Secondary Lung Tumors Workup

Updated: Feb 16, 2021
  • Author: Daniel S Schwartz, MD, MBA, FACS; Chief Editor: John Geibel, MD, MSc, DSc, AGAF  more...
  • Print

Approach Considerations

Secondary lung tumors may be identified when patients are evaluated for symptoms such as chest pain, dyspnea, cough, or hemoptysis or when patients with known primary tumors are being staged for metastases.

A clinical scenario that is not infrequently encountered is an incidental finding of secondary lung cancer of unknown origin, known as adenocarcinoma of unknown primary (ACUP), when patients are undergoing screening chest radiography, computed tomography (CT), or positron emission tomography (PET) with CT.

Radiographically, secondary lung tumors can manifest as discrete nodules (single or multiple), interstitial infiltrate(s), or endobronchial lesions with or without distal atelectasis or postobstructive pneumonitis. They often have a characteristic round appearance on chest radiographs. [4]

Diagnostic strategies for ACUP after the initial clinical and radiologic stepwise evaluation include extensive immunohistochemistry, which may yield a final classifying diagnosis in up to 50% of patients, followed by gene expression (or reverse transcription–polymerase chain reaction [RT-PCR]), which may then be expected to provide additional classifying information in the remaining patients. [5, 6, 7, 8]

The clinical decision to pursue tissue diagnosis depends on whether confirmation of clinical findings would alter treatment. Treatment of secondary lung tumors can be performed for curative intent, to reduce or eliminate tumor burden, or to palliate disease.


Laboratory Studies

The usual preoperative laboratory workup of any thoracic patient should include a coagulation profile consisting of a platelet count, international normalized ratio (INR), and activated partial thromboplastin time (aPTT). A complete blood count (CBC) and electrolyte count should also be performed, to screen for any hematologic derangements such as anemia or electrolyte abnormalities (eg, hypokalemia) that could impact anesthesia.

Cancer-specific tumor markers

Follow-up of cancer-specific tumor markers in serum is rarely clinically useful for diagnosis or prognosis. Examples of tumors in which serum markers can help to increase the specificity of imaging studies for establishing the diagnosis of pulmonary metastases include the following:

  • Nonseminomatous testicular germ cell tumors - In which the elevated levels of alpha fetoprotein and/or the beta subunit of human chorionic gonadotropin can help to predict tumor recurrence
  • Well-differentiated papillary or follicular thyroid cancer - By identification of elevated thyroglobulin levels
  • Prostate cancer - In which any detectable prostate-specific antigen in the serum after initial treatment suggests persistent disease or recurrence

Chest Radiography

Chest radiography using high-quality posteroanterior and lateral radiographs remains the most common imaging study in the initial staging evaluation of lung cancer patients. However, because of poor yield, it is rarely recommended as a part of the initial workup for common cancers (eg, breast cancer and colon cancer) at an early stage.

This is reflected by the observation that lung metastases have been detected with radiography in only 0.1% of the patients with stage I breast cancer. Chest radiographs are limited by the potential to overlook lesions located in the lung apices or posterior sulci or against the heart or mediastinum and by their overall poor sensitivity for lung nodules of less than 1.6 cm in diameter (far lower sensitivity than that of CT).

Overall, approximately 25% of the total lung volume is not readily accessible for visual examination with plain posteroanterior chest radiography.

However, recognition of secondary pulmonary tumors has increased with advances in this modality. Improvements in technique, including the use of Advanced Multiple Beam Equalization Radiography (AMBER) and a digital slot-scan charge-coupled device (CCD) system, have increased the utility of this simple and inexpensive staging modality.

Kroft et al showed that AMBER and CCD digital film systems were equivalent in detecting phantom nodules in or around the mediastinum (135/288 [46.9%] and 128/288 [44.4%], respectively). Both of these technologies were superior to older Bucky screen film technology (65/288 [22.6%]). [9]

However, other studies, using chest CT as the criterion standard, failed to confirm that these techniques had a significant advantage over standard chest radiographs. [10]


Computed Tomography

Since the introduction of CT in the 1970s, remarkable advances have been made not only in clinicians’ ability to diagnose lung cancer but, more important, in clinical staging. CT can define the location, size, and anatomic characteristics of a tumor far better and more precisely than chest radiography can, [10]  and it is used to delineate the locoregional extent and distal spread of a lung tumor.

The major advantages of CT are related to its axial format, higher-density resolution, and wider dynamic range. Continuous technical improvements and the development of more powerful and faster computers are responsible for the fact that current CT examinations of the chest produce a large amount of detailed imaging information in a very short time. Because of this evolution in technique, the development of new therapeutic strategies for lung cancer, and the introduction of PET, the contribution of CT to the staging of patients with lung cancer has been fluid. [11]

The identification of smaller lesions with CT offers the opportunity for improved diagnosis and earlier treatment of metastatic disease, which are likely to be beneficial. However, the magnitude of benefit has not been clearly documented by the literature. The increased sensitivity of CT has also resulted in an increased frequency of identification of nonmalignant lesions, which must be distinguished from true malignancies.

Sensitivity and specificity

Conventional CT of the chest from the level of the superior thoracic aperture to the adrenal glands is superior to plain chest radiography for the detection of pulmonary nodules and mediastinal lymph node involvement. Spiral (helical) CT further increases the odds of detecting pulmonary nodules.

Increased sensitivity comes at the cost of somewhat decreased specificity in comparison with standard chest radiography (posteroanterior and lateral) and conventional CT. However, the specificity of a test is strongly influenced by clinical circumstances. Thus, in highly selected patients (eg, those with osteogenic sarcoma or soft-tissue sarcoma, which are tumors that have a high propensity for metastasizing to the lungs), 95% of nodules on the CT scan have been shown to represent metastases.

Technetium-99m (99mTc)-labeled somatostatin analogue depreotide single-photon emission CT (SPECT) is used for evaluation of pulmonary nodules and staging of lung cancer, with reported sensitivity and specificity comparable to those of PET.

Indium-111 (111In)-labeled somatostatin analogue octreotide scanning is recommended for localization of carcinoid tumors. Whole-body iodine-131 (131I) scanning is recommended for the diagnosis of metastatic thyroid cancer.

Metastases vs benign lesions and primary cancers

In a patient with a known extrathoracic malignancy and a solitary pulmonary nodule on CT, various scenarios to identify metastatic lesions have been proposed.

With a history of sarcoma or melanoma, the pulmonary nodule is more likely to be a metastasis. In the case of underlying head and neck cancer or breast cancer, a second primary cancer in the lung is more likely. With other malignancies, the nodule is equally likely to be a primary lung cancer or metastatic disease.

Malignant lesions account for 3-10% of CT-detected pulmonary nodules. In an older patient, a solitary nodule is more likely to be malignant (lung cancer, in particular); in a younger patient, multiple nodules are more likely to be metastases. However, the number of pulmonary nodules is generally not helpful in distinguishing between benign and malignant lesions.

Generally, the larger the nodule, the more likely it is to be malignant (80% of solitary nodules >3 cm in diameter were malignant, compared with 20% of nodules < 2 cm), though autopsy data show that 57% of all metastases are 1-5 mm in diameter. Most of the nodules resected at the time of thoracotomy but not seen on a CT scan are small, fibrous lesions.

The mass-vessel sign (ie, a vessel entering the medial aspect of a discrete nodule) indicates hematogenous metastasis. Irregular nodule margins indicate a poor prognosis. An ill-defined margin is observed in choriocarcinoma and in other cancers after chemotherapy, indicating hemorrhage.

Calcification, cavitation, and doubling time

Calcified pulmonary metastases are observed with osteogenic sarcoma, chondrosarcoma, synovial sarcoma, ovarian cancer, breast cancer, colon cancer, and thyroid cancer. Cavitation occurs in pulmonary metastases of sarcomas and squamous cell carcinoma, as well as after treatment.

Patterns of calcification strongly suggestive of a benign nature of a nodule are diffuse homogenous calcification, central calcification, laminated concentric calcification, and popcorn calcification.

A doubling time of between 20 and 400 days is consistent with a malignant lesion. Doubling of the volume means that a nodule 0.5 cm in diameter increases by 0.12 cm in diameter, a 1-cm nodule increases by 0.26 cm in diameter, a 2-cm nodule increases by 0.52 cm in diameter, a 3-cm mass increases by 0.78 cm in diameter, and so forth.

Absence of any changes in size over a 2-year follow-up period is generally accepted as evidence of the benign nature of the nodule. Thin-section CT with three-dimensional (3D) reconstruction of the nodule is a particularly accurate method for assessing size changes.

Lymph nodes

Mediastinal nodes are considered positive on CT on the basis of size criteria—namely, whether the short axis is 1 cm or greater. Nineteen percent of nodes from 0.5-1 cm have been reported positive for micrometastases. Seventy-five percent of lymph nodes with cancer involvement are 1 cm or greater in diameter.

High-resolution CT is the imaging procedure of choice for lymphangitic carcinomatosis. Characteristic findings include thickened septal lines, prominent reticular patterns, nodular thickening of bronchovascular bundles, polygonal lines, and beaded septa. Hilar or mediastinal lymphadenopathy, lung masses, and lung nodules are also commonly identified.

Compared with sarcoidosis (a model of benign interstitial lung disease), lymphangitic carcinomatosis is more commonly unilateral or markedly asymmetric and is associated with fewer nodules and less distortion of surrounding lung parenchyma.


Magnetic Resonance Imaging

Magnetic resonance imaging (MRI) of pulmonary pathology offers little improvement over CT, with a few exceptions. MRI is often superior to other imaging modalities in the investigation of paravertebral tumors and superior sulcus tumors. In paravertebral tumors, imaging of the spinal canal without contrast media is possible. The use of routine MRI for all lung cancer is probably superfluous and not cost-efficient. MRI should be reserved for times when local tumor invasion of the mediastinum, thoracic inlet, or paravertebral region is questioned on CT. [12]


Positron Emission Tomography

PET is a physiologic imaging modality that is fundamentally based on the detection of positrons emitted by isotopes of atoms with low atomic weights. Fluorodeoxyglucose (FDG), a D-glucose analogue, is the compound most commonly used for PET. It is a D-glucose labeled with positron-emitting fluorine-18 (18F). Cells take up and phosphorylate FDG as if it were glucose. However, FDG is not metabolized further and tends to accumulate intracellularly. [13]

In general, malignant cells have a higher rate of glucose metabolism than normal cells do. Thus, the intracellular accumulation of FDG, coupled with the preferential accumulation of glucose or its analogue in malignant cells, leads to the visualization of malignancies on PET.

PET is currently used as a diagnostic and staging tool in cancer. In particular, PET is being applied to staging lung cancer. [14]  This modality has a high likelihood of assessing the malignant potential in a pulmonary nodule, particularly if the nodule is solid and larger than 1 cm in diameter. A standard uptake value of greater than 3 is sensitive and specific for cancer. [15]

Limitations of PET include an inability to detect brain metastases, false-negative results in diabetic patients and in patients with malignant lung nodules less than 1 cm in diameter (size has not been shown to play a role in the detection of mediastinal lymph node metastases), and false-positive results in persons with granulomatous or inflammatory diseases. Cost remains an important consideration in ordering this test.

Metabolic imaging of the lungs, such as with PET, is now widely used in clinical practice. The ultimate aim of various advances in lung cancer imaging is to enable clinicians to distinguish between malignant and nonmalignant lesions without the need for tissue sampling. This goal has not yet been achieved. However, these newer imaging modalities play an increasingly important role in clinical decision-making algorithms, research, and drug development. [14, 15, 16]


In a study of 138 patients with hepatocellular carcinoma, Lee et al found that FDG-18 PET and CT together were extremely useful in detecting lung metastases larger than 1 cm, as well as bone metastases. PET had a 92.3% detection rate for secondary pulmonary nodules of 1 cm or greater, but only a 20% detection rate for lung metastases smaller than 1 cm. [17]

The investigators also found that chest CT was significantly more accurate than PET in detecting lung metastases and that PET was significantly more accurate than bone scanning in detecting bone metastases.


Whole-Body PET (Oncologic PET-CT)

The fusion of CT and PET (integrated CT-PET ) is now widely available and is very commonly used in clinical practice. Integrated CT-PET has been shown to be superior in anatomic localization and metabolic characterization of lesions as compared with CT alone, with PET scanning alone, or with using CT and PET and visually correlating the abnormalities. [18]

Types of PET-CT scanners

The terminology for PET-CT software and hardware can be confusing. The three primary modalities of PET-CT scanners are hybrid, fusion, and visually correlated. The hybrid, or integrated, PET-CT scanner creates two images, with one relying on CT and the other on PET. A computer then merges the two scans into a single image. This is the most accurate and specific system to date for the staging of non-small cell lung cancer (NSCLC). It is more expensive than PET, CT, or fusion software alone.

Fusion PET-CT scanners use software to create a 3D model of the CT study and a 3D model of the PET study; the scanners then use an algorithm to compare and provide an overlay of the images. This modality is less costly than hybrid PET-CT, but it may not be as accurate as integrated PET-CT for NSCLC. [19]

With fusion software, the CT and PET scans may be obtained on different dates; however, this increases the artifact, because there is different positioning, respiration, and other movement between scans. The fusion software can also be used with MRI.

With visually correlated PET-CT, the radiologist visually and manually compares CT and PET scans side by side. The examinations can be performed on different dates or at different facilities; however, this modality has been shown in several studies to be far less accurate. [19, 20]

False positives/negatives

With any PET-CT modality, the clinical stage often differs from the pathologic stage, meaning that significant false positives and negatives remain. The value of PET-CT scans is that they help to direct the surgeon toward targets for biopsies to rule out nodal or systemic disease. All suspicious areas should be biopsied, but the practice of using a positive PET or PET-CT scan as definitive evidence of cancer is absolutely wrong.


PET-CT has enhanced the ability to spatially identify structures that could more accurately evaluate the stage, as well as the individual T, N, and M status, in patients with NSCLC. [13, 21]

Only FDG-18 is currently used in PET-CT scans, but new radiopharmaceuticals and the prospects for developing other new radiotracers for imaging seem to be promising. [22]  As before, any new radiotracer must be carefully assessed to determine its accuracy at each nodal station and at each metastasis site.

In a retrospective study of 50 patients with lung lesions suspicious for cancer, integrated CT-PET correctly predicted T status in 86% of patients, N status in 80% of patients, M status in 98% of patients, and TNM status in 70% of patients. In comparison, correct prediction rates with CT alone were lower, reaching 68%, 66%, 88%, and 46%, respectively. With PET alone, the correct prediction rates were 46%, 70%, 96%, and 30%, respectively, and with visually correlated CT and PET scans, the correct prediction rates were 72%, 68%, 96%, and 54%, respectively.



Mediastinoscopy is the criterion standard for the diagnosis of mediastinal lymph node metastatic disease. Reported specificity of the procedure is as high as 100%, with a sensitivity of approximately 90%.

Cervical mediastinoscopy by the Carlens method is used for the diagnosis of right-side paratracheal, precarinal, and subcarinal lymphadenopathy. Left-side parasternal mediastinoscopy is used for the diagnosis of anterior mediastinal and aortopulmonary window lymph node metastases.

Mediastinoscopy is an outpatient procedure with a reported complication rate of 2% and a procedure-related mortality of 0.2%.


Aspiration and Biopsy

Transthoracic needle aspiration biopsy

Transthoracic needle aspiration (TTNA) biopsy remains the initial procedure for the diagnosis of pulmonary nodules.

A 1999 meta-analysis of 48 studies reported a pooled sensitivity for malignant lesions of 86.1% (range, 83.8-88.4%), with a pooled specificity of 98.8% (range, 98.4-99.2%). [23]  CT-guided TTNA biopsy was more sensitive than fluoroscopy-guided TTNA biopsy, though other factors are used to determine which procedure is more suitable for an individual patient. Also, aspiration biopsy needles were shown to yield better results than cutting needles.

Other authors consider bronchoscopy and TTNA biopsy to be complementary procedures and advocate their sequential use. TTNA biopsy has been reported to have a high yield for malignant nodules after an indeterminate bronchoscopy.

Pneumothorax is the most consistently reported complication of the procedure. The meta-analysis reported a pooled rate of 24.5% (range, 3.1-41.7%). The pooled rate of pneumothorax requiring chest tube drainage was 6.8% (range, 0-16.6%). Bleeding of varying severity, air embolism, myocardial infarction, and local iatrogenic spread of the tumor have also been reported following the procedure.

Transbronchial needle aspiration

Bronchoscopy with transbronchial needle aspiration (TBNA) for mediastinal lymphadenopathy or peripheral lung lesions, forceps biopsy, brush biopsy, brush-needle biopsy, bronchial aspirate, bronchial washing, or bronchoalveolar lavage (BAL) is used for the diagnosis of endobronchial tumor, lymphangitic cancer, and pulmonary nodule(s), with decreasing order of yield.

The overall yield of noninvasive bronchoscopic specimens (ie, bronchial aspirates, bronchial washings, BAL) for diagnosis of peripheral lesions is just less than 50%. The highest yield of BAL is in lymphangitic carcinomatosis.

The diagnostic yield of fiberoptic bronchoscopy depends on the lesion location and size, the character of the border, and the ability to perform all sampling methods. Diagnostic yield for lesions less than 2 cm in diameter is 54%, compared with 80% for those more than 3 cm in diameter. For lesions located in the lower-lobe basilar segments or in the upper-lobe apical segments, yield is 58%, compared with 83% for other locations, and for lesions with sharp borders, the yield is 54%, compared with 83% for lesions with fuzzy borders. Only one of the sampling methods was positive in 24% of bronchoscopies.

The overall yield of invasive bronchoscopic specimens for diagnosis of peripheral lesions is 52% for brush, 57% for transbronchial biopsy, and 51% for transbronchial needle aspiration.


Combining TBNA with PET has been shown to obviate the need for mediastinoscopy for mediastinal staging of non-small cell lung cancer with mediastinal lymphadenopathy in most patients.

In a retrospective study of patients with enlarged mediastinal lymph nodes, the combination of TBNA and PET demonstrated higher sensitivity, negative predictive value, and accuracy than did either modality alone. The study used histopathology by surgical lymph node dissection as the criterion standard and found that the combined TBNA and PET scan had 100% sensitivity, 94% specificity, 79% positive predictive value, 100% negative predictive value, and 95% accuracy in the detection of malignant lymph nodes. For PET alone, these rates were 68%, 89%, 46%, 95%, and 86%, respectively; for TBNA alone, these rates were 54%, 100%, 100%, 91%, and 92%, respectively.

Electromagnetic navigation bronchoscopy with biopsy

This uses technology that allows the operator to approach a peripheral lung mass using electromagnetic navigation (EMN) based on virtual bronchoscopy and real-time 3D CT images. The technology has been shown to be capable of reaching peripheral lung masses beyond the reach of the standard bronchoscope in an animal model. [24]

However, a multicenter study by Ost et al, using data from the AQuIRE (ACCP [American College of Chest Physicians] Quality Improvement Registry, Evaluation, and Education) registry to measure and identify the determinants of diagnostic yield for bronchoscopy in patients with peripheral lung lesions, found that EMN had a lower-than-expected yield. [25]  

Endobronchial ultrasonography with biopsy

Endobronchial ultrasonography (EBUS) has been widely adopted by pulmonologists and thoracic surgeons and is poised to replace mediastinoscopy in the future. For thoracic surgeons, the technique can be easily learned, and it may be important to do so to maintain its traditional and important role in the diagnosis and staging of thoracic malignancies.

EBUS has been associated with a low (< 1%) rate of serious adverse effects, and the procedure is touted as being highly accurate, with reported false-negative rates in the range of 6-9%. The aforementioned study by Ost et al found radial EBUS to have a lower-than-expected yield in patients with peripheral lung lesions. [25]

EBUS-guided fine-needle aspiration (FNA) biopsy of mediastinal nodes offers a less invasive alternative for histologic sampling of the mediastinal nodes.

Esophagoscopy with ultrasonographically guided needle aspiration

Esophagoscopy with ultrasonographically guided needle aspiration of accessible lymph nodes is an alternative to TBNA of lymph nodes accessible from the esophagus. It appears to be complementary to EBUS. [26]

Video-assisted thorascopic surgery with biopsy

Video-assisted thoracoscopic surgery (VATS) with lung biopsy is an inpatient procedure with a high diagnostic yield and a low complication rate. It can also be used for curative resection.


Immunohistochemistry and Gene Expression


The combination of a stepwise approach, with initial clinical and radiologic evaluation and a biopsy procedure, followed by histologic evaluation with extensive immunohistochemistry, may yield a final classifying diagnosis in up to 50% of patients who have not been otherwise diagnosed. [5]

Gene expression

In the remaining patients who have not been otherwise diagnosed, further classification based on gene expression may be expected to provide additional classifying information. [27, 28, 29]

The other approach would be a rather simultaneous method in which the gene expression profile is determined up front. Complete replacement of histologic and immunohistochemical evaluation by these methods has been suggested. Both strategies may have pros and cons in terms of accuracy, time frames, and costs.