Simon P.G. Padley, Olga Lazoura
Pulmonary Neoplasms
Bronchogenic Carcinoma
Bronchogenic carcinoma remains the most common cause of cancer death worldwide. The incidence of the disease has accelerated over the past century, closely in step with tobacco smoking, to peak towards the end of the last century. In the United States in 2011 there were almost 160,000 deaths from lung cancer, with a continuing decline in incidence in men of nearly 30% since a peak in 1990.1 The incidence of lung cancer in females is following a different trajectory, beginning to increase rapidly in the mid-1960s with only a minimal fall since the peak in 2000. Lung cancer is a disease of the elderly: 60% of cancers are diagnosed in people of more than 65 years of age and 70% of cancer deaths occur after 65 years of age.2
Histopathology
All types of lung cancer are related to cigarette smoking. The predominant cell types are small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). NSCLC is divided into three main subtypes, squamous cell carcinoma, adenocarcinoma and large cell cancer. The strongest link with cigarette smoking is seen with squamous cell carcinoma.3 Thirty years ago squamous cell carcinoma was much more common than adenocarcinoma but over the past three decades the frequency of adenocarcinoma has greatly increased and the ratio is now 1.4 to 1.4
Genetic Factors
There have been important advances in the understanding of the genetics of different forms of lung cancer. There are a number of specific genetic mutations that can now be routinely identified and which convey prognostic and treatment implications. The best known of these is the expression of epidermal growth factor receptor (EFGR) mutations in adenocarcinomas of non-smokers.
Although a major risk factor for lung cancer development is smoking, the disease is increasingly recognised in never smokers. These patients make up as much of 25% of the lung cancer patients in some populations.5 These patients are more commonly female with adenocarcinoma, and this type of disease is particularly common in Asian patients, with some studies reporting as many as 40% of patients to be non-smokers.6
Epidermal Growth Factor Receptor (EGFR)
Identification of mutations in oncogenes associated with non-squamous (NSCLC) can help guide targeted therapy. Currently the most important oncogene, which is now routinely sought with genetic profiling assays, is epidermal growth factor receptor (EGFR). This protein stimulates tyrosine kinase. EFGR overexpression and mutations in the tyrosine kinase domain of the EGFR gene can directly lead to tumour growth and progression; therefore, EGRF has itself become a target for chemotherapy and a number of specific agents have been developed that prevent activation, block the relevant signalling pathways and improve response rates to therapy. The relevant EGFR mutations, associated with sensitivity to tyrosine kinase inhibitors, are most commonly encountered in non-smoking females with adenocarcinoma. Therefore the detection of these mutations will predict a response rate of up to 70%, making targeted therapy prescription of these relatively costly treatments more cost-effective.7
K-Ras
This protein stimulates signalling pathways downstream from EGFR. Specific mutations lead to the production of activated K-Ras protein, which continues to stimulate tumour growth. Although tyrosine kinase may block EGFR activation, it does not block the activity of mutated K-Ras proteins. Therefore patients with specific K-Ras mutations will not respond to tyrosine kinase inhibitors. Identification of K-Ras mutations is more common in smoking patients with adenocarcinoma, usually Caucasian rather than Asian. K-Ras mutations confer a poor prognostic outlook.
ALK
Mutations within the anaplastic lymphoma kinase (ALK) gene are associated with NSCLC (usually adenocarcinoma). Patients with ALK specific rearrangements do not benefit from tyrosine kinase inhibitors but may respond favourably to other therapies such as crizotinib, the first approved ALK inhibitor.
There are a number of other mutations of potential importance in non-cell lung cancer. These can only be identified by molecular testing, which in turn may allow targeted therapy or prediction of resistance. EGRF, K-Ras and ALK mutations are usually mutually exclusive. Application of lung cancer mutation panel tests is therefore becoming increasingly commonplace, particularly as funding for this investigation becomes more widely approved.
Lung Cancer and Other Environmental Factors
Smoking
Smoking, by a large margin, is the major risk factor for lung cancer, and a dose relationship has been reconfirmed several times since the hallmark study of Doll and Hill.8 Cigarettes have changed in composition since the 1950s. Cigarette smoke is complex in nature and there are up to 60 identified carcinogens in tobacco smoke. Filters are now commonplace and nicotine levels have fallen in the tobacco varieties now produced. Nicotine itself is not thought to cause tumours, but seems to promote their growth. Since nicotine is the major dependent pharmacological agent in cigarettes, lowering levels may have resulted in a habit of greater depths of inhalation and in total numbers of cigarettes consumed.
Passive Smoking
Passive smoking has received considerable attention and is now recognised as a major contributor to worldwide morbidity and mortality related to lung cancer. A number of studies have demonstrated that non-smoking spouses of smokers have a 20–30% increase in lung cancer9 and a dose–response relationship has been demonstrated.10
Huge efforts have been made to reduce smoking rates, and clearly never commencing smoking is the aim. There are now almost as many former smokers as active smokers in the United States,11 and cessation of smoking reduces all lung cancer risk, especially those most strongly associated with smoking, namely small cell lung cancer and squamous cell carcinoma. It is estimated that the risk of lung cancer will have dropped by 50% 15 years after ceasing to smoke.12
General Environmental Pollutants
General environmental pollutants have been suggested as a further risk factor for lung cancer development. The influence of air pollutants has been long recognised as an environmental issue. More recently, attention has been paid to air quality and particularly to the concentrations on fine particles within the air that we breathe.13 Particles of less than 2.5 µm in diameter are strongly associated with lung cancer, especially in non-smokers. These particles are particularly associated with diesel engine exhaust.
Asbestos
Asbestos has also long been associated with increased lung cancer risk as well as being a known trigger for non-malignant lung and pleural disease. Chrysotile fibres are most closely linked with lung and pleural malignancy. The exposure to both asbestos and tobacco is particularly carcinogenic and estimates of a 15- to 50-fold increased risk of developing lung cancer are frequently quoted.14
Radon
Perhaps the forgotten aetiological agent is 222radon, second only to smoking as a cause of bronchogenic carcinoma. Since radon is in the earth’s crust, there is little that can be done to alter exposure levels. Certain geographical areas, for geological reasons, result in greater exposure levels.3 Data related to radon risk largely come from underground workers, where levels are high.
Lung Cancer Screening
Chest Radiographic Screening
There is a considerable literature on the use of chest radiography as a screening tool for early lung cancer. These trials have been of all varieties, including randomised controlled trials (RCTs). Some have been undertaken in conjunction with sputum cytology. Two early important trials were performed in Japan.15,16 Although designed differently and coming to slightly different conclusions in large numbers of patients, these mass screening programmes concluded there was a benefit associated with chest X-ray screening and sputum cytology compared to non-screening. These studies were both case-controlled studies.
The next landmark study was the Mayo Lung Project performed between 1971 and 1983, a randomised controlled trial on nearly 11,000 patients. Rather than a case-control study, this RCT also set out to determine whether chest X-ray and sputum cytology provided an effective means of screening for lung cancer. Perhaps unusually, patients were randomised into two different screening regimes, screened annually or every 4 months for 6 years, with further follow-up. This study demonstrated a clear stage shift in the more frequently screened patients, resulting in a better 5-year survival but at 20-year follow-up there was no overall survival benefit.17 Further large studies from the Memorial Sloane Kettering Hospital,18 the Johns Hopkins lung project19 and a Czechoslovakian study20 all had insufficient statistical power to demonstrate a reduction in mortality between screened and non-screened patients. Since the question remained open to doubt, the prostate, lung, colorectal and ovarian cancer screening trial, recruiting between 1993 and 2001 at 10 screening centres across the United States, attempted to resolve whether chest X-ray could be used as an effective screening technique for lung cancer and result in subsequent reduction in mortality.21 Unlike previous studies, many of the enrolees were never smokers (45%), or non-smokers (42%) and only about 10% were current smokers. Of the total 154,000 men and women enrolled, only 24,000 were considered to be of high risk for lung cancer. In this study, despite a slight stage shift in the screening group, there is no difference in lung cancer deaths between the screen group and the usual care group. From these various trials, providing some contradictory data, it can be reasonably concluded that chest X-ray alone has no useful role in lung cancer screening. Although lung cancers may be detected, often at a slightly earlier stage, the eventual outcome between screened and non-screened groups is almost identical.
CT Screening
More recently the debate has concentrated on lung cancer screening with CT, more recently using a low-dose technique. Again, it was the Japanese that led the way with two early trials.22,23 These studies set the pattern of subsequent CT screening projects. High-risk patients were identified, and a volumetric CT was undertaken. Detected nodules were followed up according to protocol. Very small nodules were not followed up, intermediate-sized nodules would continue along a screening pathway and subsequent CT at varying intervals, to detect growth. Larger nodules would be immediately sampled. Subsequent studies, including the landmark Early Lung Cancer Action Project (ELCAP)24 concluded that low-dose CT results in lung cancers being diagnosed at an earlier stage and with a higher cure rate than lung cancers detected as a result of symptoms. There has since followed at least 20 subsequent studies using CT as a screening tool.25 A full discussion of these trials is beyond the scope of this chapter. However, some have had a greater impact than others, most importantly the National Lung Cancer Screening Trial (NLST).26 This large and well-funded trial, by the US National Cancer Institute, enrolled 53,454 smokers or ex-smokers between the ages of 55 and 74, randomised to low-dose CT or PA chest radiography on an annual basis for 3 years. This design is slightly unusual, since there is no non-screening arm, but compares two different screening methods. There have been a number of important findings to come out of this large trial:
• More cancers were detected in the CT arm than the chest X-ray arm.
• Screen-detected cancers are at an earlier stage in the CT arm compared with the chest X-ray arm.
• Adenocarcinomas are more common in the screening population than in the symptomatic population.
• There is a 20% reduction in lung cancer specific mortality in the CT group.
This study, it is generally agreed, has demonstrated that CT screening applied to a carefully targeted group of patients can reduce lung cancer mortality. The cost of this approach looms large amongst the other questions that await a definitive answer. Cost-effective analyses are still underway.
This is not the end of the story—there are a number of ongoing European randomised controlled trials25 underway or in preparation from Italy, France, Holland, Denmark and the UK.
All of these trials have been designed slightly differently, but the question, as yet unanswered, is should this form of screening be rolled out across the general at-risk population. Amongst the many other questions that require consideration are the issues around the psychological burden of telling a patient they have a small nodule that requires follow-up to exclude lung cancer, with no answer likely for at least 2 years.
CT screening techniques are not perfect and the accuracy of CT as a detection technique has resulted in considerable debate. In the early days, 10-mm contiguous slices were utilised. Technology has since evolved, with multi-detector CT being routine and the ability to produce contiguous 1-mm collimation images now being commonplace. A simple, cheap but very effective technique, now widely employed on these narrow section data sets, has been the use of maximum intensity projection image reconstructions. This technique can be used on all CT workstations and most picture archiving and communication system (PACS) workstations and has been demonstrated to greatly improve conspicuity of nodules.
There has also been considerable resource expended on the development of computer-assisted diagnosis (CAD). Scrolling through many hundreds of images in an attempt to detect small nodules leads fairly rapidly to reader fatigue. CAD systems have been shown to augment the ability of a radiologist to detect all relevant lesions, by highlighting candidate lesions and allowing the radiologist to include or dismiss them as appropriate.27 The corollary of CAD utilisation is high initial false-positive rates: many lesions that are highlighted by the CAD system are subsequently dismissed by the radiologist. This may not result in a faster assessment, but does improve accuracy overall.28
In both the screening studies and in general practice, once a nodule has been detected, typically between 5 and 8 mm in size, the usual practice is to undertake follow-up CT studies and most radiologists will follow the Fleischner guidelines for the follow-up of lung nodules (Table 15-1).1
TABLE 15-1
Fleischner Society Guidelines for Nodule Follow-up
| Nodule Size (mm) | Low-Risk Patient | High-Risk Patient |
| <4 | No follow-up needed | Follow-up CT at 12 months; if unchanged, no further follow-up |
| >4–6 | Follow-up CT at 12 months; if unchanged, no further follow-up | Initial follow-up CT at 6–12 months, then at 18–24 months if no change |
| >6–8 | Initial follow-up CT at 6–12 months, then at 18–24 months if no change | Initial follow-up CT at 3–6 months, then at 9–12 months if no change |
| >8 | Follow-up at around 3, 9 and 24 months, dynamic contrast-enhanced CT, PET, biopsy | Same as for low-risk patients |
Adapted from the MacMahon et al.1
The vast majority of detected lung nodules, at least 98%, will be of no clinical significance, particularly in low-risk patients. Most of these nodules will not increase in size. However, being small, reproducible measurements are potentially problematic and therefore accurate determination of genuine increase becomes of critical importance in managing further follow-up intervals. Most radiologists routinely employ 2D calibre measurements, but for small nodules this is notoriously inaccurate, and is not reproducible either across different readers or between the same readers on different occasions, and the technique also does not lend itself to non-spherical lesions. Therefore, as well as being able to assist in the detection of nodules, computer-assisted characterisation tools have also become commonplace, through automatic segmentation and volume calculation. This technique, of producing a semi-automated nodule volume, is much more reproducible than 2D calibre techniques29 even though others have shown that the technique itself, when repeatedly measuring the same nodule, may give varying results.30 The practice of routine volumetric assessment is, however, beginning to become routine as CT workstation vendors provide the relevant software as a standard (Fig. 15-1).
Radiation Dose Considerations
It is important at all times, and particularly in a screening population where there is a high likelihood of detected nodules being of no significance, to reduce radiation exposure to a minimum. The use of a ‘low-dose’ CT technique should be automatic. This can be achieved by reducing tube current and tube voltage and increasing pitch. It is also important not to over-investigate screening or incidentally detected nodules, and to time follow-up studies appropriately. Furthermore, it is sometimes preferable to target the follow-up examination to only examine the nodule in question, rather than repeating the CT study of the entire thorax.
The Future of Screening
The NLST screening study has demonstrated a decrease in mortality from lung cancer in patients undergoing low-dose CT assessment. The resources required to roll out a CT screening programme are huge. The medical community is currently awaiting to see if the NLST trial results are confirmed by one or more of the other ongoing studies in Europe. None of these are individually as large as the NLST study, but in combination they might provide sufficient evidence to recommend screening across the general smoking population. What is not yet defined is how wide the screening net should be cast, whether the parameters used by the NLST are appropriate, and which nodules require further follow-up and further investigation. Currently, both in Europe and the United States, it is agreed that the patients undergoing screening CT should do so according to agreed guidelines. These guidelines also advise in detail about the management of screening results. The American Lung Association has also recently issued guidance on lung cancer screening to patients and physicians.7,31,32 In the UK a recent opinion piece from the UK Lung Screening (UKLS) group neatly defines the problems that remain unresolved as psychosocial and cost-effectiveness issues, harmonisation of CT acquisition techniques, management of findings, screening frequency and subject selection.33
Pulmonary Nodules
Management of Small Pulmonary Nodules
Nodule detection, now an everyday occurrence in patients undergoing multi-slice CT, raises a series of management problems for the referring physician and reporting radiologist. The wildly utilised Fleischner Society Recommendations (Table 15-1) have provided clear and frequently utilised guidance since their publication in 2005.1
In evaluating a pulmonary nodule it is helpful to bear in mind likely causes. Assessment of nodular morphology and characteristics can also provide useful information.
Nodule Size
Small nodules are very unlikely to be due to malignancy. Indeed screening studies have demonstrated that malignancy in nodules of less than 5 mm in size (4 mm or less) is so low, and these nodules are so common, that follow-up is not generally recommended. However this does not hold true in patients with a known primary malignancy elsewhere. Most benign nodules measure less than 2 cm in diameter and the smaller the nodule is, the more likely it is to be benign.34–36
Location, Shape and Morphology
Perifissural nodules are a recognised entity following CT screening studies. These small subpleural nodules have been shown to frequently represent intrapulmonary lymphoid tissue or complete intraparenchymal lymph nodes.37 Characteristic features of intraparenchymal lymph nodes (Fig. 15-2)38 are nodules less than 15 mm from the pleural surface, being ellipsoid in shape and usually being connected to the pleural surface by a fine linear opacity.39 Follow-up studies of nodules of this variety, detected during the Nelson Screening Trial demonstrated that no nodules with these features developed into lung cancer.40 Nodule outline can also be helpful when other features typical of intrapulmonary lymphoid material are absent. Concave surfaces on all sides or a straight surface of contact with the pleura has also been shown to represent benign features.41 The less spherical a nodule is, particularly on volumetric assessment, the less likely a malignant aetiology. Flat or tubular nodules are more likely to be benign than round nodules. Therefore, solid, subpleural, polygonal nodules with a low sphericity index are highly unlikely to be malignant.
Cavitation within a nodule can occur and be both benign and malignant in aetiology. Malignant cavitation is often associated with a thick and irregular internal cavity wall, compared to the more uniform cavitation associated with benign nodules, although this is not a reliable distinguishing feature.42
Nodule Contour
Nodules without obvious benign morphology may be smoothly marginated, lobulated or spiculated. Smoothly marginated nodules are more likely to be benign or metastatic. Lobulated nodules are more likely to be malignant43 but there is considerable overlap (Fig. 15-3). Therefore the presence of smooth borders is of little practical value. In distinction, spiculation is predictive of a malignant aetiology.44
The presence of central air bronchograms or soap bubble lucency centrally within a nodule has been previously evaluated. Multiple spherical areas of air may be present in adenocarcinoma, due to the lepidic growth pattern of these lesions, where tumour cells have grown along the alveolar walls and adjacent airways without filling the alveolar spaces. In comparison, the presence of air bronchograms,45 rather than bubble-like lucencies, may also be seen in lymphoma, organising pneumonia and alveolar sarcoidosis.
Nodule Density
Certain patterns of calcification are recognised as being highly predictive of a benign aetiology. Recognised benign patterns are lamellated, solid, central and popcorn-like. Central or lamellated calcification is typically indicative of previous granulomatous disease44 and, similarly, dense solid calcification is likely to indicate previous granulomatous infection (Fig. 15-4). Popcorn calcification usually indicates the presence of a hamartoma46 and these lesions may also contain convincing evidence of internal fat density. When fat is present in a lesion of less than 2.5 cm in diameter, then, particularly if the lesion is PET negative, further evaluation is not required, but most hamartomas do not demonstrate this helpful characteristic. Calcification, which is eccentric or stippled within an area of soft tissue density, may be seen in malignancy. Very occasionally metastases from bone-forming or cartilage-forming tumours may present a benign pattern of calcification, but usually there is a relevant history.
Ground-Glass Nodules
A focal area of increased lung attenuation, which may be well or poorly defined but through which normal structures can still be discerned, is typically referred to as a ground-glass density. If localised, the opacity may be described as a ground-glass opacity (GGO) or ground-glass nodule (GGN) (Fig. 15-5). The term ‘pure ground glass nodule’ is the preferred descriptive term if there is no soft-tissue component.46 If an area of ground-glass density includes a solid component which does obscure lung architecture, this may be termed a part-solid ground-glass nodule. In the literature both pure ground-glass nodules and part-solid ground-glass nodules may be grouped together under the term sub-solid nodules. The Fleischner Society have now published guidelines on the management of these sub-solid nodules, recognising that such sub-solid nodules may represent early forms of adenocarcinoma.47 Because of the relevance of the current classification of lung adenocarcinoma to the management of sub-solid nodules the new classification will be considered here.48 The new classification eliminates the term bronchoalveolar carcinoma and mixed subtype adenocarcinoma and now divides adenocarcinoma into the following categories:
Prognosis of patients with adenocarcinoma in situ or minimally invasive adenocarcinoma (characterised on CT as pure ground-glass lesions) is excellent; these patients should have almost 100% disease-free survival.48 Invasive adenocarcinoma has a variable outlook, and to some extent this depends on the histological subtype. Detailed discussion is beyond the scope of this chapter.
The new recommendations for the management of sub-solid pulmonary nodules detected by CT from the Fleischner Society, reflecting the new classification of adenocarcinoma, are given in Table 15-2.
TABLE 15-2
Recommendations for the Management of Sub-Solid Pulmonary Nodules Detected at CT:
a Statement from the Fleishner Society
| Nodule Type | Management Recommendations | Additional Remarks |
| <5 mm | No CT follow-up required | Obtain contiguous 1-mm-thick sections to confirm that nodule is truly a pure GGN |
| >5 mm | Initial follow-up CT to confirm persistence; then annual surveillance CT for a minimum of 3 years | FDG-PET is of limited value, potentially misleading and therefore not recommended |
| Solitary part-solid nodules | Initial follow-up CT at 3 months to confirm persistence. If persistent and solid component <5 mm, then yearly surveillance CT for a minimum of 3 years. If persistent and solid component >5 mm, then biopsy or surgical resection | Consider PET/CT for part solid nodules >10 mm. |
| Multiple sub-solid nodules Pure GGNs <5 mm | Obtain follow-up CT at 2 and 4 years | Consider alternative causes for multiple GGNs <5 mm |
| Pure GGNs >5 mm without a dominant lesion(s) | Initial follow-up CT at 3 months to confirm persistence and then annual surveillance CT for a minimum of 3 years | FDG-PET is of limited value, potentially misleading and therefore not recommended |
| Dominant nodules with part-solid or solid component | Initial follow-up at 3 months to confirm persistence. If persistent, biopsy or surgical resection is recommended, especially for lesions with >5 mm solid component | Consider lung-sparing surgery for patients with dominant lesion(s) for lung cancer |
Note: These guidelines assume meticulous evaluation, optimally and contiguous thin sections (1 mm) reconstructed with narrow and wide and/or lung windows to evaluate the non-solid component of nodules. If indicated, when electronic calipers are used for bi-dimensional measurements, both the solid and ground-glass components of lesions should be obtained as necessary. The use of a consistent low-dose technique is recommended, especially in cases for which prolonged follow-up is recommended, particularly in younger patients. With serial scans, always compare with the original baseline study to detect subtle indolent growth.
Reproduced with permission from Naidich et al.47
In essence there are six current recommendations regarding sub-solid nodules:
• Solitary pure ground-glass nodules measuring 5 mm or less do not require follow-up surveillance.
• Solitary part-solid ground-glass nodules, with a solid component of more than 5 mm, should be considered malignant until proven otherwise if there has been growth or no change at the 3-month follow-up study (Fig. 15-6).
• In cases in which pure multiple ground-glass nodules are identified (Fig. 15-7), at least one of which is larger than 5 mm, and in the absence of a dominant lesion, an initial follow-up CT examination in 3 months is recommended followed by yearly surveillance CT examinations for at least 3 years.
Other Forms of Nodule Assessment
Nodule Follow-Up
Since repeat CT is a common recommendation in the management of both solid and sub-solid pulmonary nodules, it is important that subsequent CT examinations are undertaken using an identical protocol in order to accurately estimate doubling times. This is particularly so in cases of ground-glass nodules where very long doubling times are recognised in low-grade malignancies.49
Nodule Enhancement
Contrast-enhanced CT has been demonstrated as an effective management tool in the assessment of lung nodules. In essence, malignant nodules will demonstrate enhancement, as will some benign nodules. If there is no enhancement, malignancy is effectively excluded.50 This technique works well in soft-tissue density nodules, but is less applicable to nodules containing calcification, cavitation or ground-glass opacity. The possibility of generating a virtual non-contrast data set, using dual-energy CT, has been investigated as a means to reduce radiation exposure yet provide similar information to a pre- and post-contrast acquisition.51 Although nodule enhancement has been demonstrated to be practical, in practice it has largely been superseded by PET/CT evaluation.
PET/CT
PET/CT is now an essential tool in the management and the work-up of patients with possible pulmonary malignancy. It forms part of the standard staging in patients with proven or suspected lung cancer but is also a frequently utilised tool in the work-up of patients with an indeterminate lung nodule. The commonest isotope utilised is 2-deoxy-2-[18F]fluoro-d-glucose, a glucose analogue, with the positron-emitting radioactive isotope fluorine-18 substituted for the normal hydroxyl group at the 2′ position in the glucose molecule, commonly referred to as FDG. The technique relies on increased uptake in neoplastic nodules (Fig. 15-8). Increased uptake also occurs within many inflammatory processes. Nevertheless, the utility of PET/CT has been documented, in a number of studies,52 to have sensitivities of 90% and specificities of 83% for a diagnosis of malignancy. However, in the context of small nodules, FDG results must be interpreted with some caution, since nodules of less than 1 cm are more likely to result in a false-negative interpretation, particularly with certain lower-grade adenocarcinoma subtypes53 and carcinoid tumours. The application of PET/CT for nodules of less than 6 mm is currently not justified. In the context of lung disease, sarcoidosis, granulomatous infection and a number of inflammatory processes are recognised as resulting in significant PET FDG avidity.
Tissue Sampling
Tissue sampling of nodules which are both accessible and of a size, shape and morphology to suggest the possibility of malignancy, may be undertaken transbronchially, surgically or percutanously with radiological guidance. Central lesions may be amenable to bronchoscopic biopsy but even perihilar nodules, in the absence of an endoluminal component, remain very challenging for bronchoscopic diagnosis.54 Bronchoscopic biopsy of peripheral lung cancer using conventional techniques is also highly problematic.55 Percutaneous fine needle or cutting needle biopsy has been demonstrated as an accurate technique for the identification of malignancy, but not all nodules are suitable for this approach. When a final diagnosis of lung cancer is eventually established, a previous aspiration biopsy has a 90% likelihood of providing confirmation of a malignant diagnosis. There is a very low false-positive rate but there is a recognised and troubling false-negative rate. Transthoracic needle biopsy remains an essential tool in the management part of indeterminate nodules. Multidisciplinary team discussion of the relative merits of a follow-up strategy, further imaging, percutaneous or bronchoscopic diagnostic procedures or surgical resection for an individual lung nodule remains the ideal management step in this common problem.
Decisions regarding further investigation of pulmonary nodules can also be helpfully guided by the pre-test probability of malignancy: most importantly, previous known malignancy, patient age and smoking history. The use of risk management modelling has been previously studied,56,57 incorporating various risk factors, but all methods reported to date confirm that a significant current or previous smoking history in an elderly patient with a nodule diameter of more than 1 cm are all highly suggestive of a malignant aetiology. It should be remembered that 99% of all nodules of 4 mm or less in diameter will turn out to be benign on follow-up or resection. These lesions are so frequently seen on CT that follow-up is no longer recommended in the low-risk patient and a single follow-up CT is required in the high-risk patient. Between 4 and 8 mm in size, surveillance CT is the recommended approach, with the goal of demonstrating stability over a 24-month period. The frequency of interval CT will depend on the initial size of the lesion and the patient’s malignant risk (see Table 15-1). A more aggressive approach, including contrast enhancement CT, PET/CT, needle biopsy or resection is adopted for lesions of more than 8 mm.
Lung Cancer Staging—the 7Th Edition of the TNM Staging System for Lung Cancer
The new 7th edition of the TNM staging system is based on a much larger database of patients than previously available, comprised of over 100,000 cases from 45 centres in 20 countries of which more than 80,000 had sufficient information to contribute to the analysis. The 7th edition was published in 2009.58 Radiologists are divided as to whether it is desirable to give a TNM description in radiology staging reports, but if this information is provided then an intimate understanding of the TNM staging system is required. TNM staging is now an essential part of the multidisciplinary team decision process. The changes to the previous staging system are highlighted below.
The T descriptor not only describes the absolute size of the tumour but also assesses local invasion. The criteria for local invasion were not revised due to insufficient patient numbers and lack of validation. The main discrimination to be made is between T3 and T4 disease. Chest wall invasion, parietal pleural invasion, mediastinal pleural invasion and parietal pericardial invasion all remain T3 descriptors. Visceral pleural invasion remains a T2 descriptor. Even a tumour of less than 3 cm with parietal pleural invasion is described as a T3 lesion and this is important since these patients will receive adjuvant chemotherapy following resection.59 Lesions that invade across a fissure into an adjacent lobe have remained as T2 lesions.60 Assessment of pleural and chest wall invasion has long been recognised as a difficult task for CT analysis. Where there is obvious soft-tissue extension into the intercostal muscles or bone destruction, the issue is easily resolved. More subtle chest wall parietal pleural invasion is more difficult to define and, as in the past, local chest wall pain is known to be more specific than CT findings for chest wall and parietal pleural involvement.61 Efforts to more clearly define the presence or absence of pleural invasion have included MRI imaging, high-resolution targeted ultrasound and even diagnostic artificial pneumothorax.62 Invasion of the diaphragm is classified in the same manner as parietal pleural invasion and is a T3 descriptor.
Direct extension into the heart, great vessels, trachea, oesophagus, vertebral body and mediastinal fat are all T4 descriptors. Phrenic nerve involvement is a T3 descriptor but recurrent laryngeal nerve involvement, indicating direct mediastinal infiltration, is a T4 descriptor. Pancoast’s (superior sulcus) tumours are individually staged according to the involved tissues. For example an apical tumour with parietal pleural involvement is defined as a T3 lesion, but a similar tumour extending into a vertebral body or involving subclavian vessels becomes a T4 lesion. Because of the importance of this differentiation, the utility of multiplanar reformatted images on CT and targeted MRI examination has been highlighted. As in the 6th edition, bronchogenic lesions extending into a mainstem bronchus, but more than 2 cm from the carina, remain T2 lesions, less than 2 cm from the carina T3 lesions and invading the carina T4 lesions. An airway lesion causing peripheral atelectasis is now classified as T2a or b (depending on the size of the primary lesion) if less than the whole lung is involved and T3 if the whole lung is involved.
Additional Pulmonary Nodules in the Presence of Lung Cancer
Previously, a nodule in the same lobe as the primary tumour confirmed T4 status and if in a different lobe M1 status. This has been revised in the 7th edition (see Table 15-3). Improved outcomes for patients in this situation led to reclassification of nodules in the same lobe as the primary tumour as a T3 status, and for similar reasons a nodule in the ipsilateral lung but in a different lobe, whilst predicting a poor 5-year survival, confirms a slightly better outlook than M1 disease of other types. Therefore the combination of a primary lesion with a further nodule within an ipsilateral different lobe is now described as T4 disease in the 7th edition.
N Descriptors
Nodes are described as either N0 (no involvement), N1 (nodes up to and including hilar stations), N2 (ipsilateral mediastinal nodes) or N3 (contralateral mediastinal or more distant nodes). Before the 7th edition, there were two different nodal maps in common usage. These were largely similar but there were some important discrepancies, particularly with regard to subcarinal lymph nodes. These discrepancies have been resolved with the production of a new nodal map with anatomically defined borders. Relevant changes are supraclavicular and sternal notch nodes being designated station 1 and the shift of the midline for right and left level 2 and level 4 nodes from the midline of the trachea to the left lateral border of the trachea. Therefore a lymph node lying directly anterior to the trachea using the 7th edition nodal map would be designated as a right paratracheal lymph node.
This work has also led to the proposal of lymph node zones (Table 15-4), not incorporated into the current staging nomenclature but likely to be used in future iterations.63 Potential advantages are to facilitate description of large nodal masses crossing nodal stations. This nodal zonal concept also allowed an analysis to be undertaken of the significance of skip metastases where an N2 node is involved without evidence of N1 disease. This nuance has revealed that the presence of station 5 (aortopulmonary) nodes in association with a left upper lobe tumour but without N1 disease confers a better outcome status than other forms of N2 disease. The same was not found for right paratracheal (N2) nodes with right upper lobe tumours. The zonal mapping system also demonstrated better outcomes with patients of single zone N1 disease compared to multiple zone N1 disease and, similarly, single zone N2 disease confers a better 5-year survival than multiple zone N2 disease. The outcome of this proposal is the definition of three distinct groups of patients with progressively worse outcomes: namely, single zone N1 disease, multiple-zone N1 or single-zone N2 disease and multiple-zone N2 disease. These changes have not been incorporated into the 7th edition.
