodes. Metastasis to lymph nodes in the peribronchial or the ipsilateral hilar region, or both, including direct extension. Metastasis to ipsilateral mediastinal lymph nodes and subcarinal lyrnph nodes. Metastasis to contralateral mediastinal lymph nodes. contralateral hilar lymph nodes, or ipsilateral or contralateral scalene or supraclavicular lymph nodes.
DISTANT METASTASIS MO M1
(~t)
No (known) distant metastasis. Distant metastasis present.
Our focus is on the appropriate clinical strategy which should be used when staging nonsmall-celllung cancer. Other articles will follow from the Section on Lung Cancer of the ACCP on the pathology, diagnosis and treatment of both small-cell and non-small cell lung cancer. Staging is the clinical strategy of determining and identifying the anatomic extent of disease in a patient with a malignancy. Stratification of patients by stage groupings is crucial; however, when reporting clinical experiences, stratification of patients by precise TNM classifications is essential because of the prognostic differences of patients within stage groupings. If this is not done, meaningful conclusions and comparisons CHEST I 97 I 6 I JUNE, 1990
1431
Table I-Stage Grouping Occult carcinoma Sta~t
,O
Stage I Sta~e
II
Sta~e
iliA
Stage I1IB Stage IV
TX TIS TI T2 TI T2 1'3 1'3 T1-3 AnyT T4 AnyT
NO MO Carcinoma in situ NO MO NO MO Nl MO NI MO NO MO Nl MO N2 MO N3 MO AnyN MO AnyN Ml
between reports are not possible . In addition, in the case of lung cancer, the stage of disease correlates very well with the patient's prognosis and, therefore, serves to guide development of therapy; however, translating the need for accurate staging into clinical practice poses difficulties. Should every possible test be performed in every patient? Are all tests unnecessary and simply a thorough history and physical examination sufficient? Our goal in this review is to present "state-of-the-art" answers to these questions. The aim of clinical staging is primarily to determine the AJeC stage . Although the TNM system is presented in the usual order of T~N~M, staging in clinical practice is usually done in the reverse order of M~N~T. If a metastasis is found, the patient has M1 disease which is classified as stage IV, and special efforts to stage the mediastinum or hilum (N) and primary tumor (T) are otiose, as results will not change the patient's stage . Similarly, if there are no metastases and N3 nodes are involved , then the classification of the primary tumor cannot change the stage, which is I1IB; however, there is heterogeneity even within
FIG UHE 2 . Contralateral pulmonary metastasis. PAche st rad iograph of a 72-year-old man with left upper lobe atelectasis due to squamous carcinoma and several large and small pulmonary parenchymal metastases in the right lung. Arrow indicates the largest of the metastases.
stage groupings, as the prognosis of all patients is not identical; for example, within stage IlIA, patients with T3NOMO disease differ in many ways from patients with TlN2MO extent of disease. This distinction between T and N classifications, therefore, mayor may not be important in determining care for a particular patient. If not important, the staging endeavors cease when the clinician believes that enough information is available to determine therapy.
Flt ; UHE 1. Metastatic disease demonstrated on conventional chest radiograph . (A, left). PA radiograph of a 66-year-old man with a 3 em left upper lobe adenocarcinoma and destructive lesion of the anterior right fourth rib . (B, right). Close-up of the de stroyed anterior rib . Compare the superior cortical margins of the rib immed iately above and below the involved rib (arrows) .
1432
Staging Patients with Non-small Cell Lung Cancer (UttJe. Stitik)
STAGING FOR METASTATIC DISEASE
The first major question which requires answering when a patient presents with known or presumed lung cancer is: does the patient have a distant organ metastasis? Obtaining a thorough history from the patient and combining it with a careful physical examination answers this question in many instances. 2.3 Symptoms, such as a new headache or bone pain, or 6ndings during the physical examination, such as an enlarged liver, direct the physicians attention to specific organs which should be evaluated with a focused examination such as a brain CT or a technetium-99m polyphosphate bone scan. Simple laboratory tests, such as a CBC, hepatic function tests, and serum calcium determination, may suggest the presence of distant metastases and should be pursued appropriately. The standard PA and lateral chest roentgenogram may also provide evidence of metastatic disease to the contralateral lung, pleura, or thoracic osseous structures (Fig 1 and 2). The more difficult issue is how to stage a patient who does not have symptoms, signs, physical findings, or laboratory values suggesting the presence of metastatic disease. Radionuclide Scanning
Two radionuclide scans can be useful in screening for clinically silent metastases. The first is the technetium-99m polyphosphate bone scan, which is sensitive for occult involvement of bone, a relatively common site of metastases. Routine bone scanning is often performed, even though the detection rate of truly occult metastases (without even minimal signs and symptoms or laboratory abnormalities) is low There is a false-positive rate, which means that roentgenograms of the involved bone or other examinations are required in some patients to disprove the scan. This rate is less than 20 percent in one series, and therefore it is argued that it does not unnecessarily prolong the clinical staging process. 4 In addition to the bone scan, the gallium-67 citrate scan is useful in detecting unsuspected and silent metastases but only in specific and, unfortunately, limited circumstances," The primary tumor must be 2 em or greater in size and avidly take up gallium. The test must be performed with a high dose of the radiopharmaceutical, and the gastrointestinal tract must be completely evacuated to eliminate it as a source of isotope retention; however, most importantly, the scan must be performed with a large crystal gamma camera which obtains tomographic images. These scanners are not available at many institutions, and without them, this examination is of limited usefulness; however, when they are available, results are useful. In one report, gallium scanning detected occult metastases to distant organs in 11 percent of
100 patients without the occurrence of any falsepositive scans.6 Computed Tomography
Computed tomography of the chest and upper abdomen, down to and including the liver and adrenal glands, should be considered a routine examination to be performed in all patients in whom no metastatic disease is identified by history or physical examination. In addition to information about the primary tumor and the lymph glands within the mediastinum, cr detects metastatic pulmonary nodules, small pleural effusions, metastatic deposits in the liver, and metastatic involvement of the adrenal glands. There are few data assessing the value of this scanning technique as a screen for metastatic disease in asymptomatic patients. One retrospective review of 85 patients found that CT confirmed the presence of metastases causing symptoms in five patients and detected asymptomatic metastases in seven patients (8 percent)." Computed tomography of the chest is recommended for assessment of the primary tumor and the regional lymph nodes in every patient in whom neither the history, physical examination, nor chest roentgenogram identify Ml disease. Extending the scan to include the liver and adrenal glands is a simple, quick, and easy means to increase the sensitivity for detection of metastases. Some physicians believe that CT of the brain should be used routinely, particularly in patients with proven or suspected adenocarcinoma, because of the predilection of these tumors for metastasis to this site. In one clinical study," albeit retrospective, 16 (18 percent) of 89 patients had abnormal CT scans of the brain; however, nine of the 16 had evidence of CNS disease by history or physical examination, and the other seven had strong clinical evidence of disseminated disease. In other words, no abnormal cr scans were found in patients with completely normal clinical examinations. Therefore, there are no data to support routine brain cr in patients in the absence of symptoms suggestive of brain involvement. Magnetic Resonance Imaging
Magnetic resonance imaging (MRI) has become an important method in the diagnosis of many diseases. The value of this relatively new technology in patients with lung cancer is constantly being defined. Magnetic resonance imaging is frequently compared to cr in assessing patients with lung cancer. In the thorax, MRI is generally considered equal to Cf,9-18 although in certain limited areas, as will be discussed in this article, MRI may be more useful. Currently, cr is the standard for patients with known or suspected lung cancer, with MRI playing a problem-solving role. The technology for CT scanning of the chest has CHEST I 97 I 6 I JUNE. 1990
1433
remained stable for a number of years; however, the technology for MRI is rapidly improving. In years to come, MRI may supplant CT of the chest in some or all of the current indications. In general, MRI has the following advantages over CT : (1) no ionizing radiation; (2) better contrast resolution; (3) no intravenous iodinated contrast material necessary; (4) the ability to image in many planes, ie, coronal and sagittal planes, without significant loss of spatial resolution; (5) less degradation by metallic clips than CT ; and (6) the potential for tissue characterization. On the other hand, CT (1) has better spatial resolution; (2) is quicker (30 minutes per examination for CT vs 60 minutes for MRI); (3) is less expensive; (4) is more readily available in many hospitals; (5) is degraded less by vascular and cardiac motion; and (6) has the ability to detect calcium, which MRI does not. Magnetic resonance imaging may be more sensitive in detecting brain metastases, although it has generally not replaced CT in the evaluation of patients with signs or symptoms suggesting CNS metastases. We currently employ MRI to detect possible brain metastases when the CT findings are equivocal or the patient is allergic to iodine. Magnetic resonance imaging is probably equal to CT in the detection of hepatic metastases, although some investigators believe that MRI may be more sensitive;'? however, this is generally a moot point, as the liver is studied at the time of chest CT. Magnetic resonance imaging of the liver could be used to solve problems unresolved by CT or in those patients who are allergic to iodinated contrast material. In the future, MRI may play an important role in evaluating patients who have lung cancer with suspected adrenal metastases. Asymptomatic adrenal masses, likely nonfunctioning cortical adenomas, are often present in the normal population, with this being an incidental finding on abdominal CT in approximately 1 percent of patients.P' These adrenal masses are well seen by CT but cannot be characterized as metastases or benign conditions. Early studies have suggested that MRI , with its better contrast resolution , may be able to separate benign from malignant disease ." Other investigators report a significant overlap in imaging characteristics." The investigational radiopharmaceutical, NP-59, may help further separate benign from malignant adrenal masses. 23
stage IV and relinquishing hopes for curative therapy; for example, in one study, evaluation with an open rib biopsy of the ambiguous finding of a solitary rib abnormality on a bone scan, using a guided operative technique, proved the absence of metastatic disease in 50 percent of the cases." Half of the patients would have been inappropriately considered to have M1 disease without this aggressive surgical approach oriented around requiring histologic documentation of the nature of the tissue when the results of noninvasive examination are ambiguous. Similarly, Oliver et al25 reported the findings in 330 patients with non-small cell lung cancer who underwent CT of the thorax, including the upper abdomen. In 25 patients an adrenal mass was discovered as the only possible evidence of metastatic disease . In these 25 patients, eight were proven to have true metastatic disease by CT-guided adrenal biopsy or follow-up, while the majority (17) were shown to have benign disease by CT-guided biopsy, surgery, or follow-up." STAGING OF REGIONAL LYMPH NODES
Computed Tomography As mentioned, chest CT should be obtained in all patients. When chest CT does not show any mediastinal lymph node larger than 1 em in diameter, the chance of nodal metastasis is low, less than 10 percent for the mediastinum as a whole, and invasive mediastinal staging is not required (Fig 3).26-2H When results are analyzed for different anatomic nodal basins within the mediastinum, regional differences in this negative predictive index are found . The lowest accuracy is in the region of the aortopulmonary window, where the subaortic nodes are located .2H Fortunately, these pa-
Invasive Staging When a noninvasive test or a combination of noninvasive tests unequivocally establishes the presence of Ml disease, then histologic documentation is not required; however, if an abnormal finding is not unequivocally believed to represent a metastasis, then tissue confirmation of the nature of the abnormality should be sought before consigning the patient to 1434
FIGURE 3. False-negative cr scan of a 64-year-old man with a 2 cm adenocarcinoma in the right upper lobe and multiple small pretracheal and aortopulmonary window lymph nodes, none measuring more than 1.5 cm . At medi astinoscopy; the small lymph nodes conta ined metastatic adenocarcinoma. Staging Patients with Non-small Cell Lung Cencer (Little. Slitik)
a single enlarged node." Calcium in enlarged nodes because of granulomatous disease will not he detected by MRI.
Invasive Staging
FIGURE 4. False-positive cr scan of a 70-year-old man with squamous carcinoma causing left upper lobe atelectasis and pneumonia. A large pre-aortic lymph node measuring 2 cm is present (armw) . At thoracotomy, the enlargement was due to reactive hypertrophy, likely secondary to the obstructive pneumonia.
tients have a better prognosis than patients with involvement of other mediastinal nodal basins, so surgical treatment is still reasonable and understaging in this region does not result in inappropriate therapy;29 however, when lymph nodes are enlarged greater than 1 cm in diameter, invasive staging with mediastinoscopy needs to be performed , as approximately 30 percent of these enlarged lymph nodes will be proven to be due to reaction and inflammation, rather than metastatic disease (Fig 4). Finally, it is important to note that the reliability of CT staging varies according to cell type, as in one study the accuracy rate for adenocarcinoma was 94 .7 percent, compared to 70.6 percent for squamous cell carcinoma.P' The previous considerations represent a rea sonable assessment of the value of CT in patients with suspected N2 or N3 disease . Many controversies exist, as reviewed by several authors.v-"
Magnetic Resonance Imaging Currently, the value of MRI in the detection of nodal metastases suffers from the same lack of sensitivity and specificity as CT scanning and for exactly the same reasons; both only characterize a node by its size and location, and neither is a reliable indicator of histology. It was hoped that with MRI's ability to characterize tissue, benign and malignant nodes could be separated . Several studies have shown a significant overlap in values from benign and malignant hilar and mediastinal nodes studied in vitro, as well as in vivo .9 • 1ll •32 .:J.1 Because of its ability to image in the coronal plane, MRI may he ahle to demonstrate the size of suhaortic nodes to a better extent than CT. I" Magnetic resonance imaging may he of value in the assessment of hilar nodes in patients allergic to iodinat ed contrast material. Because of the poor spatial resolution of MRI , occasionally small lymph nodes are not recognized as discrete nodes and are visualized as
The supraclavicular and scalene (N3) lymph nodes are accessible to manual palpation and should he carefully examined in this fashion . Although "blind" scalene node biopsy is inappropriate because of the low yield in the ahsence of palpable adenopathy, all distinctly abnormal nodes are excisionally biopsied for staging. This does not mean that every lymph node that can be palpated is removed, just those that are suspicious because of their size or consistency. Some surgeons continue to strongly defend the role of routine mediastinoscopy in all patients;:J.'J ·34 however, although mediastinoscopy, both cervical and parasternal approaches, is very safe, the yield is too low to justify its use when the identified mediastinal lymph nodes are less than 1 cm in diameter on CT. When larger mediastinal nodes are seen on CT, mediastinoscopy is appropriate, as the positive predictive accuracy of the scan's results is only approximately 70 percent. Therefore, if the finding of enlarged nodes is accepted as proof of involvement of N2 or N3 nodes, fully 30 percent of patients will be inappropriately staged . The main value of CT demonstrating enlarged nodes is that it can be used as a road map and help to select the best test for histologic evaluation. The choice of mediastinoscopy, median sternotomy, transbronchial needle aspiration biopsy, percutaneous needle aspiration biop sy, or thoracotomy can be facilitated by careful analysi s of the CT scan, local physician expertise, and a thorough knowledge of the new staging system , and, therefore, treatment ramifications should the enlarged nodes be positive. The location of enlarged nodes on CT, or positive nodes found at mediastinoscopy or thoracotomy, should be carefully recorded using the AJCC' or the ATS lymph node map ." A better understanding of the significance oflymph node metastasis may help define future staging systems and therefore modify treatment and prognosis. PRIMARY TUMOR
When the patient has neither chest wall nor pleuritic pain in the location of the primary tumor and both the chest roentgenogram and CT show that the tumor is neither at the limits of the visceral pleura nor a centrally located lesion, then the tumor is, by definition, either Tl or T2 . The distinction between Tl or T2 tumors does not affect the stage and is not crucial for making initial therapeutic decisions. Patients with localized chest wall pain in the region of their primary tumor can he assumed to have a T3 or T4 lesion which is transgressing and irritating the parietal pleura; CHEST I 97 I 6 I JUNE, 1990
1435
FIGURE 5. Superior vena caval obstruction. (A, left). PA chest radiograph of a 74-year-old white man presenting with superior vena caval obstruction. A small-cell undifferentiated carcinoma is seen in the nght paratracheal area with moderate displacement of the trachea to the left . (B, right). cr scan after bolus injection of contrast material demonstrating the large mass, compression and displacement of the azygos vein and superior vena cava (arrows), and mult iple collateral vein s in the chest wall.
however, when the primary tumor abuts the chest wall on the chest roentgenogram, but the patient has no local symptoms, then it is difficult to accurately identify the level of penetration of the tumor, ie, does it simply touch, or does it actually invade the parietal pleura?
Computed Tomography Chest CT may unambiguously identify invasion of the soft tissues or bony structures of the chest wall or mediastinum, thus identifying the primary tumor as T3 or T4, depending upon the involved structure and the depth of invasion (Fig 5 and 6). When CT (or chest roentgenogram) shows definite rib destruction, chest wall invasion can be assured (Fig 7). If the mass is merely contiguous with the pleura, invasion cannot be documented or excluded (Fig 8);36-311 however, this is fortunately not always a crucial distinction to make . If the patient has no metastases (MO) and neither the N2 nor N3lymph nodes are involved (Nl or NO), then, regardless of whether the primary is T3 or not, the appropriate therapy is surgery if the patient's physiologic status permits. The surgeon, aware of the possibility of a 1'3 primary tumor, simply plans the operation so that the tumor is not violated, and an en-bloc anatomic resection of chest wall can be performed, if necessary. There is an identical precaution for the use ofCT to judge mediastinal invasion. Definite interdigitation must be present on CT before a diagnosis of invasion can be made. All others need surgical evaluation . Computed tomography has been shown to unrellably predict the necessity of a lobectomy or pneumonectomy before surgery" as well as defining T3 central tumors in general. 4(1 In addition to the difficulty in 1436
establishing invasion, there is poor visualization of central submucosal and transfissural spread of tumor. The difficulty in evaluating central spread of tumor has been shown by Naidich et al." Computed tomography may detect very small pleural effusions not readily apparent on standard chest roentgenograms; however, these small effusions could be due to only a postobstructive pneumonia or lymphatic blockage in the hilum, rather than represent direct spread of the tumor. The realization that some pleural effusions, therefore , do not necessarily indicate unresectability has led the AJCC to recommend cytologic confirmation of a possible malignant effusion before the patient is deemed to have a T4, HIB, and, therefore, an un resectable lesion.
Magnetic Resonance Imaging Characterization of the airways by MRI is less
FIG URE 6. Mediastinal invasion . cr scan after intravenous bolu s of contrast material in a 56-year-old woman with oat-cell carcinoma. The left lower lobe mass has encased the left pulmonary artery, aorta, and left main bronchus.
Staging Patients with Non-small cell Lung cancer (UttJe, Stitik)
FIGURE 7. Definite chest wall invasion demonstrated hv L'T. This pain . CT 77-year-old man presented with severe chest and should~r shows an adenocarcinoma involving the posterior rib, chest wall and vertebral body.
accurate for tumor staging than CT, even though coronal reconstructions can show the entire airway on one image . Pleural effusions and rib destruction are better seen on CT than MRI; however, MRI holds promise to better delineate chest wall invasion (Fig 9). This is especially true in superior sulcus tumors, where increased contrast resolution and the ability to visualize the apex in coronal and sagittal planes may be very helpful. 14.42-44 Endoscopic Staging Bronchoscopy is important in patients with central tumors for staging as well as diagnostic reasons. When the tumor is in the main bronchus within 2 em of the carina, it is a T3 lesion. When it is within a lobar bronchus or a main bronchus more than 2 cm beyond the carina, it is a T2 primary. Bronchoscopic assessment of the main carina was previously considered important as a means of identifying metastasis to subcarinal lymph nodes . This involved identification of widening or splaying of the carina and direct biopsy or transbronchial needle aspiration; however, CT is now the standard for detection of mediastinal adenopathy and has made routine biopsy and aspiration unnecessary, except
FIGURE 8 . Indeterminate cr for medi astinal invasion. Biopsyproved adenocarcinoma is seen in contlgultv with the aortic arch and anterior chest wall. No definite invasion is demonstrated bv cr. Patients with these findings should undergo operative sta~itl~ :
FI
-
accp section report
~III.~_----------
Clinical Staging of Patients with Non-small Cell Lung Cancer* Alex G. Little, M.D., F.C.C.R;t and Frederick R Stitik, M.D.+
This article is the first in a planned series from the Section on Lung Cancer of the ACCP addressing the important and clinically relevant aspects of what is now the most common malignancy in the world, lung cancer. This initial report addresses the problem of staging of lung cancer. Staging, or identifying the anatomic extent of disease according to the AJCC TNM classification scheme, is the first clinical activity in caring for a patient with known or presumed
of this report is not to simply present T hethepurpose recently modified AJCC TNM staging definitions for lung cancer, which are amply described in the following tabulation and Table 1:1
lung cancer because the results determine appropriate types of therapy. This is, therefore, a critically important aspect of the patient's care which forms the foundation for subsequent treatment. In addition, consistent use of this system, based on appropriate clinical and pathologic staging, in stratifying patients in clinical reports is mandatory; otherwise, meaningful comparisons and conclusions are impossible. (Chest 1990; 97:1431-38)
T4
PRIMARY TUMOR (T) TX
TO TIS TI
T2
T3
Tumor proven by the presence of malignant cells in bronchopulmonary secretions but not visualized roentgenographically or bronchoscopically, or any tumor that cannot be assessed, as in a retreatment staging. No evidence of primary tumor. Carcinoma in situ. A tumor that is 3.0 cm or less in greatest dimension, surrounded by lung or visceral pleura, and without evidence of invasion proximal to a lobar bronchus at bronchoscopy, A tumor more than 3.0 ern in greatest dimension, or a tumor of any size that either invades the visceral pleura or has associated atelectasis or obstructive pneumonitis extending to the hilar region. At bronchoseopy the proximal extent of demonstrable tumor must be within a lobar hronchus or at least 2.0 cm distal to the carina. Any associated atelectasis or obstructive pneumonitis must involve less than an entire lung. A tumor of any size with direct extension into the chest wall (including superior sulcus tumors), diaphragm, or the mediastinal pleura or pericardium without involving the heart, great vessels, trachea, esophagus, or vertebral body, or a tumor in the main
*From the Department of Surgery, University of Nevada School of Medicine, Las Vegas; and the Department of Radiology; Eastern Virginia Medical Center, Norfolk, Va. tProfessor and Chairman, Department of Surgery. :l:Professor of Radiology, EVMS. Reprint requests: Dr. Little, 2040 West Charleston, Las ~gas, Nevada 89102
bronchus within 2 em of the carina without involving the carina. A tumor of any size with invasion of the mediastinum or involving heart, great vessels, trachea, esophagus, vertebral body, or carina or presence of malignant pleural effusion.
NODAL INVOLVEMENT (N) NO NI
N2 N3
No demonstrable metastasis to regional lymph nodes. Metastasis to lymph nodes in the peribronchial or the ipsilateral hilar region, or both, including direct extension. Metastasis to ipsilateral mediastinal lymph nodes and subcarinal lyrnph nodes. Metastasis to contralateral mediastinal lymph nodes. contralateral hilar lymph nodes, or ipsilateral or contralateral scalene or supraclavicular lymph nodes.
DISTANT METASTASIS MO M1
(~t)
No (known) distant metastasis. Distant metastasis present.
Our focus is on the appropriate clinical strategy which should be used when staging nonsmall-celllung cancer. Other articles will follow from the Section on Lung Cancer of the ACCP on the pathology, diagnosis and treatment of both small-cell and non-small cell lung cancer. Staging is the clinical strategy of determining and identifying the anatomic extent of disease in a patient with a malignancy. Stratification of patients by stage groupings is crucial; however, when reporting clinical experiences, stratification of patients by precise TNM classifications is essential because of the prognostic differences of patients within stage groupings. If this is not done, meaningful conclusions and comparisons CHEST I 97 I 6 I JUNE, 1990
1431
Table I-Stage Grouping Occult carcinoma Sta~t
,O
Stage I Sta~e
II
Sta~e
iliA
Stage I1IB Stage IV
TX TIS TI T2 TI T2 1'3 1'3 T1-3 AnyT T4 AnyT
NO MO Carcinoma in situ NO MO NO MO Nl MO NI MO NO MO Nl MO N2 MO N3 MO AnyN MO AnyN Ml
between reports are not possible . In addition, in the case of lung cancer, the stage of disease correlates very well with the patient's prognosis and, therefore, serves to guide development of therapy; however, translating the need for accurate staging into clinical practice poses difficulties. Should every possible test be performed in every patient? Are all tests unnecessary and simply a thorough history and physical examination sufficient? Our goal in this review is to present "state-of-the-art" answers to these questions. The aim of clinical staging is primarily to determine the AJeC stage . Although the TNM system is presented in the usual order of T~N~M, staging in clinical practice is usually done in the reverse order of M~N~T. If a metastasis is found, the patient has M1 disease which is classified as stage IV, and special efforts to stage the mediastinum or hilum (N) and primary tumor (T) are otiose, as results will not change the patient's stage . Similarly, if there are no metastases and N3 nodes are involved , then the classification of the primary tumor cannot change the stage, which is I1IB; however, there is heterogeneity even within
FIG UHE 2 . Contralateral pulmonary metastasis. PAche st rad iograph of a 72-year-old man with left upper lobe atelectasis due to squamous carcinoma and several large and small pulmonary parenchymal metastases in the right lung. Arrow indicates the largest of the metastases.
stage groupings, as the prognosis of all patients is not identical; for example, within stage IlIA, patients with T3NOMO disease differ in many ways from patients with TlN2MO extent of disease. This distinction between T and N classifications, therefore, mayor may not be important in determining care for a particular patient. If not important, the staging endeavors cease when the clinician believes that enough information is available to determine therapy.
Flt ; UHE 1. Metastatic disease demonstrated on conventional chest radiograph . (A, left). PA radiograph of a 66-year-old man with a 3 em left upper lobe adenocarcinoma and destructive lesion of the anterior right fourth rib . (B, right). Close-up of the de stroyed anterior rib . Compare the superior cortical margins of the rib immed iately above and below the involved rib (arrows) .
1432
Staging Patients with Non-small Cell Lung Cancer (UttJe. Stitik)
STAGING FOR METASTATIC DISEASE
The first major question which requires answering when a patient presents with known or presumed lung cancer is: does the patient have a distant organ metastasis? Obtaining a thorough history from the patient and combining it with a careful physical examination answers this question in many instances. 2.3 Symptoms, such as a new headache or bone pain, or 6ndings during the physical examination, such as an enlarged liver, direct the physicians attention to specific organs which should be evaluated with a focused examination such as a brain CT or a technetium-99m polyphosphate bone scan. Simple laboratory tests, such as a CBC, hepatic function tests, and serum calcium determination, may suggest the presence of distant metastases and should be pursued appropriately. The standard PA and lateral chest roentgenogram may also provide evidence of metastatic disease to the contralateral lung, pleura, or thoracic osseous structures (Fig 1 and 2). The more difficult issue is how to stage a patient who does not have symptoms, signs, physical findings, or laboratory values suggesting the presence of metastatic disease. Radionuclide Scanning
Two radionuclide scans can be useful in screening for clinically silent metastases. The first is the technetium-99m polyphosphate bone scan, which is sensitive for occult involvement of bone, a relatively common site of metastases. Routine bone scanning is often performed, even though the detection rate of truly occult metastases (without even minimal signs and symptoms or laboratory abnormalities) is low There is a false-positive rate, which means that roentgenograms of the involved bone or other examinations are required in some patients to disprove the scan. This rate is less than 20 percent in one series, and therefore it is argued that it does not unnecessarily prolong the clinical staging process. 4 In addition to the bone scan, the gallium-67 citrate scan is useful in detecting unsuspected and silent metastases but only in specific and, unfortunately, limited circumstances," The primary tumor must be 2 em or greater in size and avidly take up gallium. The test must be performed with a high dose of the radiopharmaceutical, and the gastrointestinal tract must be completely evacuated to eliminate it as a source of isotope retention; however, most importantly, the scan must be performed with a large crystal gamma camera which obtains tomographic images. These scanners are not available at many institutions, and without them, this examination is of limited usefulness; however, when they are available, results are useful. In one report, gallium scanning detected occult metastases to distant organs in 11 percent of
100 patients without the occurrence of any falsepositive scans.6 Computed Tomography
Computed tomography of the chest and upper abdomen, down to and including the liver and adrenal glands, should be considered a routine examination to be performed in all patients in whom no metastatic disease is identified by history or physical examination. In addition to information about the primary tumor and the lymph glands within the mediastinum, cr detects metastatic pulmonary nodules, small pleural effusions, metastatic deposits in the liver, and metastatic involvement of the adrenal glands. There are few data assessing the value of this scanning technique as a screen for metastatic disease in asymptomatic patients. One retrospective review of 85 patients found that CT confirmed the presence of metastases causing symptoms in five patients and detected asymptomatic metastases in seven patients (8 percent)." Computed tomography of the chest is recommended for assessment of the primary tumor and the regional lymph nodes in every patient in whom neither the history, physical examination, nor chest roentgenogram identify Ml disease. Extending the scan to include the liver and adrenal glands is a simple, quick, and easy means to increase the sensitivity for detection of metastases. Some physicians believe that CT of the brain should be used routinely, particularly in patients with proven or suspected adenocarcinoma, because of the predilection of these tumors for metastasis to this site. In one clinical study," albeit retrospective, 16 (18 percent) of 89 patients had abnormal CT scans of the brain; however, nine of the 16 had evidence of CNS disease by history or physical examination, and the other seven had strong clinical evidence of disseminated disease. In other words, no abnormal cr scans were found in patients with completely normal clinical examinations. Therefore, there are no data to support routine brain cr in patients in the absence of symptoms suggestive of brain involvement. Magnetic Resonance Imaging
Magnetic resonance imaging (MRI) has become an important method in the diagnosis of many diseases. The value of this relatively new technology in patients with lung cancer is constantly being defined. Magnetic resonance imaging is frequently compared to cr in assessing patients with lung cancer. In the thorax, MRI is generally considered equal to Cf,9-18 although in certain limited areas, as will be discussed in this article, MRI may be more useful. Currently, cr is the standard for patients with known or suspected lung cancer, with MRI playing a problem-solving role. The technology for CT scanning of the chest has CHEST I 97 I 6 I JUNE. 1990
1433
remained stable for a number of years; however, the technology for MRI is rapidly improving. In years to come, MRI may supplant CT of the chest in some or all of the current indications. In general, MRI has the following advantages over CT : (1) no ionizing radiation; (2) better contrast resolution; (3) no intravenous iodinated contrast material necessary; (4) the ability to image in many planes, ie, coronal and sagittal planes, without significant loss of spatial resolution; (5) less degradation by metallic clips than CT ; and (6) the potential for tissue characterization. On the other hand, CT (1) has better spatial resolution; (2) is quicker (30 minutes per examination for CT vs 60 minutes for MRI); (3) is less expensive; (4) is more readily available in many hospitals; (5) is degraded less by vascular and cardiac motion; and (6) has the ability to detect calcium, which MRI does not. Magnetic resonance imaging may be more sensitive in detecting brain metastases, although it has generally not replaced CT in the evaluation of patients with signs or symptoms suggesting CNS metastases. We currently employ MRI to detect possible brain metastases when the CT findings are equivocal or the patient is allergic to iodine. Magnetic resonance imaging is probably equal to CT in the detection of hepatic metastases, although some investigators believe that MRI may be more sensitive;'? however, this is generally a moot point, as the liver is studied at the time of chest CT. Magnetic resonance imaging of the liver could be used to solve problems unresolved by CT or in those patients who are allergic to iodinated contrast material. In the future, MRI may play an important role in evaluating patients who have lung cancer with suspected adrenal metastases. Asymptomatic adrenal masses, likely nonfunctioning cortical adenomas, are often present in the normal population, with this being an incidental finding on abdominal CT in approximately 1 percent of patients.P' These adrenal masses are well seen by CT but cannot be characterized as metastases or benign conditions. Early studies have suggested that MRI , with its better contrast resolution , may be able to separate benign from malignant disease ." Other investigators report a significant overlap in imaging characteristics." The investigational radiopharmaceutical, NP-59, may help further separate benign from malignant adrenal masses. 23
stage IV and relinquishing hopes for curative therapy; for example, in one study, evaluation with an open rib biopsy of the ambiguous finding of a solitary rib abnormality on a bone scan, using a guided operative technique, proved the absence of metastatic disease in 50 percent of the cases." Half of the patients would have been inappropriately considered to have M1 disease without this aggressive surgical approach oriented around requiring histologic documentation of the nature of the tissue when the results of noninvasive examination are ambiguous. Similarly, Oliver et al25 reported the findings in 330 patients with non-small cell lung cancer who underwent CT of the thorax, including the upper abdomen. In 25 patients an adrenal mass was discovered as the only possible evidence of metastatic disease . In these 25 patients, eight were proven to have true metastatic disease by CT-guided adrenal biopsy or follow-up, while the majority (17) were shown to have benign disease by CT-guided biopsy, surgery, or follow-up." STAGING OF REGIONAL LYMPH NODES
Computed Tomography As mentioned, chest CT should be obtained in all patients. When chest CT does not show any mediastinal lymph node larger than 1 em in diameter, the chance of nodal metastasis is low, less than 10 percent for the mediastinum as a whole, and invasive mediastinal staging is not required (Fig 3).26-2H When results are analyzed for different anatomic nodal basins within the mediastinum, regional differences in this negative predictive index are found . The lowest accuracy is in the region of the aortopulmonary window, where the subaortic nodes are located .2H Fortunately, these pa-
Invasive Staging When a noninvasive test or a combination of noninvasive tests unequivocally establishes the presence of Ml disease, then histologic documentation is not required; however, if an abnormal finding is not unequivocally believed to represent a metastasis, then tissue confirmation of the nature of the abnormality should be sought before consigning the patient to 1434
FIGURE 3. False-negative cr scan of a 64-year-old man with a 2 cm adenocarcinoma in the right upper lobe and multiple small pretracheal and aortopulmonary window lymph nodes, none measuring more than 1.5 cm . At medi astinoscopy; the small lymph nodes conta ined metastatic adenocarcinoma. Staging Patients with Non-small Cell Lung Cencer (Little. Slitik)
a single enlarged node." Calcium in enlarged nodes because of granulomatous disease will not he detected by MRI.
Invasive Staging
FIGURE 4. False-positive cr scan of a 70-year-old man with squamous carcinoma causing left upper lobe atelectasis and pneumonia. A large pre-aortic lymph node measuring 2 cm is present (armw) . At thoracotomy, the enlargement was due to reactive hypertrophy, likely secondary to the obstructive pneumonia.
tients have a better prognosis than patients with involvement of other mediastinal nodal basins, so surgical treatment is still reasonable and understaging in this region does not result in inappropriate therapy;29 however, when lymph nodes are enlarged greater than 1 cm in diameter, invasive staging with mediastinoscopy needs to be performed , as approximately 30 percent of these enlarged lymph nodes will be proven to be due to reaction and inflammation, rather than metastatic disease (Fig 4). Finally, it is important to note that the reliability of CT staging varies according to cell type, as in one study the accuracy rate for adenocarcinoma was 94 .7 percent, compared to 70.6 percent for squamous cell carcinoma.P' The previous considerations represent a rea sonable assessment of the value of CT in patients with suspected N2 or N3 disease . Many controversies exist, as reviewed by several authors.v-"
Magnetic Resonance Imaging Currently, the value of MRI in the detection of nodal metastases suffers from the same lack of sensitivity and specificity as CT scanning and for exactly the same reasons; both only characterize a node by its size and location, and neither is a reliable indicator of histology. It was hoped that with MRI's ability to characterize tissue, benign and malignant nodes could be separated . Several studies have shown a significant overlap in values from benign and malignant hilar and mediastinal nodes studied in vitro, as well as in vivo .9 • 1ll •32 .:J.1 Because of its ability to image in the coronal plane, MRI may he ahle to demonstrate the size of suhaortic nodes to a better extent than CT. I" Magnetic resonance imaging may he of value in the assessment of hilar nodes in patients allergic to iodinat ed contrast material. Because of the poor spatial resolution of MRI , occasionally small lymph nodes are not recognized as discrete nodes and are visualized as
The supraclavicular and scalene (N3) lymph nodes are accessible to manual palpation and should he carefully examined in this fashion . Although "blind" scalene node biopsy is inappropriate because of the low yield in the ahsence of palpable adenopathy, all distinctly abnormal nodes are excisionally biopsied for staging. This does not mean that every lymph node that can be palpated is removed, just those that are suspicious because of their size or consistency. Some surgeons continue to strongly defend the role of routine mediastinoscopy in all patients;:J.'J ·34 however, although mediastinoscopy, both cervical and parasternal approaches, is very safe, the yield is too low to justify its use when the identified mediastinal lymph nodes are less than 1 cm in diameter on CT. When larger mediastinal nodes are seen on CT, mediastinoscopy is appropriate, as the positive predictive accuracy of the scan's results is only approximately 70 percent. Therefore, if the finding of enlarged nodes is accepted as proof of involvement of N2 or N3 nodes, fully 30 percent of patients will be inappropriately staged . The main value of CT demonstrating enlarged nodes is that it can be used as a road map and help to select the best test for histologic evaluation. The choice of mediastinoscopy, median sternotomy, transbronchial needle aspiration biopsy, percutaneous needle aspiration biop sy, or thoracotomy can be facilitated by careful analysi s of the CT scan, local physician expertise, and a thorough knowledge of the new staging system , and, therefore, treatment ramifications should the enlarged nodes be positive. The location of enlarged nodes on CT, or positive nodes found at mediastinoscopy or thoracotomy, should be carefully recorded using the AJCC' or the ATS lymph node map ." A better understanding of the significance oflymph node metastasis may help define future staging systems and therefore modify treatment and prognosis. PRIMARY TUMOR
When the patient has neither chest wall nor pleuritic pain in the location of the primary tumor and both the chest roentgenogram and CT show that the tumor is neither at the limits of the visceral pleura nor a centrally located lesion, then the tumor is, by definition, either Tl or T2 . The distinction between Tl or T2 tumors does not affect the stage and is not crucial for making initial therapeutic decisions. Patients with localized chest wall pain in the region of their primary tumor can he assumed to have a T3 or T4 lesion which is transgressing and irritating the parietal pleura; CHEST I 97 I 6 I JUNE, 1990
1435
FIGURE 5. Superior vena caval obstruction. (A, left). PA chest radiograph of a 74-year-old white man presenting with superior vena caval obstruction. A small-cell undifferentiated carcinoma is seen in the nght paratracheal area with moderate displacement of the trachea to the left . (B, right). cr scan after bolus injection of contrast material demonstrating the large mass, compression and displacement of the azygos vein and superior vena cava (arrows), and mult iple collateral vein s in the chest wall.
however, when the primary tumor abuts the chest wall on the chest roentgenogram, but the patient has no local symptoms, then it is difficult to accurately identify the level of penetration of the tumor, ie, does it simply touch, or does it actually invade the parietal pleura?
Computed Tomography Chest CT may unambiguously identify invasion of the soft tissues or bony structures of the chest wall or mediastinum, thus identifying the primary tumor as T3 or T4, depending upon the involved structure and the depth of invasion (Fig 5 and 6). When CT (or chest roentgenogram) shows definite rib destruction, chest wall invasion can be assured (Fig 7). If the mass is merely contiguous with the pleura, invasion cannot be documented or excluded (Fig 8);36-311 however, this is fortunately not always a crucial distinction to make . If the patient has no metastases (MO) and neither the N2 nor N3lymph nodes are involved (Nl or NO), then, regardless of whether the primary is T3 or not, the appropriate therapy is surgery if the patient's physiologic status permits. The surgeon, aware of the possibility of a 1'3 primary tumor, simply plans the operation so that the tumor is not violated, and an en-bloc anatomic resection of chest wall can be performed, if necessary. There is an identical precaution for the use ofCT to judge mediastinal invasion. Definite interdigitation must be present on CT before a diagnosis of invasion can be made. All others need surgical evaluation . Computed tomography has been shown to unrellably predict the necessity of a lobectomy or pneumonectomy before surgery" as well as defining T3 central tumors in general. 4(1 In addition to the difficulty in 1436
establishing invasion, there is poor visualization of central submucosal and transfissural spread of tumor. The difficulty in evaluating central spread of tumor has been shown by Naidich et al." Computed tomography may detect very small pleural effusions not readily apparent on standard chest roentgenograms; however, these small effusions could be due to only a postobstructive pneumonia or lymphatic blockage in the hilum, rather than represent direct spread of the tumor. The realization that some pleural effusions, therefore , do not necessarily indicate unresectability has led the AJCC to recommend cytologic confirmation of a possible malignant effusion before the patient is deemed to have a T4, HIB, and, therefore, an un resectable lesion.
Magnetic Resonance Imaging Characterization of the airways by MRI is less
FIG URE 6. Mediastinal invasion . cr scan after intravenous bolu s of contrast material in a 56-year-old woman with oat-cell carcinoma. The left lower lobe mass has encased the left pulmonary artery, aorta, and left main bronchus.
Staging Patients with Non-small cell Lung cancer (UttJe, Stitik)
FIGURE 7. Definite chest wall invasion demonstrated hv L'T. This pain . CT 77-year-old man presented with severe chest and should~r shows an adenocarcinoma involving the posterior rib, chest wall and vertebral body.
accurate for tumor staging than CT, even though coronal reconstructions can show the entire airway on one image . Pleural effusions and rib destruction are better seen on CT than MRI; however, MRI holds promise to better delineate chest wall invasion (Fig 9). This is especially true in superior sulcus tumors, where increased contrast resolution and the ability to visualize the apex in coronal and sagittal planes may be very helpful. 14.42-44 Endoscopic Staging Bronchoscopy is important in patients with central tumors for staging as well as diagnostic reasons. When the tumor is in the main bronchus within 2 em of the carina, it is a T3 lesion. When it is within a lobar bronchus or a main bronchus more than 2 cm beyond the carina, it is a T2 primary. Bronchoscopic assessment of the main carina was previously considered important as a means of identifying metastasis to subcarinal lymph nodes . This involved identification of widening or splaying of the carina and direct biopsy or transbronchial needle aspiration; however, CT is now the standard for detection of mediastinal adenopathy and has made routine biopsy and aspiration unnecessary, except
FIGURE 8 . Indeterminate cr for medi astinal invasion. Biopsyproved adenocarcinoma is seen in contlgultv with the aortic arch and anterior chest wall. No definite invasion is demonstrated bv cr. Patients with these findings should undergo operative sta~itl~ :
FI
