Abstract
Advances in bronchoscopy have contributed valuable tools to the diagnosis and staging of lung cancer. Detection of lesions at the premalignant microscopic stage has become possible with autofluorescence bronchoscopy and narrow band imaging. Bronchoscopy also allows for sampling of visible intra-bronchial lesions and for transbronchial needle aspiration of lesions in pulmonary parenchyma. With endobronchial ultrasound guidance, real-time evaluation and biopsy of mediastinal and pulmonary lesions can be achieved, enabling accurate clinical and pathological T-staging and N-staging without the need for surgery. In combination with advanced imaging techniques, Navigational bronchoscopy allows for the targeting and biopsy of the most peripheral lesions that are located in the smallest airways. For patients in whom tumor genetics are important, bronchoscopic-guided transbronchial biopsy can provide sufficient material for molecular analysis. As minimally invasive technology continues to evolve and improve, bronchoscopic techniques are poised to continue to be essential for the diagnosis and staging of lung cancer.
Introduction
Accurate staging of non-small cell lung carcinoma (NSCLC) is of utmost importance because it dictates management and determines prognosis [Health ACOCP Committee SP, 2003]. The comprehensive staging system proposed by the American Joint Committee on Cancer aims to determine the degree of local invasion (T stage), loco-regional lymph node involvement (N stage), and distant metastatic spread (M stage) [Detterbeck et al. 2009]. Imaging modalities such as computed tomography (CT) of the chest and abdomen, whole body positron emission tomography (PET), and magnetic resonance imaging (MRI) of the brain are usually first performed to rule out distant metastases [Groth et al. 2008]. In the absence of metastatic spread, the next step involves in-depth T staging and N staging to determine the feasibility and utility of surgical resection.
T stage is determined by the size of the lesion, its relationship to major airways, and its degree of invasion of adjacent organs [Detterbeck et al. 2009]. Tumor size and gross local invasion are the only components of the T stage that can be adequately assessed by CT. Other determinants such as precise relationship to the airway, subtle invasion of vascular structures, and the presence of preinvasive disease usually require interventional bronchoscopic techniques [Banerjee, 2009].
N stage is determined by the presence of cancer-positive lymph nodes and their location within the chest. The presence of positive intrapleural lymph nodes (N1 disease) does not usually preclude surgical resection. In contrast, patients with positive extrapleural mediastinal lymph nodes (N2 disease) should be referred to chemoradiation prior to any operative treatment [Detterbeck and Sukumar, 2008]. Clinical N staging is suboptimal with CT (sensitivity 47–54%) and PET (sensitivity 69–79%), even when these two imaging modalities are combined as a PET-CT (sensitivity 83–96% and specificity 78–91%) [Freudenberg et al. 2010]. Surgical mediastinoscopy remains the gold standard with a high sensitivity (80%) and specificity (100%) and a very low false negative rate (7%) [Rami-Porta and Call, 2012]. Although they occur very rarely, complications of mediastinoscopy such as great vessel injury and recurrent laryngeal nerve palsy cannot be overlooked [Shrager, 2010]. In many cases, tissue biopsy under bronchoscopic guidance can obviate the need for mediastinoscopy and avert the risk of surgery and general anesthesia.
Flexible bronchoscopy is the vehicle of choice for delivery of advanced staging technology, such as autofluorescence bronchoscopy (AFB), narrow band imaging (NBI), endobronchial ultrasound (EBUS), and transbronchial needle aspiration (TBNA). This review will focus on the utility of each of these modalities with emphasis on EBUS in the T staging and N staging of NSCLC.
Autofluorescence bronchoscopy
The utility of conventional bronchoscopy (white light bronchoscopy) is restricted to situations of airway invasion by gross tumor. In cases of preinvasive disease (carcinoma
Narrow band imaging
NBI exploits the fact that dysplastic tissue, through angiogenesis and neovascularization, displays a denser network of blood vessels and capillaries than normal surrounding tissue [Shibuya et al. 2003]. The NBI system uses two distinct wavelengths of light to enhance visualization of blood vessels: a blue narrow band (390–445 nm) that is absorbed by superficial mucosal capillaries and a green narrow band (530–550 nm) that is absorbed by the thicker deeper blood vessels [Perin and Zaric, 2012]. When visualized under NBI, dysplastic tissue can be distinguished from normal surrounding tissue due to the characteristic appearance of dense and tortuous blood vessels. Compared with white light bronchoscopy, NBI allows for a 23% increase in sensitivity for detection of premalignant airway lesions [Vincent et al. 2007]. Like autofluorescence, NBI is particularly useful in surveillance of high-risk patients who have no evidence of gross tumor. Compared with AFB, NBI shows a slightly lower sensitivity, but significantly higher specificity in diagnosing high-grade dysplasia [Herth et al. 2009].
Transbronchial needle aspiration
TBNA of mediastinal structures was first reported in 1983, and is finding increased utility in the N staging of lung cancer [Wang et al. 1983; Mehta et al. 1989]. Using guidance from pertinent findings on the patient’s CT scan images, and associating with anatomical landmarks on bronchoscopy, TBNA can be performed on enlarged hilar and mediastinal lymph nodes which are believed to harbor tumor [Wang et al. 1983]. Because of the absence of real-time image guidance at the time of biopsy, the success of TBNA greatly depends on the operator’s skill and their capability of image correlation. For that reason, the diagnostic yield of TBNA for lymph node staging varies greatly in the published literature, with reported sensitivity ranging between 14% and 100% and a false-negative rate ranging between 0 and 66%. The high false-negative rate makes TBNA less useful for staging of the mediastinum in patients without extensive lymphadenopathy on imaging. In contrast, in patients with radiographic evidence of enlarged mediastinal lymph nodes adjacent to the airways TBNA has a reported specificity of 100% and a false-positive rate of 0%. Positive results are reliable but negative TBNA results cannot exclude mediastinal nodal involvement [Wang et al. 1983; Wang and Terry, 1983].
Endobronchial ultrasound
The development of EBUS started in the 1980s and was introduced into clinical practice in 1992 [Hurter and Hanrath, 1992]. The initial EBUS was the radial probe EBUS (RP-EBUS), which is used through the working channel of the flexible bronchoscope. The convex probe EBUS with ability of real-time EBUS-guided TBNA (EBUS-TBNA) was first introduced in 1994 [Yasufuku et al. 2004a] and quickly moved to play a prominent role in lymph node staging of lung cancer [Yasufuku et al. 2004b].
Radial probe endobronchial ultrasound
RP-EBUS is a 20 MHz rotating mechanical probe which produces high-resolution 360° images. It has a penetration depth of 5 cm. RP-EBUS is inserted through the working channel of a flexible bronchoscope. The central-type RP-EBUS is fitted with a 2.6 mm balloon sheath carrying a water-filled balloon at the tip. Using this probe, detailed imaging of the central airway, bronchial wall, and peribronchial structures is possible. The peripheral-type RP-EBUS is used for the assessment of peripheral nodules [Yasufuku, 2010] (Figure 1). Given the penetration depth of RP-EBUS, its use allows assessment of depth of tumor infiltration into the airways. The cartilaginous portion of the bronchial wall has been described as a five- to seven-layer structure. RP-EBUS contributes to the T staging of lung cancer by enabling detailed visualization of the different layers of the bronchial wall, and reliably differentiating between airway invasion and airway compression by the tumor [Herth et al. 2003b; Tanaka et al. 2000].

Radial probe endobronchial ultrasound. Computed tomography scan image of a left upper lobe pulmonary nodule (a) and corresponding radial probe endobronchial ultrasound image showing irregular margin and heterogeneous internal echoes (b).
RP-EBUS-guided TBNA has also been described, whereby suspicious lymph nodes are first identified with RP-EBUS, then the probe is withdrawn allowing for a conventional TBNA at the same endobronchial location. In a report of 242 patients with enlarged mediastinal lymphadenopathy, Herth and colleagues successfully sampled 86% of the cases with conventional TBNA following lymph node localization by RP-EBUS. Diagnosis was obtained in 72% without complications [Herth et al. 2003a]. A randomized study comparing the yield of TBNA under RP-EBUS guidance with conventional TBNA showed that the overall yield of RP-EBUS-TBNA was higher compared with conventional TBNA (85%
Convex probe endobronchial ultrasound
CP-EBUS is a flexible bronchoscope integrated with a convex transducer on the tip, which scans parallel to the insertion direction of the bronchoscope. Visualization of the peribronchial structures can be achieved either via direct EBUS contact with the bronchial wall or via an inflated tip of the probe water-filled balloon. CP-EBUS (BF-UC180F, Olympus) is connected to a dedicated ultrasound scanner which produces excellent quality ultrasound images. Power and color Doppler modes are available for precise characterization of examined structures. Needles, 21- and 22-gauge, with an internal stylet for tip clearance are used for tissue aspiration during real-time EBUS-TBNA (Figure 2). The reach of CP-EBUS within the mediastinum is complimentary to the cervical mediastinoscope [Yasufuku and Keshavjee, 2010]. CP-EBUS can access the upper paratracheal (#2R, #2L), lower paratracheal (#4R, #4L), the subcarinal (#7), the hilar (#10), the interlobar (#11), and the lobar (#12) nodal stations. CP-EBUS cannot access the prevascular and the retrotracheal (#3), the subaortic (#5), the para-aortic (#6), the paraesophageal (#8), and the pulmonary ligament (#9) nodes. Some authors have used the CP-EBUS endoscope through the esophagus to sample the paraesophageal (#8), and the pulmonary ligament (#9) nodes [Herth et al. 2010; Hwangbo et al. 2010]

Convex probe endobronchial ultrasound. Computed tomography scan image of a patient with a right upper lobe primary tumor (a) along with marginal right hilar and mediastinal adenoapthy (b–d) and corresponding convex probe endobronchial ultrasound images of right interlobar lymph node station 11R (e), subcarinal lymph node station 7 (f), right lower paratracheal lymph node station 4R (g), right upper paratracheal lymph node station 2R (h), left lower paratracheal lymph node station 4L (i) and right upper lobe primary tumor (j). The lymph nodes are evaluated in a systematic way. In this case, endobronchial ultrasound guided transbronchial needle aspiration (EBUS-TBNA) was performed from lymph node stations 4L, 2R, 4R, 7 and 11R in this order. EBUS-TBNA was also performed from the primary tumor.
Multiple studies have demonstrated high sensitivity, specificity and diagnostic accuracy of CP-EBUS in diagnosis of mediastinal lymphadenopathy in patients with lung cancer. The first prospective study to evaluate mediastinal lymph node staging in patients with lung cancer demonstrated sensitivity of 94.6%, specificity of 100%, positive predictive value of 100%, negative predictive value of 89.5% and diagnostic accuracy of 96.3%. Not only did EBUS-TBNA provide reliable and pathological staging, but it also was shown to significantly affect treatment decisions in the management of lung cancer. Out of 105 patients, EBUS-TBNA was successfully performed with a diagnostic accuracy rate of 96% and served to avoid 29 mediastinoscopies, eight thoracotomies, four VATS, and nine CT-guided percutaneous biopsies [Yasufuku et al. 2005]. Compared with conventional imaging (CT and PET) for N staging, EBUS-TBNA has been shown to have higher sensitivity (92%
The largest multicenter, prospective study evaluated 502 patients with lung cancer with documented mediastinal and hilar lymphadenopathy. The sensitivity of EBUS-TBNA for malignancy detection was 94% after 572 nodal biopsies and positive results in 535 nodes. However, given the high malignancy prevalence in the studied population (98.2%), the negative predictive value was only 11% [Herth et al. 2006]. Multiple large series and meta-analyses involving hundreds of patients have consistently demonstrated the safety, efficacy, and high diagnostic yield of EBUS-TBNA, especially in patients with radiological evidence of enlarged mediastinal lymph nodes [Yasufuku et al. 2005; Yasufuku, 2006; Hwangbo et al. 2009; Herth et al. 2006, 2008; Gu et al. 2009; Adams et al. 2009]. Based on this evidence, EBUS-TBNA was introduced into the second edition of the American College of Chest Physicians’ evidence-based practice guidelines for the invasive staging of lung cancer [Health ACOCP Committee SP, 2003].
A lingering question that has not been resolved yet is how the diagnostic accuracy of EBUS-TBNA compares with that of cervical mediastinoscopy, long considered the ‘gold standard’ of all mediastinal staging procedures. A recent prospective controlled trial, in which 153 patients each underwent EBUS-TBNA and mediastinoscopy under general anesthesia in the same setting, failed to show any significant differences between the two procedures [Yasufuku et al. 2011]. The sensitivity, negative predictive value, and diagnostic accuracy for mediastinal lymph node staging for lung cancer were 81%, 91%, 93% respectively for EBUS, and 79%, 90%, 93% respectively for mediastinoscopy. No significant differences were found between EBUS-TBNA and mediastinoscopy in determining the true pathologic N stage. Because both EBUS-TBNA and mediastinoscopy are highly operator dependent, it remains to be seen whether the results of this trial can be reproduced. In all likelihood, EBUS-TBNA will serve to reduce the number of mediastinoscopies performed in the future; however, due to its inherent rate of false-negative results, it is not clear that EBUS-TBNA will completely replace mediastinoscopy for definitive mediastinal staging.
Complications related to EBUS-TBNA are similar to those of conventional TBNA, including pneumothorax, pneumomediastinum, hemomediastinum, mediastinitis, bacteremia, and pericarditis. Unlike in mediastinoscopy, no major complications for EBUS-TBNA have been reported in the literature.
Navigational bronchoscopy
Electromagnetic navigational bronchoscopy is a novel technique that was first reported in humans in 2006 [Schwarz, 2006]. The procedure is carried out on a table which emits electromagnetic waves that are captured by a minute receptor probe introduced into the flexible bronchoscope. The system is also linked to a computer where the patient’s preloaded CT scan serves as a virtual plan for the procedure. Major anatomical landmarks (such as carinas) can be identified on bronchoscopy and registered as datasets on the CT scan. The computer software then automatically aligns the patient’s anatomy with the CT scan, creating a virtual roadmap for the bronchoscopist. This effectively enables the bronchoscope to be guided through the smallest airways right into the part of the lung that contains a lesion of interest. The lesion can then be biopsied, imaged, or marked for treatment by surgery or radiation. The majority of studies have looked at the use of navigational bronchoscopy for transbronchial biopsy of peripheral lung nodules [Eberhardt et al. 2007], but it has also been applied for guidance during mediastinal lymph node biopsy [Gildea et al. 2006].
Molecular staging
Customizing treatment based on histology and molecular typing has become the standard of care for management of patients with metastatic NSCLC. Thus, epidermal growth factor receptor (EGFR) genotyping and pathology subtyping should be considered routine in newly diagnosed patients with metastatic NSCLC. Patients with somatic EGFR activating mutations can be treated initially with EGFR tyrosine kinase inhibitor until progression, followed by standard chemotherapy. Other targeted therapies such as crizotinib in tumors harboring the EML4-ALK gene rearrangement have also recently been highlighted [Leighl, 2012]. Most patients with metastatic NSCLC have inoperable disease and therefore surgical samples from the primary tumor or metastatic lymph nodes are not available for molecular analysis. Clinicians must therefore rely on biopsy specimens for initiating molecular typing. This may be challenging from small needle biopsy specimens.
EBUS-TBNA can target metastatic mediastinal and hilar lymph nodes in a minimally invasive way and has been shown to provide an adequate specimen for molecular and genetic analysis. In patients with lung cancer with N2 or N3 disease proven by EBUS-TBNA, DNA extracted from paraffin-embedded samples has been shown to be feasible for the detection of EGFR mutations [Mohamed et al. 2008]. Methylation analysis as well as extraction of RNA from EBUS-TBNA samples has been demonstrated, although not standardized. RNA from metastatic lymph nodes can be used for aberrant fusion gene detection (EML4-ALK fusion gene) [Sakairi et al. 2010]. EGFR mutation analysis using EBUS-TBNA samples was first reported in 2007 [Nakajima et al. 2007] followed by multi-gene mutation analysis of metastatic lymph nodes [Nakajima et al. 2011]. Owing to the spread of the technology, an increasing number of centers are now capable of reproducing the data [Navani et al. 2012]. However, this requires expertise and the way that the samples are handled becomes crucial. It is important to optimize the methodology of specimen handling during the procedure [Nakajima and Yasufuku, 2011]. For cytological evaluation, smears can be processed by an air-dry method as well as ethanol fixation. Histological cores can be used to make formaldehyde fixed paraffin embedded samples for histological diagnosis, which would enable pathologists to perform immunohistochemistry. Flow cytometry and
Summary
Bronchoscopic techniques have evolved to occupy a central role in all aspects of T staging and N staging of NSCLC. Be it for the detection of premalignant lesions through AFB and NBI, or for the accurate delineation of tumor invasion by EBUS, or for pathological and molecular lymph node staging of the mediastinum by EBUS-TBNA, flexible bronchoscopy will continue to find wide-ranging utility in the staging and treatment of patients with lung cancer.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
KY has received educational and research support funding from Olympus Medical Systems. WCH declares that there is no conflict of interest.
