Introduction
The lung is a common site for metastasis of malignant tumors from other organs. Logic would suggest that the lung is such a common destination for metastasis because it alone receives the entire cardiac output. Indeed, this likely does play a role in the eventual seeding of the lungs when cancers metastasize via the hematogenous route. However, a more detailed understanding of the molecular mechanisms dictating organ-specific metastases has made it clear that specific properties of the primary tumor and the microenvironment of the tissue of origin and of the lung play a significant role in directing the process of metastasis to the lung. Paget initially proposed what became known as the “seed and soil” hypothesis in 1889 based upon his observations of the nonrandom nature of breast cancer metastasis. While we can now begin to define mechanisms of organ-specific metastases at a molecular level, we still lack therapeutic tools specifically aimed at preventing metastases. This chapter reviews the epidemiology of lung metastasis, advances in our knowledge of the pathophysiology of metastasis, approaches to help distinguish metastasis from primary lung cancer, advances in diagnostic methods, and treatment options that vary from palliation to curative surgery.
Epidemiology
Estimates of the incidence of lung metastasis among patients with cancer vary from 20% to 40%. However, estimates of the incidence of lung metastasis have limitations. Most series focus on a single tumor type as the origin of lung metastases. The incidence of lung metastasis also varies depending on the means used for detection and for follow-up; a lower incidence is reported if the metastases are detected by the presence of respiratory symptoms and rises progressively when metastases are detected by routine surveillance with chest radiographs or with computed tomography (CT) scans, or autopsy data. Not surprisingly, the detection of metastases for any given solid tumor type will be greater when F-fluorodeoxyglucose positron emission tomography (PET) scanning is used as a method of surveillance. While reports of the proportion of cancers that metastasize to the lung vary, the lung is always among the most common sites of metastasis of solid tumors.
Clinical History
Lung nodules or effusions developing in the context of an extrapulmonary neoplasm can either be synchronous (discovered at the same time as the primary tumor) or metachronous (discovered at some period after the initial solid tumor, either incidentally in the course of follow up of a prior malignancy or in the context of new pulmonary symptoms). Because of the large degree of reserve pulmonary function in most individuals, metastases to the lungs rarely produce symptoms, and even patients with a large metastatic tumor burden can present with minimal or no pulmonary symptoms. When nonspecific symptoms of cough, dyspnea, or chest pain and discomfort can be attributed to metastatic lesions, they usually result from either a very large tumor burden ( Fig. 55-1 ), extensive lymphatic infiltration, airway involvement, or a large pleural effusion. The first clue to the presence of lung metastasis comes most commonly from surveillance radiographs in patients with known prior cancer. The most common presentation of lung metastasis is the finding of multiple nodules in the lower lobes. The finding of an incidental solitary pulmonary nodule as the first sign of an extrapulmonary neoplasm makes the distinction of metastasis from that of a primary lung tumor challenging. The physical examination is unlikely to disclose evidence of lung metastasis but, in the context of a solitary pulmonary nodule, the examination should include a search for malignancy elsewhere in the body (e.g., breast or abdominal masses, enlarged lymph nodes). Rarely, a metastasis can affect the larger central airways, producing localized wheezing, which an astute clinician discovers on auscultation. Malignant pleural fluid accumulation produces the physical findings of reduced breath sounds, a dull percussion note, and decreased tactile fremitus.
Mechanism of Metastasis to the Lung
When one considers that the lung receives the entire cardiac output, it is perhaps surprising that lung metastases are not even more common among patients with cancer. However, research has shown that the majority of cancer cells that circulate in the blood never result in a clinically overt metastatic focus, suggesting that mere access to the lung is not sufficient to lead to clinically evident metastasis. Additional mechanisms must therefore account for the propensity of certain tumors to seed distant organs preferentially.
One of the main processes involves expression of receptors that bind to ligands in metastatic sites. Chemokines are a group of small cytokines originally discovered as mediators of leukocyte trafficking. Chemokines mediate several important pathologic aspects of tumor biology, including angiogenesis, cell proliferation, invasion, and metastasis. The pattern of constitutive expression of chemokine receptors on cancer cells combined with expression of the corresponding chemokine ligands in various organs may be responsible for the organ-specific pattern of metastasis of a variety of solid tumor types, including breast, prostate, and lung cancer. For example, the chemokine receptors CXCR4 and CCR7 are highly expressed in human breast cancer cell lines, in primary breast tumor samples, and in their metastases. The corresponding ligands for the CXCR4 and CCR7 receptors, CXCL12 and CCL21, are constitutively expressed in organs to which these tumors commonly metastasize (lung, brain, bone, and lymph nodes). In breast cancer cell lines, signaling through the receptors CXCR4 or CCR7 promotes in vitro migration and invasion. Most interestingly, antibodies or small molecule antagonists directed against CXCR4 significantly impaired metastasis of breast cancer cells to regional lymph nodes and lung in vivo. In addition, malignant melanoma, which has a similar metastatic pattern as breast cancer but also a high incidence of skin metastases, showed high expression of the chemokine receptor CCR10 in addition to CXCR4 and CCR7. The ligand for CCR10 (cutaneous T-cell attracting chemokine, or CTACK/CCL27) is highly expressed in normal dermis. In a study of metastasis of non–small cell lung cancer, metastatic cells were found to be enriched for the expression of CXCR4 compared to cells from the primary tumors, suggesting that CXCR4-expressing cells had an advantage in reaching or surviving in the metastatic niche. Chemokine receptors on cancer cells, in concert with tissue-specific expression of their chemokine ligands, appear to have a critical role in determining the metastatic destination of circulating tumor cells.
Certain molecular processes may increase lung vascular permeability to metastatic cells. In an elegant study, transforming growth factor-β was shown to prime breast cancer cells for lung-specific metastasis by up-regulation of the gene angiopoietin-like 4 ( ANGPTL4 ). Transforming growth factor-β induction of ANGPTL4 in breast cancer cells enhanced their subsequent retention in the lungs but not in bone. Tumor cell–derived ANGPTL4 disrupted the pulmonary microvascular cell-cell junctions, increasing the permeability of lung capillaries and facilitating the passage of tumor cells into the lung parenchyma. In contrast, in the bone marrow sinusoids, which are normally more leaky microvascular beds, this mechanism does not confer an advantage.
Other processes may help support the metastatic cell in its lung niche. In another elegant mouse study, tumor cells were shown to be preceded into the metastatic niche by bone marrow derived cells (BMDCs) that express the type 1 vascular endothelial cell growth factor receptor (VEGFR1+ BMDC). These authors demonstrated that VEGFR1+ BMDCs were hematopoietic progenitor cells that (1) arrived at the metastatic site before tumor cells, (2) accumulated in the target organ of metastasis following either injection of tumor cells or treatment with tumor cell conditioned media, and (3) established a premetastatic niche in a tumor-specific fashion (i.e., they established a permissive environment for metastasis in an organ-specific pattern that varied depending on the tumor type).
Finally, other processes may enhance lung-specific adherence. As illustrated by Brown and Ruoslahti, the protein, metadherin, was found by phage display to be present on the surface of metastatic breast cancer cells that preferentially allowed adherence to the pulmonary vascular endothelium. In summary, tissue-specific metastases arise in a nonrandom fashion determined by specific molecular mechanisms that involve both tumor-derived soluble factors and microenvironment-specific ligand-receptor interactions.
A further refinement in our understanding of the mechanisms of metastasis comes from a study that examined isolated efferent blood from primary tumors in mice and found that primary tumors also shed tumor fragments consisting of both malignant cells and host stromal fibroblasts. Importantly, the viability of circulating malignant cells was nearly twofold higher in these heterotypic tumor fragments than in single tumor cells shed from the primary tumor. These chaperone stromal fibroblasts survived, proliferated, and conferred a survival advantage to lung metastatic deposits. When the investigators selectively depleted carcinoma-associated fibroblasts from mice after primary tumor removal, the growth of metastatic deposits was significantly inhibited. The study further demonstrated that human brain metastases, but not primary brain tumors, contained carcinoma-associated fibroblasts, supporting the hypothesis that the fibroblasts enhanced the metastatic malignant process. This elegantly designed and executed series of experiments, which required modern molecular techniques and lineage tracking of cells, brought the field back full circle to Paget’s seed-and-soil hypothesis because, as the authors correctly point out, the tumor cells facilitate lung metastasis by “bringing their own soil.” While these molecular insights have not yet led to specific interventions to prevent metastasis, such advances in the knowledge of molecular mechanisms that promote or facilitate organ-specific metastasis can be expected to drive the search for new treatments to prevent cancer metastasis. Metastases are nearly universally responsible for cancer mortality and, if metastases can be prevented or treated by novel strategies based on these molecular insights, survival with cancer (as opposed to a cure for cancer) is an attainable goal. A diagram illustrating potential mechanisms behind tissue specific metastasis to the lung from extrapulmonary malignancies is shown in Figure 55-2 .
Diagnosis
Differential Diagnosis
Given the range of diagnostic and therapeutic options in cases of known or suspected lung metastasis, patient care is optimal when delivered within the context of a multidisciplinary team. For example, when deciding whether to biopsy a lung nodule in a patient with a prior cancer, a chest physician is well served by speaking to the oncologist and/or the surgeon to learn whether a patient with metastatic cancer has therapeutic options and whether the results of a biopsy would alter those options. If a patient with a solitary nodule suspected to be an isolated lung metastasis is a surgical candidate and a complete resection is feasible, a preoperative tissue diagnosis may not change the treatment plan and, if so, should not be undertaken. Conversely, surgeons may prefer a preoperative tissue diagnosis to improve their discussion of risks, benefits, and alternatives with the patient. Surgeons may also prefer knowing whether a nodule is a metastasis or a primary lung cancer in planning the appropriate surgical procedure, either a lobectomy or an anatomic segmentectomy. For these and other reasons, the management of these patients is best served when clinicians are readily able to communicate patient-specific concerns to one another, such as in a multidisciplinary conference.
In general, the patient with a prior history of cancer and suspicion of a lung metastasis should be managed with knowledge of several interacting factors: the type of prior cancer, its natural history, its propensity to cause lung metastasis, its sensitivity to systemic or radiation therapy, and the number and location of lung nodules. A decision to biopsy should take into account the likelihood of an alternative diagnosis and whether the results of the biopsy will change the management options. The diagnostic possibilities for a lung nodule or multiple nodules in the patient with a history of cancer vary with the type of prior cancer. For example, the implications of finding a cavitary nodule in a patient with prior leukemia after a bone marrow transplant are quite different from finding the same cavitary nodule in an individual with prior head and neck cancer. In general, the differential diagnosis of a lung nodule in a patient with prior cancer includes the same diagnoses found in the general population, with the obvious exception that metastatic disease must be considered as well. Both primary malignant and benign causes, including infectious and noninfectious etiologies, should be considered ( Table 55-1 ).
| MALIGNANT |
|
| BENIGN |
| Infectious |
|
| Noninfectious |
|
Distinguishing Metastasis From Primary Lung Tumors
When a nodule or nodules are discovered in a patient with a prior history of cancer, it is important to distinguish lung metastasis from a primary lung cancer because the treatment can be very different. Clues such as the type of prior cancer, radiologic characteristics of the nodule, and pathologic appearance of subsequent biopsy material can be used to guide treatment decisions. Multiple pulmonary nodules with smooth borders located in the lower (dependent) lobes are most likely to be metastatic based on clinical grounds alone. When a single nodule is the only evidence of metastasis, recognizing it as metastatic can be more difficult. In some instances, the type of prior cancer can provide a clue to the likelihood of a metastasis versus a primary lung cancer. In a retrospective study of patients with single pulmonary nodules and a prior history of cancer, multiple factors were examined as possible predictors for metastasis including the histologic characteristics of the original extrapulmonary neoplasm, patient age, and smoking history. Of 161 patients with solitary pulmonary nodules, 81 (50%) were determined to have a primary lung cancer; this outcome was more common in patients with a prior history of cancer of the head and neck, bladder, breast, cervix, bile ducts, esophagus, ovary, prostate, or stomach. Fifty patients (31%) were determined to have metastasis, a finding more common in patients with salivary, thyroid, and adrenal tumors, melanoma, and sarcoma. Primary lung cancer or metastases were equally likely in patients with prior cancer of the colon, kidney, or uterus. In those with prior lymphoma or leukemia, there was a high proportion of patients with a benign diagnosis (6 of 14), perhaps reflecting the predisposition of these patients to infections. In total, 30 of 161 patients (19%) had nodules of a benign nature. The large percentage of benign diagnoses in this study by Quint and colleagues (1994–1999) contrasts with a larger but much older study by Cahan and associates (1940–1975) in which only 11 of 800 patients (1.3%) had benign disease. The higher incidence of benign disease in the more recent study can probably be explained by two factors: (1) because of the greater use of CT scanning and thus higher rate of detection of nodules in the later study and (2) because the earlier study was a surgical series in which patients with a high pretest probability of benign disease may have been excluded. A consistent finding in these two studies representing nearly 40 total years from two large centers was the high proportion of patients with head and neck squamous or esophageal cancers in whom a primary lung cancer subsequently developed. Aerodigestive cancers share common risk factors, and the presence of a smoking history increases the likelihood that a solitary lung nodule in someone with a prior extrapulmonary neoplasm will prove to be a lung cancer. By understanding the data in these and other studies, clinicians can more accurately estimate the likelihood that a pulmonary nodule is a metastatic malignancy, a primary lung cancer, or benign.
Pathology
When a nodule is found to be malignant, the etiology of the nodule can be assessed by comparing the pathology of the prior cancer and the lung nodule. The pathologist’s approach begins with a knowledge of the prior cancer, its stage and grade (if applicable), and the pretest likelihood of lung metastasis from a tumor with such characteristics. The gross and microscopic pattern of the possible metastasis may be sufficiently similar to the original primary tumor that the diagnosis is rendered easily. In many situations, the histology is not sufficiently distinctive to permit a differentiation between metastasis and primary lung cancer, and immunohistochemical stains are used to help guide the diagnosis ( Table 55-2 ). Lung adenocarcinoma can be identified by specific immunohistochemical markers including cytokeratin-7, thyroid transcription factor 1 (TTF1, see Fig. 19-4B ), and napsin. For example, cytokeratin-7 can be useful for distinguishing adenocarcinoma of the lung from colon cancer metastasis, with cytokeratin-7 favoring lung and cytokeratin-20 favoring a colon primary. TTF1, a nuclear transcription protein expressed in embryonic and adult epithelial cells of the lung and thyroid, is detected in 75% of lung adenocarcinomas, but rarely detected in other adenocarcinomas that commonly metastasize to the lungs except for thyroid cancer. Napsin is an aspartic proteinase that is present in lung and kidney epithelial cells, and is a newer marker for lung adenocarcinoma with similar sensitivity but slightly greater specificity for lung origin than TTF1. Markers specific for other tumors can be used to identify lesions as metastatic; for example, breast cancers can generally be identified if they maintain their original pattern of staining for estrogen receptor or Her-2/Neu, melanomas can be identified by S100 or HMB45 immunoreactivity, and renal carcinoma can be identified by immunostaining for the product of paired box 2 or 8 genes, PAX2 or PAX8 . Thus, when the primary tumor stains for specific markers, these stains can be useful for identifying a lesion as a metastasis. Squamous carcinoma presents a particular challenge for distinguishing between a metastasis and a primary lung carcinoma, because there are no markers currently capable of distinguishing between squamous tumors of lung origin and those from extrapulmonary epithelial tissues.
| Cancer | Positive Markers | Negative Markers |
|---|---|---|
| Lung | CK7, TTF1, * SP-A, SP-B, Napsin, * p63 † | CK20, PAX2, PAX8 |
| Colon | CK20 | TTF1, Napsin |
| Bladder (urothelium) | S100P, GATA3 | TTF1, Napsin |
| Breast | ER, PR, HER2/neu | TTF1, Napsin |
| Prostate | PSA, prostatic acid phosphatase | TTF1, Napsin |
| Melanoma | HMB45, S100, tyrosine hydroxylase | Cytokeratin ‡ |
| Thyroid | TTF1, thyroglobulin | Napsin |
| Germ cell | Alpha-fetoprotein | TTF1, Napsin |
| Renal cell | PAX2, PAX8, RCCma | TTF1, Napsin |
† Squamous carcinoma specific.
‡ Fewer than 10% of melanoma metastases stain positively for cytokeratin.
Molecular Classification
Progress in transcriptional and proteomic characterization of tumors has provided tools to assist in determining the nature of a lung nodule. Giordano and colleagues attempted to develop a molecular classification scheme to discriminate adenocarcinomas of the lung, colon, and ovary. They identified three groups of 20 differentially expressed genes that correctly identified the origins of all but 2 of 154 tumors. Other groups of investigators have proposed classifiers for a broader representation of tumors in an effort to categorize tumors diagnosed as carcinoma of unknown primary. Tothill and colleagues developed a gene array-based classifier using a microarray platform and validated with real-time polymerase chain reaction on 229 samples of known origin (a training set). This gene classifier was then able to assign a “tissue of origin” in 9 of 11 cases of carcinoma with unknown primary. They noted that the two specimens that the gene classifier failed to classify were actually both squamous carcinomas. Subsequent review of the clinical course of the patients supported the assessment of the gene classifier. In another study, a gene expression-based classifier was devised to distinguish metastases of head and neck squamous carcinoma from primary lung squamous cancer, a task made difficult by both the histologic similarity as well as shared risks for both diseases. Ten genes were identified whose expressions enabled discrimination between tumors of the lung and of the upper aerodigestive origin with high accuracy in both the training and validation samples. Certainly, molecular techniques provide new ways of classifying tumors and, in so doing, improve the ability to identify the origin of a lung nodule. When the prior tissue is available for comparison, molecular testing may also assist identification of the new nodule as either a metastasis or a new cancer.
Finding viral antigens may help identify malignancies and reveal different roles in pathogenesis. Recent recognition of the role of certain strains of human papillomavirus (HPV) in squamous carcinogenesis of the cervix and oropharyngeal mucosa had caused some investigators to investigate whether this oncogenic virus was involved in the development of lung squamous carcinoma. However, in a recent study, investigators using both in situ hybridization and polymerase chain reaction based genotyping found that, in 132 squamous lung carcinomas, only 5 (1.5%) contained the HPV genome. All five of these patients had prior diagnoses of HPV-related squamous carcinoma in a different tissue and thus the HPV-positive tumors were considered to be metastatic. This would suggest that the finding of HPV-associated squamous cancer in the lung might serve as an indicator of metastasis, particularly in the setting of a prior HPV-positive extrapulmonary malignancy.
Options for Obtaining a Tissue Diagnosis
A diagnosis may be confidently made without the need for a biopsy when a patient with a history of prior cancer known to metastasize to the lungs presents with nodules with highly typical characteristics: multiple, new, or growing lower lobe pulmonary nodules with smooth borders. A diagnosis may also be possible without the need for a biopsy when, less commonly, there are measurable diagnostic serum markers (CA19-9, CA125, carcinoembryonic antigen, alpha-fetoprotein, β human chorionic gonadotropin, CYFRA21-1). However, for most tumors other than those of germ cell origin, these epithelial markers are not sufficiently specific and a biopsy is warranted to guide therapy. In all cases, before performing a lung biopsy, it is important to search for evidence of metastasis to other sites where a less invasive biopsy may yield the diagnosis. This should begin with a thorough physical examination directed to finding any enlarged lymph nodes, skeletal tenderness, or hepatomegaly that would guide the next imaging study or tissue sampling. In older series, much of the data on tissue biopsies is derived from surgical material, and this remains the “gold standard” due to the quantity of material available to the pathologist. Surgery is often used for diagnosis when it is the appropriate definitive therapy for patients with lung metastases from solid tumors. However, there are nonsurgical biopsy options, each with strengths and drawbacks that can be weighed to achieve the best balance between certainty, risk, and cost. Fiberoptic bronchoscopy, percutaneous CT-guided biopsy, and surgical resection are discussed herein, with each successive option providing increased accuracy (sensitivity and negative predictive value) but also increased risk and cost.
Bronchoscopy
Fiberoptic bronchoscopy is particularly useful for accessing centrally located lesions and mediastinal lymph nodes. Transbronchial needle aspiration of enlarged mediastinal lymph nodes is safe and has excellent diagnostic accuracy for malignant involvement of mediastinal lymph nodes. Use of endobronchial ultrasound (EBUS) improves sensitivity and permits the biopsy of much smaller lymph nodes in both the mediastinum and hilar lymph node stations (see Chapter 22 ). The accuracy of standard bronchoscopy for lung parenchymal lesions declines rapidly as the distance from the main-stem bronchi increases, so that traditional transbronchial sampling techniques have less than 20% sensitivity for peripheral lung nodules. The increasing use and availability of navigational bronchoscopy has allowed biopsy of small peripheral lung nodules with significantly greater accuracy. Experienced users of this newer approach report a 69% to 80% sensitivity for the diagnosis of small peripheral lung nodules (as small as 7 mm, ranging up to 8 cm, and average size less than 2 cm). Combining navigation-assisted bronchoscopy with real-time imaging using a radial probe ultrasound catheter to confirm the location suggested by virtual images further increased the accuracy of transbronchial biopsy to 88% in one series. Using bronchoscopy, multiple samples can be taken from one area or from multiple different nodules in a single procedure without an appreciable increase in the risk for pneumothorax. Given the low risk of pneumothorax (1% to 5%) and of major bleeding, the chief risk of the bronchoscopic biopsy, even when combined with electromagnetic navigation and real-time radial probe ultrasound, is that of a false-negative biopsy result.
CT-Guided Biopsy
CT-guided biopsy by experienced clinicians has excellent accuracy (see Chapter 19 , Figs. 19-1 to 19-2 ). This approach has greater sensitivity than bronchoscopy, particularly for peripheral lesions. The sensitivity of CT-guided biopsy of lung nodules varies with the size of the lesion, with a 65% to 75% accuracy for smaller lesions (<1 cm diameter) and a greater than 95% sensitivity for lesions larger than 1.5 cm in diameter. Larger size and a more peripheral location are associated with greater diagnostic accuracy. Core biopsies, in addition to or instead of cytologic material, can be obtained with larger 19-gauge needles (see Fig. 19-6 ). The major risk of CT-guided transthoracic needle biopsy is pneumothorax, which develops in up to 35% of patients and 10% to 15% requiring tube thoracostomy or catheter drainage. The risk of pneumothorax appears to increase with the length of the needle tract, as well as the presence of obstructive lung disease. In one large retrospective series, severe complications of CT-guided biopsy were quite rare. Of 9783 biopsies, 74 (0.75%) patients were reported to have severe complications, including 6 patients with air embolism (see eFig. 19-2 ), 10 with tension pneumothorax, 6 with severe pulmonary hemorrhage, 9 with hemothorax (see eFig. 19-1 ), and 6 with seeding of the needle tract. Eight deaths were reported. In general, CT-guided biopsies are safe and accurate for peripheral lesions.
Surgery
Resection of a nodule provides the greatest diagnostic sensitivity and may provide both a diagnosis and treatment in a single procedure. The selection of patients for whom nodule resection is appropriate is a matter of controversy. There are few, if any, randomized studies comparing surgical and nonsurgical treatment of lung metastasis, and the data from reported series on surgical treatment of lung metastases suffer from a lack of uniformity in selection criteria. The sensitivity of surgery for diagnosis is likely close to 100% given that it is the accepted gold standard for tissue diagnosis of lung lesions. However, very small lung lesions can be difficult for the surgeon to palpate and may be missed. The risks of surgery include those for general anesthesia in addition to the risks associated with lung resection. Thoracoscopic surgery is discussed later and is the preferred option when surgery is required for diagnosis.
Diagnosis
Differential Diagnosis
Given the range of diagnostic and therapeutic options in cases of known or suspected lung metastasis, patient care is optimal when delivered within the context of a multidisciplinary team. For example, when deciding whether to biopsy a lung nodule in a patient with a prior cancer, a chest physician is well served by speaking to the oncologist and/or the surgeon to learn whether a patient with metastatic cancer has therapeutic options and whether the results of a biopsy would alter those options. If a patient with a solitary nodule suspected to be an isolated lung metastasis is a surgical candidate and a complete resection is feasible, a preoperative tissue diagnosis may not change the treatment plan and, if so, should not be undertaken. Conversely, surgeons may prefer a preoperative tissue diagnosis to improve their discussion of risks, benefits, and alternatives with the patient. Surgeons may also prefer knowing whether a nodule is a metastasis or a primary lung cancer in planning the appropriate surgical procedure, either a lobectomy or an anatomic segmentectomy. For these and other reasons, the management of these patients is best served when clinicians are readily able to communicate patient-specific concerns to one another, such as in a multidisciplinary conference.
In general, the patient with a prior history of cancer and suspicion of a lung metastasis should be managed with knowledge of several interacting factors: the type of prior cancer, its natural history, its propensity to cause lung metastasis, its sensitivity to systemic or radiation therapy, and the number and location of lung nodules. A decision to biopsy should take into account the likelihood of an alternative diagnosis and whether the results of the biopsy will change the management options. The diagnostic possibilities for a lung nodule or multiple nodules in the patient with a history of cancer vary with the type of prior cancer. For example, the implications of finding a cavitary nodule in a patient with prior leukemia after a bone marrow transplant are quite different from finding the same cavitary nodule in an individual with prior head and neck cancer. In general, the differential diagnosis of a lung nodule in a patient with prior cancer includes the same diagnoses found in the general population, with the obvious exception that metastatic disease must be considered as well. Both primary malignant and benign causes, including infectious and noninfectious etiologies, should be considered ( Table 55-1 ).
