Procedures for tissue biopsy in patients with suspected non-small cell lung cancer

INTRODUCTION — Lung cancer is the most common cancer worldwide. Non-small cell lung cancer (NSCLC) accounts for approximately 85 percent of all lung cancers, of which, adenocarcinoma is the most common histologic subtype [1].

Traditionally, the histopathologic diagnosis of NSCLC has been made based upon information obtained from small biopsies. There has been a shift towards the use of therapies targeted at specific mutations, the identification of which can be challenging on low volume biopsy samples. This, together with the rising incidence of adenocarcinoma and its classification into subgroups, has led to greater emphasis on the yield of tissue biopsy for the diagnosis of NSCLC. (See "Pathology of lung malignancies" and "Personalized, genotype-directed therapy for advanced non-small cell lung cancer".)

This topic will discuss the accuracy of procedures commonly used to obtain tissue for diagnosis and staging in patients with suspected NSCLC. The clinical presentation, initial evaluation and staging, and approach to the selection of modality for biopsy of patients with suspected NSCLC are discussed separately. (See "Overview of the initial evaluation, diagnosis, and staging of patients with suspected lung cancer" and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer".)

SELECTION OF MODALITY FOR TUMOR BIOPSY — Selection of the optimal biopsy target and modality should be driven by evidence-based and institution-specific protocols that preferably involve a lung cancer tumor board or multidisciplinary team. When an institution is equipped with reliable expertise, one of the minimally invasive or more invasive tools listed in the section below may be used to biopsy a target lesion. Importantly, estimates of diagnostic accuracy and diagnostic yield for individual modalities are usually based upon small studies with inadequate gold standard comparators that lead to bias and significantly limit interpretation of the data. The clinician should be aware of these clinically relevant limitations when choosing among the modalities available for tissue acquisition. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer" and 'Endoscopic and percutaneous procedures' below and 'Surgical staging procedures' below.)

Prior to biopsy the following should be in place:

●A thorough history and examination with routine laboratory studies. (See "Overview of the initial evaluation, diagnosis, and staging of patients with suspected lung cancer", section on 'Initial evaluation'.)

●Imaging with contrast-enhanced computed tomographic (CT) scan of the chest and upper abdomen to assess the radiographic extent of disease prior to biopsy. Some patients may require additional imaging including positron emission tomography (PET), or integrated PET/CT to detect occult metastases or imaging of sites suspected to be involved with metastases (eg, those with symptoms or focal findings). (See "Overview of the initial evaluation, diagnosis, and staging of patients with suspected lung cancer", section on 'Suggested approaches to diagnostic evaluation and radiographic staging'.)

●Good working knowledge of the Tumor Node Metastasis (TNM) staging of NSCLC, the International Association for the Study of Lung Cancer (IASLC) lymph node map (figure 1), and the therapeutic strategies for the suspected stage. The eighth edition (table 1 and table 2) of the TNM staging system is in effect. (See "Management of stage I and stage II non-small cell lung cancer" and "Management of stage III non-small cell lung cancer" and "Overview of the initial treatment of advanced non-small cell lung cancer" and "Tumor, node, metastasis (TNM) staging system for lung cancer".)

Guidelines issued by the American College of Chest Physicians (ACCP), European Society of Thoracic Surgeons (ESTS), and International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society (IASLC/ATS/ERS) are available to guide the physician with modality selection [2-5].

The evidence for the diagnostic and staging accuracy of each procedure for NSCLC is discussed in this topic. General biopsy goals and a suggested diagnostic and staging algorithm are discussed in detail separately. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Biopsy goals' and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient'.)

ENDOSCOPIC AND PERCUTANEOUS PROCEDURES

Overview — Commonly used minimally invasive approaches employ bronchoscopic and percutaneous modalities. Bronchoscopic techniques include bronchial washings, brushings, forceps transbronchial biopsy or needle aspiration, endobronchial ultrasound (EBUS)-guided transbronchial needle aspiration (EBUS-TBNA), navigational-guided transbronchial biopsy with or without robotic bronchoscopy, and conventional TBNA. Guide sheaths, virtual bronchoscopy, and ultra-thin bronchoscopy are additional enhancements. Multiple bronchoscopic techniques may be combined in a single procedure, and this provides a potential advantage of obtaining both the diagnosis and staging at the same time.

Percutaneous approaches include transthoracic needle aspiration (TTNA) or needle/core biopsy (TTNB) of the primary tumor.

A combination of bronchoscopic and percutaneous modalities may be necessary to permit sampling from several areas of interest or when the first modality is nondiagnostic. Selecting one of these procedures or combining several modalities to achieve the most accurate histologic classification and staging information relies on knowledge of the technique's diagnostic accuracy for the target lesion(s). Critical considerations for procedure selection also include patient-specific details like cardiopulmonary status and goals of care, as well as systems factors such as available equipment, laboratory resources, and operator proficiency. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient'.)

The evidence for each minimally invasive procedure pertinent to its diagnostic and staging accuracy for NSCLC is discussed in this section. In general, the following applies:

●Bronchoscopy with EBUS-TBNA has gained wide acceptance as a first-choice diagnostic and staging procedure for thoracic lesions and enlarged mediastinal lymph nodes that are accessible by this modality. This is due to its ability to establish both a cytologic diagnosis of lung cancer as well as its high sensitivity for sampling central lesions and lymph nodes in the paratracheal, subcarinal, and hilar regions of the mediastinum [4,6-10]. Limitations of EBUS-TBNA include inability to access retrotracheal, periaortic, paraesophageal, and pulmonary ligament lymph node stations (3a, 5, 6, 8, 9) and variable operator proficiency, which may influence the rate of non-diagnostic findings. (See 'Endobronchial ultrasound' below.)

●Although transthoracic (percutaneous) needle aspiration/biopsy (TTNA/TTNB) has a high accuracy for diagnosing NSCLC, the higher risk of complications (eg, pneumothorax) is an important limitation, and it is rarely useful for mediastinal lymph node staging [2,4,11-16]. CT-guided TTNA may be especially useful for peripheral nodules and masses, especially when lesions appear to directly involve the chest wall or skeleton, or as an alternate procedure when lesions cannot be safely or reliably accessed using another modality. (See 'Transthoracic needle biopsy' below.)

●Conventional flexible bronchoscopy with forceps biopsy, brushing, or washing is best utilized for accessing large, central lesions and those with suspected airway involvement. These lesions may present with hemoptysis or postobstructive pneumonia, and direct visualization reveals endobronchial involvement or obstruction (see 'Conventional bronchoscopy' below). Conventional bronchoscopy for transbronchial needle aspiration without ultrasound guidance may be utilized for diagnosis and staging of significantly enlarged subcarinal, paratracheal, and hilar nodes (figure 1). Conventional bronchoscopic TBNA may be effectively utilized particularly in settings with cost and resource limitations [17]. However, without image guidance, TBNA alone has limited sensitivity for staging of the mediastinum. The use of ultrasound guidance for tissue biopsy enhances the ability of bronchoscopy to sample lymph nodes and tumors outside the airways, making EBUS the preferred bronchoscopic modality for diagnosis and staging [7].

●Endoscopic ultrasound (EUS)-guided fine needle aspiration (EUS-FNA) using a transesophageal approach is a sensitive staging tool for suspected NSCLC in subcarinal and paratracheal nodes. EUS-FNA can be combined with EBUS-TBNA to enhance mediastinal staging. However, it requires special expertise and is not commonly used in the United States. (See 'Transesophageal endoscopic ultrasound' below and 'Combined modalities for mediastinal lymph node staging' below.)

●Electromagnetic navigational bronchoscopy (ENB) and, most recently, robotic-assisted bronchoscopy are utilized for sampling peripheral lung lesions. These techniques are based on highly detailed digital pathways developed from CT chest imaging that are utilized in real-time during the procedure to localize the biopsy target. Both robotic bronchoscopy and ENB can be combined with other real-time imaging, including traditional fluoroscopy, enhanced or augmented multidimensional fluoroscopy, and radial probe ultrasound, to further enhance diagnostic yield. (See 'Enhanced, image-guided bronchoscopic techniques' below and "Image-guided bronchoscopy for biopsy of peripheral pulmonary lesions".)

Histologic or cytologic classification of all lung cancers remains the gold standard and the priority for tissue biopsies. Molecular and genetic profile subclassification of either the tissue specimen or blood samples to identify specific treatment targets are commonly utilized within research protocols, within established treatment protocols, and in limited circumstances for prognosis considerations (see "Personalized, genotype-directed therapy for advanced non-small cell lung cancer"). Although obtaining samples of lavage fluid or "normal"-appearing bronchial tissue for genomic analysis has been studied as a potential method to enhance the ability of bronchoscopy to rule out a diagnosis of lung cancer, the incremental value of these tests is not fully established and further study is required before these specimen types are recommended for routine use [18]. (See "Personalized, genotype-directed therapy for advanced non-small cell lung cancer".)

Guidelines issued by the American College of Chest Physicians (ACCP), European Society of Thoracic Surgeons (ESTS), International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society (IASLC/ATS/ERS), and the European Society of Gastrointestinal Endoscopy (ESGE) have incorporated the use of minimally invasive tools in the diagnostic work-up of NSCLC [19].

Bronchoscopic approaches — Although different technical bronchoscopic approaches are discussed individually in this section, in practice, most experienced bronchoscopists use multiple combined modalities to increase the overall sensitivity for the diagnosis of NSCLC.

Conventional bronchoscopy — There are several ways to obtain tissue with conventional flexible bronchoscopy (see "Flexible bronchoscopy in adults: Preparation, procedural technique, and complications", section on 'Diagnostic and therapeutic procedures' and "Flexible bronchoscopy in adults: Overview"):

●Washings, brushings, or endobronchial forceps biopsy for visually obvious lesions in the main airways.

●Transbronchial forceps for biopsy of peripheral, parenchymal lesions that are not directly visualized. Transbronchial biopsy is typically image-guided (real-time fluoroscopy or radial probe ultrasound). (See 'Enhanced, image-guided bronchoscopic techniques' below and "Image-guided bronchoscopy for biopsy of peripheral pulmonary lesions".)

●TBNA for biopsy of lesions and lymph nodes that are close to or encroaching upon the airway. Conventional bronchoscopic TBNA can be performed based on previously obtained CT imaging and anatomic relationships or may be enhanced with real-time guidance by EBUS (EBUS-TBNA or radial probe [RP]-TBNA) or by fluoroscopy. EBUS requires additional equipment, image processors, and procedural training distinct from conventional flexible bronchoscopes. (See 'Endobronchial ultrasound' below.)

●For lesions that are large or central, or for lesions that obstruct airways, the diagnostic sensitivity of conventional bronchoscopy with endobronchial or transbronchial biopsy is 65 to 88 percent [2]. For these lesions, the use of needle aspiration (bronchoscopic-TBNA) increases the diagnostic sensitivity (70 to 96 percent) [2]. Bronchoscopy has a higher sensitivity for diagnosis of the primary tumor in this setting because proximal lesions in or close to the airway are easily visualized and accessed. However, for smaller peripheral lesions that are not visible endoscopically or are located distal to sub-segmental bronchi, the sensitivity decreases [2,20]. (See "Image-guided bronchoscopy for biopsy of peripheral pulmonary lesions".)

Diagnostic and staging accuracy — For patients with central lesions and CT evidence of bronchial, carinal, or tracheal involvement, conventional bronchoscopy is essential for accurate determination of T-descriptor staging (Tumor Node Metastasis [TNM] staging). Conventional bronchoscopy with TBNA may also be sufficient in some cases where there is bulky, central mediastinal disease. However, conventional bronchoscopic TBNA without ultrasound guidance has a more limited role in mediastinal staging in most patients with hilar, paratracheal and smaller subcarinal lymph nodes because of its significantly lower sensitivity. Bronchoscopic TBNA without real-time ultrasound guidance is inferior to EBUS-TBNA for the diagnosis and staging of NSCLC (sensitivity 36 versus 69 percent) [7-9,21,22]. The limited sensitivity associated with bronchoscopic TBNA without real-time guidance for staging the mediastinum often prompts additional mediastinal biopsy procedures when testing is negative for cancer. (See 'Diagnostic and staging accuracy' below.)

Conventional bronchoscopy (preferably with TBNA if CT and/or positron emission tomography [PET] imaging demonstrate enlarged and/or hypermetabolic lymph nodes in accessible stations) can be considered when EBUS-TBNA is not available or when it is the only diagnostic modality that can be performed safely in patients in whom more invasive surgical procedures or prolonged endoscopic procedures are contraindicated [21]. For patients with smaller endotracheal tubes on mechanical ventilation and for those with tracheostomy tubes that cannot accommodate the larger diameter EBUS bronchoscope, conventional TBNA may be the only feasible modality. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient'.)

●Diagnosis of primary tumor – Conventional bronchoscopy has high diagnostic accuracy for large, central or obstructive airway lesions that can be biopsied under direct visualization using endobronchial biopsy forceps (65 to 88 percent) [2]. Although bronchial lavage and brushings can be diagnostic, these techniques generally have a low sensitivity when used in isolation (43 and 54 percent, respectively) [2]. Bronchoscopic TBNA without real-time image guidance, which involves aspirating tissue by passing a needle (19 to 21 gauge) through the working channel of a conventional bronchoscope, can sample parenchymal lesions in subcarinal, paratracheal, and submucosal locations, with sensitivity between 70 and 96 percent [23-25]. (See "Diagnostic evaluation of the incidental pulmonary nodule", section on 'Bronchoscopic techniques'.)

Bronchoscopy is less useful for small or peripheral lesions (ie, not directly visible bronchoscopically; sensitivity up to 65 percent), with higher sensitivity in lesions greater than 2 cm [2]. Thus, when bronchoscopy is nondiagnostic, and clinical suspicion remains high, further testing to pursue a diagnosis should be sought.

●Staging – Bronchoscopy is routinely performed preoperatively to assess for synchronous lesions and to delineate involvement of the central bronchi [2,26]. Specifically, bronchoscopy is necessary for assessing the proximal extent of central or endobronchial tumors (ie, tumor size [T]). Tumors that involve the mainstem bronchus (regardless of the distance from the carina but not involving the carina) are T2 lesions in the American Joint Committee on Cancer (AJCC) 8th edition staging system and tumors involving the carina are T4 lesions (table 1 and table 2). Identification of mainstem bronchial involvement can change the operative approach by the surgeon, as well as the advisability of adjuvant chemotherapy.

Conventional bronchoscopic TBNA (ie, using only anatomic landmarks and previously obtained CT images) can access paratracheal (4R, 4L), subcarinal (7), and less commonly, hilar (10R, 10L) lymph nodes (figure 1). However, a meta-analysis showed a pooled sensitivity of only 39 percent when limited to studies that enrolled patients with potentially resectable NSCLC [27]. In a pooled analysis of a more heterogeneous group of 2048 patients, including those with bulky adenopathy, the sensitivity ranged from 14 to 100 percent (median 78 percent) with a negative predictive value of 77 percent [2]. The sensitivity of conventional TBNA without image guidance is related to the size of the target lymph node with greater sensitivity for substantially enlarged nodes (eg, >2 cm) and markedly decreased sensitivity for nodes 1 cm or less.

Three to four needle passes of selected lymph nodes for staging are often needed to maximize the sensitivity. One study suggested that the diagnosis was established with the first, second, third, and fourth needle pass at 64, 87, 95, and 98 percent of targets, respectively [28]. Although the sensitivity varied widely with target site features (eg, size, location), the stepwise increment to the maximum sensitivity provided by serial passes was similar across target sites. In one meta-analysis of five studies, the addition of rapid on-site cytologic evaluation (ROSE) did not improve the diagnostic yield but did reduce the number of total needle passes and lowered the rate of complications [29].

Endobronchial ultrasound — EBUS is an endoscopic imaging tool that utilizes real-time ultrasonographic guidance to facilitate TBNA of parenchymal lesions and lymph nodes that are adjacent to the large airways, as well as forceps, needle, or brush biopsies of peripheral lesions. Although there are important technical differences between conventional and EBUS-TBNA, such as the length of the needle used, the main benefit of ultrasound guidance is related to the ability to visualize anatomy and precisely localize suspicious lymph nodes or peripheral lung lesions. There are two distinct ultrasound designs used in bronchoscopy: linear EBUS and RP-EBUS. Linear (ie, convex probe) EBUS is incorporated into the distal tip of a dedicated bronchoscope that guides TBNA. Linear EBUS-TBNA is ideally suited for sampling mediastinal lymph nodes or large, central masses. RP-EBUS is catheter-based and deployed through the working channel of a conventional flexible bronchoscope and into subsegmental and distal airways to locate peripheral lesions and then to guide forceps or brush biopsies. RP-EBUS is most commonly used for diagnosis of parenchymal nodules or masses. (See "Endobronchial ultrasound: Technical aspects" and "Image-guided bronchoscopy for biopsy of peripheral pulmonary lesions", section on 'Radial probe endobronchial ultrasound (RP-EBUS)'.)

EBUS-TBNA is most sensitive when used to sample large, central primary lesions (84 to 85 percent) and enlarged mediastinal lymph nodes (up to 96 percent) in the paratracheal, posterior tracheal, subcarinal, and hilar regions (2R, 2L, 3p, 4R, 4L, 7, 10R, 10L) (figure 1) [7-9,30-34]. High operator proficiency and multiple passes with ROSE enhance the overall sensitivity [5,35-39]. EBUS-TBNA has been shown to surpass the sensitivity of conventional flexible bronchoscopy alone for the diagnosis and staging of NSCLC [7-9]. Several studies suggest that EBUS-TBNA is at least comparable, and may be superior, to cervical mediastinoscopy [6,40-42]. (See 'Diagnostic and staging accuracy' below and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient'.)

The value of EBUS-TBNA for the diagnosis and staging of NSCLC is discussed here. The technical aspects of EBUS and other indications for its use are discussed separately. (See "Endobronchial ultrasound: Indications, contraindications, and complications" and "Endobronchial ultrasound: Technical aspects" and "Endoscopic ultrasound-guided sampling of the mediastinum: Technique, indications, contraindications, and complications".)

Diagnostic and staging accuracy — Major studies that have examined the diagnostic accuracy of radial EBUS-TBNA for NSCLC are summarized below. (See "Endobronchial ultrasound: Technical aspects" and "Endobronchial ultrasound: Technical aspects", section on 'Types of EBUS'.)

●Diagnosis of primary tumor – EBUS has a sensitivity of 73 to 85 percent for large symptomatic (eg, hemoptysis) lesions that are centrally located [7-9]. Its sensitivity declines for lesions <2 cm (71 percent), particularly those that are peripherally located (56 percent) [8,9]. Although EBUS enhances the accuracy of conventional transbronchial biopsy, CT-guided percutaneous transthoracic needle aspiration or core needle biopsy has comparable results if not superior accuracy for smaller lesions [8,9,13-15]. The sensitivity of EBUS-TBNA for identifying mediastinal lesions other than NSCLC is discussed separately. (See "Endobronchial ultrasound: Technical aspects" and "Endobronchial ultrasound: Indications, contraindications, and complications".)

●Staging – EBUS-TBNA can access upper and lower paratracheal posterior tracheal, subcarinal and hilar lymph node stations (2R, 2L, 3p, 4R, 4L, 7, 10R, 10L,) (figure 1). However, it cannot access lymph nodes in the prevascular, para-aortic, paraesophageal, or pulmonary ligament regions of the mediastinum (3a, 5, 8, 9).

•Early studies suggested that EBUS-TBNA detected malignant lymph node involvement with sensitivity and negative predictive values of 95 and 90 percent, respectively [30-33]. A subsequent 2013 meta-analysis that included 2756 patients reported that the sensitivity of EBUS-TBNA was somewhat lower than that previously reported (median 89 percent, range 46 to 97 percent) [2].

•Similarly, the negative predictive value of EBUS-TBNA is probably lower than originally thought and wide-ranging across studies due to variation in prevalence [2,40,43]. While one systematic review reported that the negative predictive value of EBUS-TBNA was 91 percent (60 to 97 percent), some studies report higher rates of false negative results and nondiagnostic findings [2,40,43].

•The variability in sensitivity and negative predictive value of EBUS-TBNA may reflect variable expertise, uncertainty related to small sample sizes, and differences in the prevalence or extent of adenopathy and in protocols for sample processing. Nonetheless, the variable estimates of observed diagnostic accuracy may also reflect natural variation in general practice.

•Most published studies have used systematic lymph node sampling that samples representative lymph nodes from each station (figure 1). The approach taken should depend upon factors including the clinical suspicion of nodal metastasis based on CT and PET imaging, the expectation of definitive surgical management of the primary lesion, the utilization of surgical mediastinal lymph node dissection as well as the expertise and resources of the management team.

Comparatively, EBUS is superior to unguided bronchoscopic TBNA for staging NSCLC [7-10,30,33,44]. Several studies that have directly compared EBUS-TBNA with the original "gold standard," cervical mediastinoscopy, suggest that EBUS-TBNA is at least comparable if not superior to mediastinoscopy [6,40,41,45]:

•One prospective, single center, cross over trial in 66 patients with suspected NSCLC showed that EBUS had a superior sensitivity to cervical mediastinoscopy (87 versus 68 percent) [6]. However, this study was performed in a center with high level expertise in interventional bronchoscopy which may not be generalizable to other centers where this level of expertise is not available. In addition, this study included a high rate of EBUS-TBNA in the subcarinal region, a region that cannot be as easily accessed by mediastinoscopy.

•Another prospective study of 153 patients with suspected NSCLC compared EBUS-TBNA to cervical mediastinoscopy in patients who were potentially resectable and in whom both procedures were performed prior to surgery [41]. The sensitivity, negative predictive value, and diagnostic accuracy were comparable for EBUS-TBNA and mediastinoscopy (81 versus 79 percent, 91 versus 90 percent, and 93 versus 93 percent).

•Another retrospective review of patients at high risk of N2 disease who had undergone both procedures reported that compared with mediastinoscopy, EBUS-TBNA had a higher rate of nondiagnostic findings (20 percent) resulting in a negative predictive value of 72 percent (95% CI 56-89 percent) [40].

Large prospective trials comparing EBUS-TBNA with mediastinoscopy are in progress to further define the procedure's sensitivity in larger, less highly selected patients.

The airway wall is well visualized with RP-EBUS [46-48]. This allows more accurate determination of the distance from the carina to the proximal portion of the tumor when the right or left main bronchi are involved. This distance is an essential part of the T component of the TNM staging system (table 1 and table 2) and will partially determine whether the lesion is T2 or T4, which can impact the disease stage. Linear EBUS does not visualize the airway wall as well as RP-EBUS; therefore, it should not be used for this purpose. (See "Overview of the initial evaluation, diagnosis, and staging of patients with suspected lung cancer" and "Tumor, node, metastasis (TNM) staging system for lung cancer".)

Limitations — The major limitations of EBUS-TBNA are that sampled specimens can be small in size, rates of operator proficiency are variable, and availability is institution-specific. In addition, EBUS cannot access periaortic and posterior lymph node stations. These limitations of EBUS-TBNA among others are listed below:

●Low volume specimen size – Compared to surgical biopsy, the amount of tissue obtained by aggregating tissue from multiple needle passes on EBUS-TBNA is small and may be insufficient for architectural and molecular assessment unless additional steps are taken.

Multiple passes (three to eight per site) with ROSE and well-defined institutional procedural protocols for processing improve the clinical utility of EBUS-TBNA [5,35-38,49-54]. The use of ROSE has also been associated with a lower need for additional procedures (eg, transbronchial biopsy) and fewer punctures during bronchoscopy [55]. Blood core clot (often obtained during EBUS-TBNA) has also been shown to contain cellular sample that can be additive diagnostically [56]. In general, only 140 to 400 ng of DNA are extracted from needle aspirations [57,58]. Estimates of amounts of total DNA extracted from EBUS-TBNA samples are approximately 280 ng, which is usually adequate for molecular testing [59-62]. Thus, with the use of multiple needle passes and ROSE by proficient endoscopists, enough tissue (up to eight needle aspirates per cell block) can be aspirated to obtain material suitable for both cytologic diagnosis as well as additional immunohistochemical and molecular testing [5,35-39,54]. This was best demonstrated in a meta-analysis of 33 studies, which reported that sufficient sample was obtained on EBUS-TBNA for molecular analysis of both EGFR and ALK mutations [63]. Another study suggested that EBUS-TBNA had a sensitivity of 72 percent for the detection of programmed cell death ligand 1 (PD-L1) expression but lower sensitivity for the detection of PD-L1 in greater than 50 percent of cells [64].

Small tissue samples are inherent to all needle aspiration techniques but may also be partly due to variable rates of expertise and minor technical differences within or between institutions [5,35-38,65]. Many interventional pulmonologists use larger needles in an attempt to aspirate more tissue. However, larger needle size (eg, 21 versus 22 gauge) has been shown in three of four studies not to have an impact on sensitivity of EBUS-TBNA [35,38,39,66].

●Availability and operator experience – Not every facility has the equipment and fully trained team of operators to perform EBUS-TBNA. Low operator experience may also contribute to low sensitivity and inaccurate staging of suspected NSCLC. Conversely, training which includes simulation practice may improve procedure performance [67].

●Low prevalence of disease – The prevalence of mediastinal metastasis is low among patients with known or suspected NSCLC who have no evidence of lymph node enlargement or hypermetabolism on imaging studies. A meta-analysis of 13 studies reported that the prevalence of mediastinal metastasis was 12 percent and the pooled sensitivity of EBUS-TBNA for identifying metastasis was only 49.5 percent (95% CI 36.4 to 62.6%) in this patient population.

●Lymph node access – EBUS cannot access all stations in the mediastinum. For example, prevascular, periaortic, paraesophageal, or pulmonary ligament lymph node stations (3a, 5, 6, 8, 9) cannot be sampled with this modality (figure 1). The access to the posterior and inferior lymph node stations by EUS-FNA has led to the combined use of EBUS and EUS for staging patients with suspected NSCLC. However, this combination is rarely used. (See 'Combined modalities for mediastinal lymph node staging' below.)

●Complications – Complications of EBUS-TBNA (eg, pneumothorax, hemorrhage) are uncommon (<1.5 percent) and discussed separately. (See "Bronchoscopy: Transbronchial needle aspiration", section on 'Complications'.)

Enhanced, image-guided bronchoscopic techniques — Image-guided techniques including virtual navigational bronchoscopy, ultrathin bronchoscopy, electromagnetic navigational bronchoscopy (ENB), and robotically assisted bronchoscopy have emerged as diagnostic modalities at many medical centers. These techniques are used primarily for accessing peripheral lesions [68].

●Electromagnetic navigation bronchoscopy – The NAVIGATE study prospectively evaluated the results of ENB biopsies in a cohort of 1215 consecutive patients undergoing ENB with 12-month follow-up [69]. The overall diagnostic yield was 72.9 percent, and the sensitivity for malignancy 68.8 percent. A meta-analysis of ENB suggested the diagnostic yield of guided bronchoscopic procedures was higher when a CT bronchus sign was present (74.1 percent) compared with when no bronchus sign was present (49.6 percent) [70]. A second meta-analysis including results from 40 studies in 3342 patients found a pooled sensitivity for lung cancer diagnosis of 77 percent, which remained consistent over both the time interval of included studies (2005 through 2020) as well as technical aspects of the navigation system used [71].

Further details regarding ENB and other image-guided techniques for the biopsy of peripheral pulmonary lesions are provided separately. (See "Image-guided bronchoscopy for biopsy of peripheral pulmonary lesions".)

Transthoracic needle biopsy — Transthoracic needle biopsy involves passing a needle percutaneously, under image guidance (usually computed tomography), to either aspirate (TTNA) or biopsy (TTNB) tissue. The needle frequently traverses pleura and lung to either aspirate or biopsy target tissue (lung or lymph node). The term TTNB in this section refers to both aspirate and biopsy samples because most studies examining sensitivity do not distinguish between these entities.

TTNB has a sensitivity of 74 to 90 percent for lung and mediastinal lesions suspected to be malignant [2,4,11,13-16,72,73]. It can access most intraparenchymal lesions and selected mediastinal lymph nodal stations (figure 1). However, traversing the pleural space and lung tissue is frequently unavoidable resulting in high rates of pneumothorax (on average 10 to 15 percent), limiting the use of TTNB as a diagnostic and staging tool [12,15,16,72]. As with all minimally invasive procedures, the relatively high false negative rates associated with needle aspiration necessitate additional biopsy procedures when the pretest probability of cancer is high but TTNB is nondiagnostic. (See 'Diagnostic accuracy' below.)

TTNB is often considered for peripheral lesions or when another procedure (eg, EBUS- or bronchoscopic-TBNA) has failed to obtain diagnostic tissue. It is also considered when the patient is at high risk for complications of video-assisted thoracoscopic surgery or cervical mediastinoscopy. The high diagnostic accuracy of TTNB should be weighed against the risk of pneumothorax and bleeding while considering patient preference. The risk benefit ratio is increased in patients with concomitant emphysema, bullous disease, or chronic respiratory failure. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient'.)

The value of TTNB for the diagnosis of NSCLC is discussed here. The indications and use of TTNB in the work-up of the solitary pulmonary nodule is discussed separately. (See "Diagnostic evaluation of the incidental pulmonary nodule", section on 'Transthoracic needle biopsy'.)

Diagnostic accuracy — Major studies that have examined the sensitivity of TTNB for NSCLC are summarized below.

●Diagnosis of primary tumor – In one meta-analysis of 46 studies, TTNB detected malignancy with a sensitivity of 90 percent [2]. The diagnostic sensitivity of TTNB for malignancy is lower for smaller lesions (<3 cm) with reported sensitivity and specificity of 67 and 100 percent respectively, and non-diagnostic results occurring in 36 percent of cases [11]. While some studies suggest that percutaneous core needle biopsy is not more sensitive than aspiration for the diagnosis of NSCLC, it has the advantage of providing the pathologist with larger amounts of material to distinguish benign from malignant pathologies and to perform molecular testing [2]. (See "Diagnostic evaluation of the incidental pulmonary nodule", section on 'Transthoracic needle biopsy'.)

●Staging – In practice, TTNB of the mediastinum is rarely performed unless there is bulky anterior lymphadenopathy. Stations 2, 4, 5, and 6 can be accessed by an anterior parasternal approach, and stations 4, 7, 8, and 9 by a posterior approach (figure 1) [3].

Compared with bronchoscopic procedures, the prevalence of pneumothorax following TTNB is high (10 to 15 versus <1 percent) [12]. The risk is greatest when the lung is traversed [12,16]. Higher rates have also been noted in older patients with COPD, in whom a higher risk of respiratory failure requiring mechanical ventilation has also been reported [12]. Most can be managed conservatively. The risk of hemorrhage is low (<1 percent) but may be higher if patients are taking anticoagulants or the target for biopsy is close to a vascular structure. Complications of TTNB are discussed separately. (See "Diagnostic evaluation of the incidental pulmonary nodule", section on 'Transthoracic needle biopsy'.)

Transesophageal endoscopic ultrasound — EUS-FNA is similar to EBUS-TBNA but may use either a larger gastrointestinal endoscope that is inserted into the esophagus or the smaller ultrasound bronchoscope (EUS-B-FNA) [19]. EUS-FNA was originally developed as a diagnostic and staging modality for gastric and esophageal cancers. However, it is a potentially useful modality for the diagnosis and staging of lung cancer, particularly, in those with mass lesions that appear to involve the posterior and inferior mediastinum as well as the left paratracheal lymph nodes. EUS-FNA requires experienced, skilled endoscopists who have received specialized training. A limitation of EUS-FNA as a stand-alone diagnostic and staging procedure is its inability to simultaneously visualize the airway and access the primary tumor mass. For this reason, EUS is performed in some medical centers in tandem with bronchoscopy and EBUS-FNA.

EUS-FNA is a sensitive staging tool for suspected NSCLC involvement in subcarinal (3, 7, 8, 9) and paratracheal nodes (2, 4) (figure 1 and figure 2) [4]. Case series also suggest a potential role for EUS-FNA in suspected M1b disease (eg, adrenal gland or liver) [74-76]. It has also been used to assess direct tumor invasion of the mediastinum, esophagus, and great vessels (T4).

The imperfect sensitivity associated with fine needle aspiration techniques necessitates additional biopsy, usually mediastinoscopy, when minimally invasive testing is negative for cancer in enlarged and/or hypermetabolic lymph nodes. This is a particularly important consideration if the patient would otherwise have local stage disease and be a candidate for curative resection. (See 'Diagnostic and staging accuracy' below and 'Sampling metastatic disease' below and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient' and 'Combined modalities for mediastinal lymph node staging' below.)

The value of EUS-FNA for the diagnosis and staging of NSCLC is discussed here. The general indications and use of EUS-FNA are discussed separately. (See "Endoscopic ultrasound-guided sampling of the mediastinum: Technique, indications, contraindications, and complications".)

Diagnostic and staging accuracy — Major studies that have examined the diagnostic accuracy of EUS-FNA for NSCLC are summarized below.

●Diagnosis of primary tumor – EUS-FNA can access tumors that invade the posterior and inferior, and occasionally, superior mediastinum that are close to the esophagus. Studies demonstrating its efficacy diagnostically compared to other modalities are lacking. In a series of 35 patients suspected of having lung cancer where bronchoscopy was nondiagnostic, EUS-FNA detected lung cancer in 25 of 26 patients and correctly distinguished benign from malignant disease with sensitivity and negative predictive values of 96 and 90 percent, respectively [77]. Other studies have demonstrated similar results [78].

●Staging – The main use of EUS-FNA is to access paratracheal (2, 4) and posterior subcarinal lymph nodes (3, 7, 8, and 9) (figure 1) [79]. Its ability to access station 5 is dependent on specific anatomic relationships as well as the size of the target lymph node.

A randomized trial of suspected lung cancer patients with 2L, 3, 4L, 7 lymphadenopathy or lung masses adjacent to the mediastinum compared the diagnostic rate of EUS-FNA with EBUS-TBNA. EUS-FNA established a specific diagnosis in 48 out of 55 patients (87 percent) compared to 50 out of 55 patients (91 percent) in the EBUS-TBNA group, a difference that was not statistically significant. Procedure time was shorter with EUS-FNA and additional equipment was not necessary since both the EUS and EBUS procedures were performed using an ultrasound bronchoscope [80].

One 2013 meta-analysis of 26 studies showed pooled sensitivity and negative predictive values of 89 and 86 percent, respectively [4]. However, the wide range of sensitivity (45 to 100 percent) and negative predictive values (68 to 100 percent) suggest high variability among centers and/or heterogeneous patient samples. Three to five passes per nodal site is suggested for adequate sampling [79].

Case series and observational studies also suggest that EUS-FNA can diagnose M1b disease due to its unique ability to access retroperitoneal and celiac nodes, the left lobe of the liver, and adrenal glands [74-77,81-87]. However, its diagnostic accuracy in this context for staging NSCLC has not been formally evaluated.

Tumor that invades the mediastinum, heart, great vessels, trachea, recurrent laryngeal nerve, esophagus, vertebral body, and/or carina, constitutes T4 disease, and in most cases is inoperable (table 1 and table 2). The preoperative identification of T4 disease can potentially result in the avoidance of futile thoracotomies. Although many case series of EUS-FNA have included patients with T4 level of disease, only one study examined its performance for this indication [88]. In this small series from a single tertiary medical center, EUS did not accurately identify mediastinal invasion (false positive rate 30 percent) such that we typically confirm mediastinal invasion by other methods before determining whether or not it is resectable.

Additional trials are required to compare the performance of EUS-FNA to all staging tools, including cervical mediastinoscopy. Smaller comparative studies are described below:

●One prospective, single center study of 104 patients reported that EUS-FNA diagnosed metastatic NSCLC in posterior lymph nodes with a sensitivity and negative predictive value of 93 and 94 percent, respectively [81].

●In a single-center trial of 60 patients with NSCLC, compared to cervical mediastinoscopy, EUS-FNA was able to detect lymph node involvement with higher sensitivity (96 versus 24 percent) [89].

●In a small study of 35 patients with NSCLC, EUS-FNA identified occult N2 or N3 disease, (not previously identified by cervical mediastinoscopy), in 37 percent of cases [90]. Most of the occult disease that was identified by EUS-FNA and missed by cervical mediastinoscopy was in the subcarinal region (7) [90].

Combined modalities for mediastinal lymph node staging — Combining modalities maximizes access to the mediastinum to improve the sensitivity of staging. More sensitive staging identifies occult disease so that multimodality treatment can be considered. The most common staging modalities that are combined are endobronchial ultrasound-guided transbronchial needle aspiration plus endoscopic ultrasound-guided fine needle aspiration (EBUS-TBNA plus EUS-FNA). This is colloquially referred to as "medical mediastinoscopy" or combined ultrasound (CUS). Less common are the combinations of EUS-FNA plus cervical mediastinoscopy and CUS plus cervical mediastinoscopy.

●EBUS-TBNA plus EUS-FNA, can access almost all lymph node stations in the mediastinum (2, 3, 4, 7, 8, 9, 10, 11, +/- 12) (figure 1). Both procedures overlap for many mediastinal lymph nodal stations (2, 3, 4, and 7) but are complimentary for stations 8, 9 (EUS), 10, 11, +/- 12 (EBUS).

●EUS-FNA plus cervical mediastinoscopy can access more mediastinal lymph node stations than either procedure alone (1, 2, 3, 4, 7, 8, 9) (figure 1). Both procedures overlap for reaching lymph nodal stations 2, 3, 4, and 7, but EUS-FNA offers the additional ability to access stations 8 and 9. Combined modalities that utilize EUS-FNA have the advantage of also accessing potential M1b disease in the retroperitoneal, celiac lymph nodes as well as the liver and adrenal glands [74-76,79].

●Combined ultrasound plus cervical mediastinoscopy can access almost all lymph node stations in the mediastinum (1, 2, 3, 4, 7, 8, 9, 10, 11, +/- 12) with overlap for stations 1, 2, 4, and 7 (figure 1).

Although combined approaches show improved sensitivity, their clinical utility in general practice has not been clearly defined. For most patients, the identification of nodal tumor metastasis in N2 or N3 stations using a single technique makes the second procedure unnecessary. Combined modalities are not routinely planned for staging NSCLC unless the initial procedure is negative and the clinical probability of nodal metastasis remains unacceptably elevated. Routine use of a combined modality for staging should be reserved for select cases, best dictated by the expertise and clinical protocols in specific institutions where they are available.

The sensitivity of combining modalities pertinent to staging NSCLC is discussed here. The value of individual techniques for the diagnosis and staging of NSCLC are discussed separately. (See 'Endobronchial ultrasound' above and 'Transesophageal endoscopic ultrasound' above.)

Diagnostic and staging accuracy — Major studies that have examined the diagnostic accuracy of combined modalities for NSCLC are summarized below [22,82,91-97].

●EBUS-TBNA plus EUS-FNA – One multicenter randomized study of 241 patients with suspected N2/N3 NSCLC reported that compared with cervical mediastinoscopy alone, combined endoscopic-endobronchial ultrasonography (CUS) plus selective mediastinoscopy (when CUS results were negative) had a higher sensitivity (94 versus 79 percent) and resulted in fewer unnecessary thoracotomies (7 versus 18 percent) [91]. However, a post-hoc analysis of long-term outcomes from this trial reported no survival advantage among those staged using CUS plus selective surgical staging compared with those who underwent surgical staging alone [98]. Another randomized trial of 160 cases of histologically confirmed or strongly suspected potentially operable NSCLC found that adding EBUS-TBNA to EUS-FNA increased the sensitivity (92 versus 60 percent) and diagnostic accuracy (97 versus 87 percent) when compared with EUS-FNA alone [99]. Another prospective study of 229 patients with suspected resectable NSCLC reported that compared with PET-CT-guided EBUS, the addition of EUS-FNA using the same bronchoscope increased the sensitivity of diagnosing N2/N3 disease from 77 to 84 percent, thereby improving the negative predictive value from 84 to 97 percent [100].

Although these modalities are complimentary, the sensitivity appears to be reduced in patients with normal-sized lymph nodes. Two prospective studies of CUS in patients with enlarged (138 patients) and normal-sized (120 patients) lymph nodes, respectively, reported that, compared with either procedure alone, CUS had higher sensitivity for staging NSCLC (93 percent for enlarged nodes versus 68 percent for normal-sized nodes) [22,93]. The negative predictive values were unaffected by node size (97 and 91 percent respectively).

One retrospective study of patients with NSCLC compared the staging performance of combined EBUS/EUS with an aggressive surgical staging modality, transcervical extended mediastinal lymphadenectomy (TEMLA) [101]. Although CUS was not as sensitive as TEMLA (96 versus 88 percent), it was a safer procedure. This study is discussed in detail separately. (See 'Transcervical extended mediastinal lymphadenectomy' below.)

Bronchoscopists and thoracic surgeons may introduce EBUS bronchoscopes into the esophagus during the same procedure as bronchoscopic EBUS-TBNA and achieve more accurate mediastinal node sampling and greater diagnostic success [80,94,95]. This approach has been termed EUS with bronchoscope-guided needle aspirate (EBUS-B-NA or EUS-B-FNA). EBUS-B-NA was reported to have sensitivity and negative predictive values of 96 and 96 percent, respectively [95].

●EUS-FNA plus cervical mediastinoscopy – There is a paucity of data regarding the staging accuracy of this combination. In one multicenter study of 107 patients with suspected NSCLC, cervical mediastinoscopy plus EUS-FNA detected more metastatic disease than either mediastinoscopy or EUS-FNA alone (36 versus 20 and 28 percent) and resulted in the avoidance of unnecessary thoracotomies in 16 percent of cases [82].

●EBUS-TBNA plus EUS-FNA plus cervical mediastinoscopy – One multicenter prospective study that primarily examined the diagnostic accuracy of CUS, reported a sensitivity of 94 percent for CUS plus cervical mediastinoscopy [91].

Together, these and other studies suggest that combined approaches have a high sensitivity for staging mediastinal NSCLC, especially when disease is suspected by CT or PET scan. However, other clinical priorities and objectives, such as reducing the number of invasive procedures (see 'Surgical staging procedures' below), remain clinically important goals. Based on available evidence, clinical practice guidelines from the European Society of Gastrointestinal Endoscopy endorse an approach of combined EBUS-TBNA and EUS-FNA (or EUS-B-FNA) over either test alone [19]. Larger randomized studies would be required to determine which patients are most likely to benefit from a combined approach and to specify which combination is preferred for staging patients with NSCLC based on accuracy, safety, and other factors (eg, cost effectiveness).

The limitations of EBUS-TBNA, EUS-FNA, and the complications of cervical mediastinoscopy are discussed separately. (See 'Limitations' above and 'Endobronchial ultrasound' above and "Surgical evaluation of mediastinal lymphadenopathy", section on 'Complications of mediastinoscopy'.)

SURGICAL STAGING PROCEDURES

Overview — Standard cervical mediastinoscopy (SCM), video-assisted thoracoscopic surgery (VATS), anterior mediastinotomy (Chamberlain procedure), and robotic-assisted thoracic surgery (RATS) are common surgical modalities used for staging NSCLC. Other surgical procedures (extended cervical mediastinoscopy [ECM], video-assisted mediastinal lymphadenectomy [VAMLA], transcervical extended mediastinal lymphadenectomy [TEMLA]) are not as well validated, and experience is more limited. Selecting one of these surgical procedures relies on physician judgment and knowledge of their diagnostic accuracy for the target lesion(s) in the context of operator proficiency, patient safety, and eventual goals for treatment. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient'.)

Guidelines issued by the American College of Chest Physicians (ACCP), European Society of Thoracic Surgeons (ESTS), and International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society (IASLC/ATS/ERS) have incorporated the use of invasive procedures in the diagnostic work-up of NSCLC. Evidence for these surgical procedures is largely derived from older retrospective case series and small studies that do not compare performance with thoracotomy. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient'.)

In general, the following applies:

●SCM is a sensitive staging procedure when sampling nodal stations easily accessed by this approach (1, 2, 3, 4, anterior 7, 10) (figure 1) [102-104]. Use of a video mediastinoscope and systematic lymph node sampling may enhance sensitivity. Although it is the historical gold standard for staging the mediastinum, disease will not be detected in lymph nodes stations that are not accessed by this modality: subaortic (5), para-aortic (6), inferior (8, 9), lobar/interlobar (11 to 14), and, occasionally, the posterior subcarinal (7) stations. SCM can be used as a staging tool for NSCLC when other staging modalities do not identify cancer, or are not available (eg, endobronchial ultrasound). (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient' and 'Standard cervical mediastinoscopy' below.)

●VATS is an accurate modality for evaluating the extent of invasion (chest wall, mediastinal) by the primary tumor (T). VATS can also sample mediastinal lymph nodes (4, 5, 6, 7, 8, 9, 10 to 14) but only those on the affected side (N1/N2), unless a bilateral thoracotomy is performed (figure 1). Bilateral thoracotomy is not typically performed due to the increased mortality and morbidity associated with this procedure. VATS is also useful for detecting pleural involvement (M1a) (table 3). VATS is typically used for tissue diagnosis when there is a suspicious peripheral lesion that is amenable to wedge resection and there is no computed tomographic (CT) and/or positron emission (PET) evidence for extrathoracic or mediastinal nodal metastasis. VATS may also be used when percutaneous or endoscopic procedures cannot access the primary tumor or are non-diagnostic [105-108]. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient' and 'Video-assisted thoracic surgery and robotically assisted thoracic surgery' below.)

●RATS is increasingly utilized in early-stage lung cancer, both for lymph node staging and resection of the primary tumor. Although many aspects of RATS overlap with VATS, technological advancements and the development of specific robotic techniques have led to improved three-dimensional optics, ergonomics, and capability in confined spaces compared with VATS. (See "Overview of minimally invasive thoracic surgery" and 'Video-assisted thoracic surgery and robotically assisted thoracic surgery' below.)

●Left anterior mediastinotomy (Chamberlain procedure) is often the only option for sampling the subaortic (aortopulmonary window) (5) and para-aortic (6) lymph node stations (although some endoscopic ultrasound techniques can reach the AP window). (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient' and 'Anterior mediastinotomy (Chamberlain procedure)' below.)

●ECM [109-111], VAMLA [112,113], and TEMLA [101,114-116], are surgical procedures that allow more extensive mediastinal sampling. Although they are sensitive procedures for detecting mediastinal disease, their use is limited because they are not widely available and are associated with increased morbidity when compared with SCM (eg, recurrent laryngeal nerve paralysis). (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Approach to the patient' and 'Extended cervical mediastinoscopy' below and 'Video-assisted mediastinal lymphadenectomy' below and 'Transcervical extended mediastinal lymphadenectomy' below.)

The evidence for each invasive procedure pertinent to its diagnostic and staging accuracy for NSCLC is discussed in this section. The surgical management of mediastinal lymphadenopathy and procedural aspects of VATS are discussed separately. (See "Surgical evaluation of mediastinal lymphadenopathy" and "Overview of minimally invasive thoracic surgery".)

Standard cervical mediastinoscopy — SCM is a surgical procedure that accesses the mediastinum and as such is primarily a mediastinal staging tool. Its major advantage, compared to minimally invasive tools (eg, endobronchial ultrasound), is direct visualization of nodes for sampling and/or dissection. In addition, use of biopsy forceps provides sufficient amounts of material for diagnosis and immunohistochemical and molecular analysis.

SCM is the historical gold standard for mediastinal lymph node staging of NSCLC with reported sensitivities of 78 to 90 percent [4,102-104]. Traditionally, this procedure has not used video-assistance for sampling lymph nodes and has sampled nodes by biopsy forceps. Much of the evidence describing its performance as a staging modality relates to this particular approach. However, there has been a paradigm shift in the use of a video-assisted mediastinoscope to access the mediastinum. Video mediastinoscopy improves operator visualization, so that systematic lymph node sampling (by biopsy forceps) can be performed. When a video mediastinoscope is used to dissect and remove an entire lymph node, this is known as VAMLA which is discussed below. (See 'Video-assisted mediastinal lymphadenectomy' below.)

SCM can easily access pretracheal (1, 3), paratracheal (2R, 2L, 4R, 4L), anterior subcarinal (7) and occasionally hilar (10) nodes (figure 1). It is not able to sample subaortic (5) or para-aortic (6), inferior (8, 9), posterior subcarinal (7), or lobar/interlobar (11 to 14) stations. Use of the video mediastinoscope can extend access to the posterior subcarinal nodes (station 7). SCM is best used to sample lymph nodes in these selected stations, particularly when other staging modalities have failed or are not available.

Evidence supporting the role of SCM in staging the mediastinum in patients with suspected NSCLC is reviewed here. The technical aspects, complications of mediastinoscopy, and comparative studies with other staging modalities are discussed separately. (See "Surgical evaluation of mediastinal lymphadenopathy", section on 'Complications of mediastinoscopy' and 'Diagnostic and staging accuracy' above and 'Endobronchial ultrasound' above and 'Diagnostic and staging accuracy' above and 'Diagnostic and staging accuracy' above.)

Staging accuracy — The largest meta-analysis evaluating the accuracy of SCM pooled data from 9267 patients with suspected NSCLC who underwent SCM for staging the mediastinum [4]. This analysis reported that SCM had a median sensitivity and negative predictive value of 78 and 91 percent, respectively. However, the sensitivity is wide ranging (32 to 92 percent). Additional techniques that may enhance the accuracy of SCM include:

●Extensive sampling and complete lymph node dissection may increase the accuracy of staging with a reported sensitivity of 90 percent [103].

●Systematic lymph node sampling is when the operator systematically samples at least one node from each accessible lymph node station (typically up to five stations). Compared to selective node sampling where lymph nodes are randomly selected by the operator, systematic lymph node sampling may result in increased sensitivity (up to 89 percent) [3,4,103,104,117,118].

There are relatively few comparative studies of endobronchial ultrasound-guided transbronchial needle aspiration or transesophageal endoscopic ultrasound-guided fine needle aspiration, with surgical modalities. (See 'Diagnostic and staging accuracy' above and 'Endobronchial ultrasound' above and 'Combined modalities for mediastinal lymph node staging' above.)

●A comparative study of 127 patients who completed both endobronchial ultrasound-transbronchial needle aspiration (EBUS-TBNA) followed by mediastinoscopy followed by open surgical resection and complete lymph node dissection for all patients with negative initial results demonstrated a sensitivity, specificity, and accuracy of EBUS-TBNA of 88, 100, and 92.9 percent compared to 81, 100, and 89.0 percent for mediastinoscopy respectively [42].

●In a highly selected group of 159 patients who underwent EBUS-TBNA and cervical mediastinoscopy, the sensitivity of both procedures was comparable (81 versus 79 percent) and there were no significant differences between the negative predictive value and overall diagnostic accuracy of the procedures (93 versus 93 percent) [41].

One limitation of these studies is that they are based on data from small cohorts. In addition, up to half of cases that are reported as falsely negative are due to the inability of SCM to access certain lymph node stations [119-124]. The variable sensitivity may also be due to poor adherence to guidelines for the performance of SCM [125,126]. One study of practice variation at four hospitals showed that guidelines for the use of mediastinoscopy in patients with suspected lung cancer, between the years of 1993 and 1999, were adequately followed in only two-thirds of cases [126]. Approximately, 17 percent of patients appeared to have unforeseen N2/3 disease, resulting in 18 percent of cases in whom thoracotomy could have been prevented had guidelines been followed [126]. However, whether this is reflective of practice in most institutions is not known.

Cervical mediastinoscopy is a surgical procedure. Potential complications include hemorrhage, vocal cord dysfunction, tracheal injury and pneumothorax [127]. The complications of cervical mediastinoscopy are discussed separately. (See "Surgical evaluation of mediastinal lymphadenopathy", section on 'Complications of mediastinoscopy'.)

Video-assisted thoracic surgery and robotically assisted thoracic surgery — VATS and RATS are typically used in the diagnosis and staging of suspected NSCLC when alternative procedures cannot access the tumor or are nondiagnostic. Both techniques may be used as a primary diagnostic modality when the probability of cancer is high and the tumor is easily accessible to thoracoscopic wedge resection.

The main advantage of these minimally invasive surgical techniques is the direct visualization of and extensive access to the ipsilateral lung, thoracic cavity, and all mediastinal structures (eg, lymph nodes, vessels, and esophagus). This allows concurrent access to both the primary tumor and to almost all lymph nodes in the ipsilateral mediastinum (4, 5, 6, 7, 8, 9, 10 to 14) (figure 1). These techniques also allow reliable evaluation of the extent of invasion by the primary tumor (T), especially that of the chest wall and mediastinal structures, malignant mediastinal lymph node involvement (N), and pleural involvement with nodules or effusion (M1a) (table 1 and table 2) [105,106]. Minimally invasive surgical approaches permit rapid intraoperative diagnosis, staging and occasionally therapy, thereby shortening the time needed to diagnose and stage NSCLC.

VATS and RATS are major surgical procedures requiring general anesthesia and single-lung ventilation. Reaching the right side of the mediastinum with VATS is technically easier than reaching the left, particularly, the left paratracheal nodes (4L). In addition, although access to mediastinal node stations is extensive, typically only one side can be sampled. Occult disease on the contralateral side can be missed unless bilateral procedures are performed. However, bilateral procedures are not routinely performed for staging because this may greatly increase the morbidity and mortality of the surgery. The complications and procedural aspects of minimally invasive thoracic surgery and surgical management of mediastinal lymphadenopathy are discussed separately. (See "Surgical evaluation of mediastinal lymphadenopathy" and "Overview of minimally invasive thoracic surgery".)

Staging accuracy of VATS — One 2013 meta-analysis of four small studies reported a median sensitivity of 99 percent (range 58 to 100 percent) and negative predictive value of 96 percent (range 88 to 100 percent) [4]. However, of the four studies included in this meta-analysis, the study that reported the lowest sensitivity (58 percent) was a multicenter study. The variability of results and low sensitivity seen in the multicenter study suggest wide differences in the accuracy of VATS for staging NSCLC or heterogeneity of patient populations. There are no studies that directly compare the diagnostic or staging accuracy of VATS to other staging modalities.

VATS has been shown to have value in the evaluation of T4 lesions. While some T4 lesions may be potentially resectable (eg, large, >7 cm and surrounded by normal lung with no nodal metastasis), most are not resectable (eg, large bulky tumors infiltrating the mediastinum) [107,128,129]. Further staging by VATS in this setting may either downstage the disease such that the patient may become a candidate for surgery or upstage it when unexpected metastatic disease of the pleura is found [106,107,129-132]. The best evidence to support this comes from one prospective study of 64 patients with NSCLC staged using the 7^th edition staging classification system (table 3) [107]. All patients were thought be ineligible for surgery based on their computed tomography (CT) scan characteristics (eg, pleural fluid, T4 lesions). Among the 30 patients with mediastinal infiltration on CT scan, nine (30 percent) were down-staged to potentially resectable status and two (7 percent) to resectable disease because mediastinal invasion was not confirmed. In addition, this study and others have reported that VATS can upstage the disease to stage IV by the identification of unsuspected pleural disease in 6 to 60 percent of cases [106,107].

VATS is frequently used to perform lobectomies in patients with peripheral nodules and suspected N0 or N1 disease. For accurate staging, these patients are routinely subjected to mediastinal lymph node sampling intraoperatively. Two controversial issues arise intraoperatively:

●Unexpected occult N2 disease – Typically, surgery is not offered as first line therapy to patients with preoperative evidence of N2 disease. However, if occult N2 disease is discovered intraoperatively, most surgeons will proceed with resection if the diseased node can be fully resected and adjuvant chemotherapy is subsequently offered [133]. (See "Management of stage I and stage II non-small cell lung cancer", section on 'Video-assisted thoracoscopic surgery'.)

●Extent of lymph node dissection – It is unknown whether lymph nodes should be sampled systematically or fully dissected intraoperatively. One randomized study of 1111 patients with NSCLC who had negative preoperative sampling of lymph nodes who also underwent VATS suggested that, compared with lymph node sampling, lymph node dissection discovered more occult N2 disease (4 versus 1 percent) [134]. However, dissection was not associated with a survival advantage (median survival 8.5 years in both groups) or decrease in local or regional recurrence. (See "Management of stage I and stage II non-small cell lung cancer", section on 'Mediastinal lymph node dissection'.)

Staging accuracy of RATS — Since 2011, the utilization of RATS appears to be increasing faster than VATS [135]. When compared with VATS, RATS for lobectomy has similar or superior safety, short- and long-term outcomes (including operative time), and mortality (both overall survival and disease-free survival) [135-138].

●A single center open-label, randomized trial of RATS versus VATS in 320 patients demonstrated a higher number of lymph nodes harvested (11 versus 10) and more lymph node stations examined (six versus five), but no differences in hospital stay or perioperative complications, including mortality [139].

●A single center retrospective cohort study of 246 patients between 2010 and 2018 with clinical N0 status comparing RATS with total thoracoscopic approach demonstrated a higher total lymph node count and lymph node stations examined without differences in complication rates [140].

Anterior mediastinotomy (Chamberlain procedure) — A left anterior mediastinotomy (Chamberlain procedure) is a surgical procedure that requires general anesthesia. The incision is made over the left second or third intercostal space to access lymph nodes in stations 5 and 6 (figure 1). Due to lymphatic drainage patterns, cancers of the left upper lobe preferentially drain to these regions (figure 1). However, these regions are traditionally difficult to access and cannot be reached by minimally invasive techniques or by standard cervical mediastinoscopy.

A left anterior mediastinotomy is the traditional surgical procedure used to stage left-sided cancer when suspected nodes are located in stations 5 and 6. It has the added advantage of accessing left upper lobe tumors for concurrent resection if there is no evidence of nodal or distant disease.

Because it is often the only option when suspicious lymph nodes are identified in this region, there is a paucity of data regarding its diagnostic accuracy. Nonetheless, one meta-analysis of four small studies reported a sensitivity and negative predictive value of 71 and 91 percent respectively [141]. No studies have directly compared its performance to ECM, VATS, or to endoscopic ultrasound-guided fine needle aspiration which can occasionally reach this region. (See 'Extended cervical mediastinoscopy' below and 'Video-assisted mediastinal lymphadenectomy' below.)

Extended cervical mediastinoscopy — ECM accesses the mediastinum and as such is primarily a mediastinal staging tool. It extends the sampling territory of standard cervical mediastinoscopy (1, 2, 3, 4, 7, 10) to the subaortic (5) and para-aortic (6) lymph node stations (figure 1). Stations 5 and 6 cannot be reached using standard cervical mediastinoscopy. Traditionally, accessing nodes in these stations requires surgical dissection via a left anterior mediastinotomy (Chamberlain procedure). Thus, the main advantage of ECM is that when access to stations 5 and 6 is desired to stage tumors of the left upper lobe, it avoids the surgical risk associated with a left anterior mediastinoscopy. The main disadvantage of ECM is that its performance is limited to the few centers with expertise for it as a staging modality.

Two case series of 377 patients with suspected cancer of the left lung compared the staging accuracy of routine and selective ECM [109,111]. ECM was considered routine when it was performed in every patient with suspected NSCLC, and considered selective when biopsy was only performed in patients with suspicious nodes identified by CT or positron emission tomography (PET). Compared to routine ECM, selective ECM had higher sensitivity (65 to 75 versus 44 to 45 percent) but a similar negative predictive value (95 versus 94 percent), although these results may be explained by spectrum bias (eg, lymph nodes were larger and/or more likely to be involved in the ECM group). The complication rate was 2 percent and included mediastinitis, ventricular fibrillation, and wound infection.

Another retrospective analysis of 55 patients with NSCLC reported that, compared with PET/CT, ECM had a higher sensitivity (69 versus 53 percent), and negative predictive value (89 versus 83 percent) [110]. ECM is not widely available and requires further study.

Video-assisted mediastinal lymphadenectomy — Video-assisted mediastinoscopes offer improved visualization and extensive sampling of mediastinal lymph nodes for complete removal (lymphadenectomy) (figure 1). VAMLA can sample the same lymph node stations as standard cervical mediastinoscopy (1, 2, 3, 4, anterior 7, and 8). In addition, they provide better access to the posterior region of the subcarinal lymph node station (7).

●One single-center study of 144 patients reported a sensitivity for VAMLA of 100 percent [112].

●A case series of 156 patients reported that, compared to standard mediastinoscopy, VAMLA was superior with higher sensitivity (94 versus 64 percent) and negative predictive value (96 versus 84 percent) [113].

●Another multicenter series of patients with clinical N1 disease who underwent VAMLA has a sensitivity of 73 percent for the detection of N2 disease when compared with standard video-assisted mediastinoscopy [142].

In most studies, the complication rate was high (3 to 6 percent) with a higher than usual rate of recurrent laryngeal nerve injury (4 to 5 percent) [112,113].

VAMLA is not widely available and requires further study.

Transcervical extended mediastinal lymphadenectomy — TEMLA uses a sternal retractor to widen surgical access to the mediastinum for open (not endoscopic or video-assisted) dissection. This open access approach allows sampling of stations 1, 2, 3, 4, 5, 6, 7, and 8 (figure 1).

One prospective study of 81 patients showed sensitivity and negative predictive values of 90 and 95 percent, respectively [114]. Compared to standard cervical mediastinoscopy, a prospective randomized study of 41 patients reported that TEMLA had higher sensitivity (100 versus 38 percent) and negative predictive values (100 versus 67 percent) [115]. However, a major caveat of this study was its early termination due to the unusually high false negative rate in the cervical mediastinoscopy group. The recurrent laryngeal nerve palsy rate appears lower than for VAMLA, at 2 percent with less than 1 percent having permanent damage [116].

One retrospective study of patients with NSCLC compared the staging performance of endoscopic modalities (EBUS-TBNA, EUS-FNA, and combined EBUS/EUS) with TEMLA [101]. In the 276 patients where TEMLA was preceded by negative endoscopic sampling for primary staging, TEMLA discovered occult disease in 50 patients and deferred surgery in three patients. Compared to endoscopic modalities, TEMLA had higher sensitivity (96 versus 88 percent) and negative predictive values (99 and 83 percent, respectively). However, the increased sensitivity of TEMLA occurred at the expense of increased morbidity (7.2 versus 6.4 percent), mostly recurrent laryngeal nerve palsy (3 versus 0 percent), and mortality (0.4 versus 0 percent). (See 'Combined modalities for mediastinal lymph node staging' above.)

Additional large randomized multicenter trials will be needed to validate TEMLA further before it can be routinely used as a staging procedure for NSCLC.

SAMPLING METASTATIC DISEASE — Sites of distant metastases in NSCLC are classified into intrathoracic and extrathoracic metastases. Common intrathoracic sites are the pleura (nodules, effusion), contralateral lung (nodule), and/or pericardium (nodules, effusion) (stage IV; M1a). Common extrathoracic sites are the liver, adrenal gland, brain and bone (stage IV; M1b when isolated and M1c when multiple) (table 1 and table 2). An important exception is that a tumor in a supraclavicular or scalene node is considered to be N3 (stage IIIB or IIIC) disease, not stage IV disease despite the fact that their location is extrathoracic. Tissue confirmation of suspected disease at any of these sites indicates NSCLC that is inoperable.

Discovery of sites involved with metastatic NSCLC are often prompted by clinical assessment and/or imaging. The imaging modalities should direct site selection for tissue confirmation of metastatic NSCLC. Typically, consideration is given to sampling suspected metastatic sites first rather than sampling the suspected primary lesion. This is especially important for solitary metastases when tissue diagnosis is essential to establish clear evidence of inoperable NSCLC. When there is evidence on imaging to support a high probability of multiple metastases, the site of tissue biopsy is selected based upon the ease of access and risks of the procedure.

The accuracy of specific procedures used to biopsy suspected metastases from NSCLC is discussed here. Imaging and approach to tissue biopsy of potential metastatic disease are discussed separately. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Suspected advanced disease' and "Overview of the initial evaluation, diagnosis, and staging of patients with suspected lung cancer", section on 'Clinical-directed imaging'.)

Intrathoracic — Metastatic NSCLC in the pleura, contralateral lung, or pericardium constitutes stage M1a disease, which is inoperable.

●Pleura – Two common clinical presentations of pleural disease are: metastases associated with pleural effusion or multiple pleural-based nodules, and direct extension of the primary tumor to the pleura or chest wall. Direct extension of the primary tumor into the visceral or parietal pleura needs to be distinguished from metastatic NSCLC of the pleural space that is separate from the primary tumor. Primary tumor that infiltrates visceral or parietal pleura is staged as a T2 or T3 lesion, respectively (table 1 and table 2). Some T2/T3 lesions are potentially resectable. In contrast, metastatic NSCLC discovered as separate solid lesion(s) in the visceral or parietal pleura or as malignant cells in an effusion represent M1a disease that is not resectable. The diagnostic accuracy of sampling pleural lesions is discussed below. (See 'Suspected pleural metastases' below.)

●Lung – Solitary nodules or masses in the contralateral lung require histologic evaluation to distinguish metastasis from a synchronous primary malignant tumor or an unrelated non-malignant diagnosis. NSCLC with a histologically similar contralateral solitary lesion is classified as M1a disease. The diagnostic accuracy for sampling nodules or masses suspicious for metastatic disease in the contralateral lung is the same as sampling a primary tumor or a solitary pulmonary nodule, which are discussed separately. (See 'Endoscopic and percutaneous procedures' above and 'Video-assisted thoracic surgery and robotically assisted thoracic surgery' above and "Diagnostic evaluation of the incidental pulmonary nodule", section on 'Nonsurgical biopsy' and "Diagnostic evaluation of the incidental pulmonary nodule", section on 'Surgical biopsy'.)

●Pericardium – Direct extension of the tumor to the parietal pericardium constitutes a T3 lesion and visceral pericardium is a T4 lesion (table 1 and table 2). A pericardial effusion due to NSCLC is M1a disease. Pericardial NSCLC with invasion of the visceral pericardium or heart itself and all cases of pericardial effusions due to NSCLC are considered unresectable. Sampling pericardial fluid or tissue is discussed separately. (See "Pericardial disease associated with cancer: Clinical presentation and diagnosis".)

Suspected pleural metastases — Pleural effusions or pleural abnormalities (solid masses/nodule/thickening) suspicious for metastatic disease are present on imaging in up to one-third of patients with NSCLC at the time of presentation [143]. Although some effusions may be benign (eg, due to post-obstructive pneumonitis or atelectasis), the majority are malignant and establish a diagnosis of M1a disease (table 1 and table 2). Sampling the pleural space (effusions or solid lesions) is essential in all patients with suspected pleural involvement. Thoracentesis is typically performed first for evaluation of pleural effusions. Pleural biopsy (image-guided or thoracoscopic) is often performed when pleural fluid cytology is negative or to obtain tissue from a solid pleural lesion.

The diagnostic accuracy of thoracentesis and pleural biopsy techniques for NSCLC are discussed here. The imaging modalities (ultrasound, computed tomography [CT], positron emission tomography [PET]) and approach to suspected pleural disease in patients with NSCLC are discussed separately. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Suspected advanced disease' and "Overview of the initial evaluation, diagnosis, and staging of patients with suspected lung cancer", section on 'Suggested approaches to diagnostic evaluation and radiographic staging' and "Imaging of pleural effusions in adults" and "Imaging of pleural plaques, thickening, and tumors".)

Thoracentesis-cytology — Thoracentesis removes fluid from the pleural space for cytologic analysis. The diagnostic accuracy of pleural fluid cytology is derived from older, small studies. In addition, many of these studies report the diagnostic yield in the setting of benign and malignant disease of varied histologic types [144-150]. Nonetheless, two meta-analyses reported a mean sensitivity for the diagnosis of malignancy of 72 percent (49 to 91 percent) [2] and 59 percent [151]. It is the imperfect sensitivity of thoracentesis that necessitates pleural biopsy when fluid cytology is negative. (See "Ultrasound-guided thoracentesis" and "Pleural fluid analysis in adults with a pleural effusion" and "Clinical manifestations of lung cancer", section on 'Pleural involvement' and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Suspected advanced disease'.)

The sensitivity of pleural fluid cytology increases with recurrent sampling. Up to 30 percent of second thoracenteses are positive for cancer when the results of the initial thoracentesis are falsely negative [147,148,152]. Other studies report variable increases in the rates of cancer diagnoses from 5 to 25 percent after the third thoracentesis [144,147].

The sensitivity also varies with tumor types. In one meta-analysis of 36 studies, the sensitivity was high for lung adenocarcinoma (83.6 percent; 95% CI 77.7-89.6 percent) and ovarian cancer (85.2 percent; 95% CI 74.2-96.1 percent), modest for breast cancer (65.3 percent; 95% CI 49.8-80.8 percent) and low for lung squamous cell carcinoma (24.2 percent; 95% CI 17-31.5 percent) and mesothelioma (28.9 percent; 95% CI 16.2-41.5 percent) [151].

There is conflicting data regarding the optimal volume required for the identification of malignant cells in pleural fluid. Studies are hampered by varied definitions of what is considered large volume versus small volume and lack comparators. While one study suggested higher sensitivity from samples >60 mL, particularly those >150 mL, another study suggested that 50 mL specimens are adequate [153,154]. Two studies also reported higher sensitivity when enough material was collected for the preparation of a cell block [148,153]. Typically, for cell block preparation, higher volumes are needed. Although undetermined, it has been suggested that the variability in optimal volume required for diagnosis may simply relate to the total number of malignant cells in the sample pleural fluid, a variable which is unknown at the time of thoracentesis. Thus, the optimal volume removed for cytologic analysis for thoracentesis varies among institutions and interventional clinicians. In general, when feasible, volumes of at least 50 to 75 mL processed by cell block preparation are recommended. (See "Ultrasound-guided thoracentesis", section on 'Technique' and "Ultrasound-guided thoracentesis", section on 'Complications' and "Large volume (therapeutic) thoracentesis: Procedure and complications" and "Measurement and interpretation of pleural pressure (manometry): Indications and technique".)

There has been a paradigm shift with the increased use of bedside thoracic ultrasound (TUS) to guide the operator in performing thoracentesis safely. TUS decreases the incidence of iatrogenic pneumothorax following thoracocentesis [155-157]. The advantage of TUS is that it can facilitate access to small effusions that would not have traditionally undergone bedside manual thoracentesis. The diagnostic accuracy of TUS for identifying malignant effusions in this context is uncertain. However, it is typically used by many clinicians for reasons of safety rather than for diagnostic certainty. (See "Ultrasound-guided thoracentesis".)

Although there is reduced sensitivity and variable assay methodology, it is technically possible to evaluate molecular tumor markers on pleural fluid specimens [158-160].

Pleural biopsy — When pleural fluid cytology is negative or pleural based abnormalities/masses are present, a pleural biopsy is indicated. Three options are available to the clinician to biopsy the pleura: thoracoscopy (surgical or medical), image-guided biopsy (CT or ultrasound), and closed pleural biopsy.

●Surgical or medical thoracoscopy – Surgical thoracoscopy is usually performed in the operating room under video assistance and general anesthesia. Medical thoracoscopy (pleuroscopy) is an endoscopic procedure usually done in the endoscopy suite under conscious sedation. Both procedures are capable of obtaining biopsy material from the parietal and visceral pleura under direct visualization. The operator can randomly sample the pleura as well as biopsy macroscopically visible lesions. Among the three options for pleural biopsy, thoracoscopy has the highest diagnostic accuracy. Reported sensitivities range from 80 to 99 percent with negative predictive values of 93 to 96 percent [102,143,161-166]. Lower sensitivity values may be due to a higher percentage of false negative results in patient cohorts that include mesothelioma, which is typically difficult to diagnose histologically on pleural biopsy [164]. Thoracoscopy has the added advantage of providing therapeutic options to avoid recurrent effusion such as pleurodesis or the insertion of an in-dwelling catheter at the time of surgery. (See "Medical thoracoscopy (pleuroscopy): Equipment, procedure, and complications" and "Medical thoracoscopy (pleuroscopy): Diagnostic and therapeutic applications", section on 'Metastatic lung cancer'.)

●Image-guided biopsy – Image-guided biopsy is performed in the interventional radiology suite with or without sedation. Biopsy material is obtained using CT-guidance, or less commonly, ultrasound-guidance [167-170]. This modality is best utilized to biopsy focal pleural lesions/abnormalities with or without associated effusions. The reported sensitivity ranges from 76 to 88 percent with negative predictive values of 75 to 80 percent [167-169]. The diagnostic utility of the type of cutting needle used (Cope or Abrams needle) has not been comprehensively studied. However, one study of cytology-negative pleural effusions used an Abrams needle specifically to biopsy associated pleural-based masses/thickening under CT-guidance. Compared to thoracoscopy, no difference was observed in the sensitivity for identifying lung cancer (100 versus 93 percent) [163]. Another study reported that CT-guided Abram's needle biopsy had a higher sensitivity than ultrasound-guided biopsy using a cutting needle (82 versus 67 percent) [171].

●Closed pleural biopsy – Closed pleural biopsy is rarely performed for staging NSCLC in contemporary practice. It is a blind procedure typically performed with an Abrams needle. It is a bedside procedure that can only be used when pleural fluid is present and can only biopsy the parietal pleura. Pleural metastases are more common on the visceral pleura and are sparse and focal on the parietal pleura. This feature, together with dwindling utilization and expertise in closed biopsy, may contribute to the low sensitivity of this modality, thereby limiting its utility.

Extrathoracic — Clinical history, examination and imaging typically alert the clinician to suspected metastatic disease distant from the lung. (See "Overview of the initial evaluation, diagnosis, and staging of patients with suspected lung cancer", section on 'Clinical-directed imaging'.)

Histologic confirmation of extrathoracic NSCLC is necessary to establish M1b (isolated metastasis) and M1c (multiple metastases) disease (table 1 and table 2). In contrast, NSCLC in a supraclavicular node represents N3 nodal involvement and is staged at IIIB or IIIC (depending on the T stage). Most forms of extrathoracic disease are inoperable. The following sections discuss the modalities used to biopsy suspected metastases in NSCLC:

●Liver (see "Approach to liver biopsy" and "Evaluation and management of the adrenal incidentaloma", section on 'Fine-needle aspiration biopsy' and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Suspected advanced disease')

●Adrenal gland (see "Evaluation and management of the adrenal incidentaloma", section on 'Evaluation for malignancy' and "Evaluation and management of the adrenal incidentaloma", section on 'Fine-needle aspiration biopsy' and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Suspected advanced disease')

●Brain – Biopsy of brain lesions suspected to be metastatic NSCLC is rarely performed but their evaluation is discussed separately (see "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Suspected advanced disease')

●Bone (see "Bone tumors: Diagnosis and biopsy techniques", section on 'Biopsy techniques' and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Suspected advanced disease')

●Supraclavicular lymph node (see "Evaluation of peripheral lymphadenopathy in adults", section on 'Localized lymphadenopathy' and "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer", section on 'Suspected advanced disease')

SPUTUM CYTOLOGY — Sputum cytology is a noninvasive tool that has diagnostic value in a small population of patients with suspected NSCLC who are unable or unwilling to undergo other diagnostic procedures. However, it does not directly provide staging information for NSCLC, nor it is it likely to provide ideal specimens for immunohistochemical or molecular studies [172].

Pooled data from small observational series report sensitivity values of 66 percent (range 42 to 97 percent) for the diagnosis of NSCLC [2]. Sensitivity varies by location of the primary tumor, being highest for large, centrally located lesions, and lower for smaller or peripheral lesions [2].

Although cytologic specimens can be diagnostic of NSCLC, not every patient produces sputum. Thus, a negative test does not exclude the diagnosis of NSCLC and a second procedure will often be necessary for this purpose. Similarly, when sputum cytology is positive, additional testing will often be necessary for staging, especially when mediastinal or distant metastases are suspected.

Thus, in practice, sputum cytology is typically most useful for patients in whom avoiding a biopsy is desirable (eg, multiple comorbidities, contraindications to invasive biopsy, overwhelming evidence of metastatic disease). In this context, diagnostic sputum cytology may permit palliative or other therapy without the risk associated with biopsy.

Should sputum cytology be positive, chest computed tomography is indicated to look for a culprit lesion. In general, further diagnostic investigations, targeted at the primary lesion or at metastases is typically necessary for diagnostic staging. For those with negative chest imaging, white light bronchoscopy with or without autofluorescence bronchoscopy in indicated. (See "Flexible bronchoscopy in adults: Overview" and "Detection of early lung cancer: Autofluorescence bronchoscopy and investigational modalities" and "Detection of early lung cancer: Autofluorescence bronchoscopy and investigational modalities", section on 'Sputum atypia and normal radiographic imaging'.)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Non-small cell lung cancer (The Basics)")

●Beyond the Basics topics (see "Patient education: Non-small cell lung cancer treatment; stage I to III cancer (Beyond the Basics)" and "Patient education: Non-small cell lung cancer treatment; stage IV cancer (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Scope – There are a number of minimally invasive and invasive procedures that can be used by the clinician to biopsy patients with suspected non-small cell lung cancer (NSCLC). Selecting a modality depends upon knowledge of the diagnostic accuracy of a modality for the target lesion(s), in the context of operator proficiency, patient safety and eventual goals for treatment. While minimally invasive techniques are generally preferred to more invasive procedures, their imperfect sensitivity necessitates additional biopsy when testing is negative for cancer. Sensitivity and specificity estimates for individual modalities are usually based upon data from small studies with inadequate gold standard comparators that lead to bias and significantly limit interpretation of the data. The clinician should be aware of these flaws when choosing among the modalities available for tissue acquisition. (See "Selection of modality for diagnosis and staging of patients with suspected non-small cell lung cancer".)

●Bronchoscopic techniques – Bronchoscopic techniques often employ multiple modalities (eg, forceps biopsy, brushings, ultrasound, needle aspiration) to increase the sensitivity for the diagnosis and staging of NSCLC. Bronchoscopy is best utilized for accessing large, visible, or central lesions and suspicious paratracheal and subcarinal nodes. As examples (see 'Bronchoscopic approaches' above):

•Conventional bronchoscopy with endobronchial washings, brushings, and biopsy is favored for large central tumors, and EBUS-TBNA is typically paired with conventional bronchoscopy to biopsy accessible lymph node stations during the same procedure.

Endobronchial-ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is the first-choice modality for sampling suspicious mediastinal lymph nodes in the paratracheal, subcarinal, and hilar regions (2R, 2L, 3p, 4R, 4L, 7, 10R, 10L, 11R, 11L) (figure 1). Limitations of EBUS-TBNA include inability to access prevascular, periaortic, paraesophageal, or pulmonary ligament lymph node stations (3a, 5, 6, 8, 9), and variable proficiency among operators, which may influence the rate of nondiagnostic findings. (See 'Endobronchial ultrasound' above.)

•Other image-guided bronchoscopic techniques are useful for sampling peripheral lung lesions. (See 'Enhanced, image-guided bronchoscopic techniques' above and "Image-guided bronchoscopy for biopsy of peripheral pulmonary lesions".)

●Percutaneous techniques – Transthoracic (percutaneous) needle aspiration/biopsy (TTNA/B) is a sensitive procedure for acquiring tissue for most intraparenchymal lesions, particularly peripheral masses and nodules. It is often used to access peripheral lesions and can also be used when other procedures have failed. Results are sometimes nondiagnostic and there is a relatively high risk of pneumothorax. (See 'Transthoracic needle biopsy' above.)

●Other endoscopic ultrasound techniques – Endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) is a sensitive staging tool for suspected NSCLC involvement in subcarinal (3, 7, 8, 9) and paratracheal nodes (2, 4). EUS-FNA can be combined with EBUS-TBNA to enhance mediastinal staging. (See 'Transesophageal endoscopic ultrasound' above and 'Combined modalities for mediastinal lymph node staging' above.)

●Cervical mediastinoscopy – Standard cervical mediastinoscopy (SCM) is a sensitive staging procedure when sampling nodal stations easily accessed by this approach (1, 2, 3, 4, anterior 7, 10) (figure 1). Use of a video mediastinoscope and systematic lymph node sampling/dissection may enhance the accuracy of this modality. (See 'Standard cervical mediastinoscopy' above.)

●Video-assisted thoracoscopic surgery – Video-assisted thoracoscopic surgery (VATS) is an accurate modality for evaluating the extent of invasion (chest wall, mediastinal) by the primary tumor (T), mediastinal lymph node involvement (4, 5, 6, 7, 8, 9, 10 to 14) and pleural involvement (M) (figure 1). It is a surgical procedure that requires general anesthesia and carries higher morbidity and mortality than other diagnostic or staging modalities. VATS is most often used when alternative procedures cannot access the primary tumor or are non-diagnostic. (See 'Video-assisted thoracic surgery and robotically assisted thoracic surgery' above.)

●Other surgical procedures

•Robotic-assisted thoracic surgery is increasingly utilized in early-stage lung cancer both for lymph node staging and resection of the primary tumor. (See 'Video-assisted thoracic surgery and robotically assisted thoracic surgery' above.)

•Left anterior mediastinotomy (Chamberlain procedure) is often the only option for sampling the subaortic (5) and para-aortic (6) lymph node stations. (See 'Anterior mediastinotomy (Chamberlain procedure)' above.)

•Extended cervical mediastinoscopy (ECM), video-assisted mediastinal lymphadenectomy (VAMLA), and transcervical extended mediastinal lymphadenectomy (TEMLA) are surgical procedures that allow more extensive mediastinal sampling than standard cervical mediastinoscopy. They are not widely available and may be associated with greater morbidity than standard cervical mediastinoscopy. Using any one of these procedures is best dictated by the expertise of practitioners within the institution. (See 'Extended cervical mediastinoscopy' above and 'Video-assisted mediastinal lymphadenectomy' above and 'Transcervical extended mediastinal lymphadenectomy' above.)

●Sampling metastatic disease – Sampling site(s) of suspected metastatic disease is necessary for tissue confirmation. Sampling the pleural space (effusions or solid lesions) is essential in all patients with suspected pleural involvement. Thoracentesis is typically performed first for evaluation of pleural effusions. Pleural biopsy (image-guided or thoracoscopic) is often performed when pleural fluid cytology is negative or to obtain tissue from a solid pleural lesion. (See 'Sampling metastatic disease' above and 'Suspected pleural metastases' above.)

●Sputum – Sputum cytology has diagnostic value in a small population of patients with suspected NSCLC. Sensitivity is highest for large, centrally located lesions, and lower for small peripheral lesions. It is not useful for staging NSCLC. (See 'Sputum cytology' above.)