Lung cancer staging by combined endobronchial ultrasound (EBUS) and endoscopic ultrasound (EUS): The gastroenterologist's perspective

1. Introduction 

Lung cancer is the leading cancer-related cause of death in United States in both men and women [1]. Non-small cell lung cancer (NSCLC) usually metastasises first to hilar and mediastinal lymph nodes and subsequently, haematogenous metastasis to other distant areas may occur. Patient survival depends on the stage of the cancer and is inversely correlated to stage. It is critical to stage the disease accurately as choice of therapy is dictated by stage of the disease. Lung cancer is staged by TNM classification where ‘T’ stands for local tumour extension, ‘N’ for lymph node metastasis and ‘M’ for distant metastasis similar to other cancers. The distribution and mapping of lymph nodes in the mediastinum (Fig. 1) are used to describe ‘N’ factor in TNM staging [2].

View Large Image Download to PowerPoint

Fig. 1. Mediastinal and hilar lymph nodes stations. Lymph nodes accessible to endoscopic ultrasound scope include level 2 right and left of trachea, 4 R/L, 5, 7, 8 and 9. Lymph nodes anterior to the trachea (levels 2 and 4) are accessible only to endobronchial ultrasound scope.

In the absence of distant metastasis or mediastinal involvement (N2–3), patients are considered surgical candidates. Also in some patient with limited ipsilateral lymph node involvement (N2) surgery may cure the disease. In those patients who have extensive (bulky) N2 or N3 (contralateral side relative to tumour) or metastatic disease, surgery is not an option and chemotherapy with or without radiation is considered the standard therapy. Nevertheless, there is some evidence suggesting survival of patients with more advanced disease after resection [3].

This article outlines the basic principles of lung cancer (NSCLC) staging by these minimally invasive procedures and assists the gastroenterologist performing endoscopic ultrasound (EUS).

Non-invasive methods for lung cancer staging mainly include computed tomography (CT) and positron emission tomography (PET) scanning. These modalities, though safe, have limited sensitivity and specificity, with positive predictive values (PPVs) of only 56–79% and negative predictive values (NPVs) of 83–93% for detection of mediastinal lymph node metastases [2]. One study reported 23% of patients having unresectable disease at surgical staging, a third of these because of malignant lymph nodes being missed by preoperative PET staging [4]. Similarly, patients with negative CT for mediastinal adenopathy can have up-to 35% prevalence of malignant mediastinal lymph nodes [5]. In order to limit false positive (FP) and false negative (FN) results of these imaging modalities, tissue sampling of mediastinal lymph nodes is considered essential for accurate staging.

The American College of Chest Physicians (ACCP) guidelines for lung cancer staging, recommend patients with abnormal lymph nodes on CT or PET, or centrally located tumours without mediastinal adenopathy, should undergo invasive staging [6]. The most common invasive staging includes mediastinoscopy/thoracoscopy, transbronchial needle aspiration (TBNA), endoscopic ultrasound-fine needle aspiration (EUS-FNA), and endobronchial ultrasound-fine needle aspiration (EBUS-FNA).

Mediastinoscopy, a modality considered to be the diagnostic standard (NPV, 89%; PPV, 100%), is an invasive procedure with significant limitations [6]. It is a surgical procedure performed under general anaesthesia and is best suited to sample pretracheal and paratracheal lymph nodes. It has limited access to the inferior and posterior mediastinum as well as aortopulmonary window. Though generally safe, mediastinoscopy has a 2% risk of major morbidity and a 0.08% risk of mortality and it is substantially more costly than EUS-FNA [6], [7], [8]. Proponents of mediastinoscopy cite its high NPV, particularly for patients without enlarged mediastinal lymph nodes [6]. However, a recently published study by Wallace et al. showed combined NPV of EUS and EBUS to be 97% in patients with non-small cell lung cancer with or without enlarged mediastinal lymph nodes, approaching that of thoracotomy with mediastinal lymph node dissection [9]. In another study, Larsen et al. compared EUS-FNA with mediastinoscopy in patients with subcarinal or paratracheal lymphadenopathy and showed EUS-FNA to be significantly more accurate than mediastinoscopy, specifically in the subcarinal area. However, this study did not evaluate patients with enlarged pretracheal lymph nodes, a region in which mediastinoscopy is expected to do better [10].

The mediastinal lymph nodes are surrounded by great vessels, heart, lungs, bones and this makes tissue sampling challenging. All of these lymph nodes are not accessible by either EUS or EBUS alone and that may be the reason for FN samples with one modality alone. For lung cancer staging, EUS and EBUS can access the majority of these lymph nodes complementing each other. These minimally invasive methods can be used for tissue confirmation of suspected metastatic mediastinal lymph nodes.

2. Equipment 

EUS equipment consists of two types of echoendoscope transducers for imaging; a radial and curvilinear array transducer. With radial echoendoscope the image is oriented perpendicularly in a radial format relative to the echoendoscope; the linear echoendoscope has the image oriented parallel with the scope. Radial echoendoscope is well suited for lung cancer staging; however, the linear scope is required to perform FNA as it allows visualisation of the needle entering into the target lesion. Linear echoendoscopes (Fig. 2) are produced by Olympus (Tokyo, Japan), Pentax (Golden, CO) and Toshiba (New York, NY). These scopes have standard accessory channels (2.0–2.8mm) and larger accessory channels (3.7mm) capable of delivering needles and other therapeutic devices. The echoendoscopes require attachment to a specific ultrasound processor which cannot be used with other makers of echoendoscopes. A recently developed linear echoendoscope with forward viewing ultrasound and optics is in the process of evaluation but not widely available for clinical use. Initially radial scanning catheter probes placed through the bronchoscope were used for imaging of lung mass, bronchial walls and lymph nodes. The flexible ultrasonic bronchoscope (Olympus BF-UC160F-OL8, Tokyo, Japan) developed in 2004, is a linear EBUS scope which gives image parallel to the scope (Fig. 3). It has an outer diameter of insertion tube of 6.7mm with biopsy channel of 2mm. The ultrasonic frequency is 7.5MHz with penetration depth of 4–5cm and is well suited for FNA of lymph nodes and lung masses through the trachea and bronchi.

View Large Image Download to PowerPoint

Fig. 2. Curvilinear echoendoscope with 22 gauge needle for the performance of fine needle aspiration (Olympus GF-UC140P-AL5, Olympus, Tokyo, Japan).

View Large Image Download to PowerPoint

Fig. 3. Curvilinear (Olympus GF-UC140P-AL5) (A), endobronchial (Olympus BF-UC160F-OL8) (B), and radial array (Olympus GF-UE160_AL5) (C), echoendoscopes in comparison.

For EUS-FNA, different sized needle are available (19–25g). Usually, 22g needle is used for biopsy of mediastinal lymph nodes. There is no proof that 22g needle is the best needle for sampling the mediastinal lymph nodes. Indeed, there is now a tendency to move from 22g to 25g needle because of a better accuracy for most FNA indications. For EBUS a 22g needle is used to perform FNA. All needle systems are luer-locked to the accessory channel of the echoendoscope and are advanced into the tissue with a piston like handle. The needle is occluded with stylet during passage through the GI tract wall and bronchial wall to minimise the contamination from passage through the structures.

3. Description of the procedure 

EUS-FNA is performed through the oesophagus and ultrasonic rays do not penetrate air-filled structures, thus regions immediately anterior to the trachea (areas 2 and 4) are a “blind spot” for EUS-FNA and FN sample appears to be lymph node metastases located in sites inaccessible to EUS-FNA [11]. For this reason EUS is best suited for sampling of the lymph nodes in the posterior mediastinum. As compared to EUS, EBUS on the other hand is best suited for sampling of the lymph nodes in areas anterior and parallel to the trachea, as it is performed via trachea and does not share the same ‘blind spot’.

The examination for lung cancer staging can be performed by either radial or linear scope. The radial scope can be used for staging purpose and then a linear echoendoscope is used for targeted EUS-FNA. Alternatively, the entire examination can be performed with a linear echoendoscope and indeed this may be preferable, since echo-characteristics alone have limited accuracy and EUS staging of lung cancer almost always requires FNA of lymph nodes. Informed consent is obtained before the procedure. EUS and EBUS can be performed under conscious sedation or general anaesthesia. After sedation, the patient is intubated with the echoendoscope. The most distant sites of potential metastases are surveyed first. Initially the liver and if possible right adrenal gland can be inspected through the duodenal bulb and lesser curve of the stomach. The left adrenal gland can be identified 2–4cm left of the aorta at the level of the celiac artery origin. With the linear echoendoscope, it can be identified by rotating echoendoscope 30° clockwise at the origin of celiac artery. Examination of the mediastinum with the radial echoendoscope is a simple pullback procedure. The subcarina is located immediately superior to the left atrium and the aortopulmonary window is located immediately inferior to the aortic arch. With the linear echoendoscope, the mediastinum is surveyed by first finding the descending aorta at the gastro-oesophageal junction, and then rotating 360° through the mediastinum until the aorta again comes into view. The instrument is withdrawn 2–3cm and the manoeuvre repeated until the entire mediastinum has been inspected. Subcarinal lymph nodes are seen immediately under the echoendoscope, between the left atrium and pulmonary artery (Fig. 4). Aortopulmonary window lymph nodes are seen by following the descending aorta cephalad to just below the aortic arch, then rotating clockwise 90° and tipping the up–down dial in the upward direction (Fig. 5).

View Large Image Download to PowerPoint

Fig. 4. Subcarinal lymph node is seen immediately under the echoendoscope (curvilinear), between the left atrium and pulmonary artery. A needle is seen passing into the lymph node (arrow).

View Large Image Download to PowerPoint

Fig. 5. Aortopulmonary window lymph node seen with curvilinear echoendoscope. Note the landmarks of aorta and pulmonary artery.

EUS-FNA is usually performed using a 22g or 25g needle with pre-loaded stylet. The node is punctured under real time ultrasound guidance and then stylet is withdrawn. The needle is moved back and forth through the lymph node to obtain lymph node tissue. Suction is not necessary to obtain tissue specimen, but can be used if enough specimen is not obtained with initial puncture. The specimen is placed onto a slide for preparation by the endoscopist or cytotechnician. In many institutions cytopathologists prepare the slides themselves at the time of tissue acquisition. Usually 3–4 needle passes are sufficient to obtain adequate tissue, but more passes can be made if cytopathologist requires more tissue for cytological evaluation.

EBUS is performed by introducing the EBUS scope via vocal cords in a patient under conscious sedation and via endotracheal tube if procedure is performed under general anaesthesia. The tip of the scope is placed under direct visual guidance in different positions against the bronchial wall. The aim is to outline 360° along the main bronchi and the trachea, starting from the segmental bronchi to the distal end of the endotracheal tube. With EBUS-TBNA the paratracheal (stations 2 and 4) (Fig. 6), subcarinal (station 7) hilar and intrapulmonary nodes (stations 10 and 11) can be reached. EBUS-TBNA can be performed with the transducer in direct contact with the wall of the trachea or bronchus. After the lesion is identified, needle is introduced through the biopsy channel of the echoendoscope. Doppler immediately before the puncture can be used to prevent unintended puncture of blood vessels. After the puncture, suctioning can be applied to the syringe and needle is moved back and forth through the lesion to obtain tissue specimen. Before retracting the needle from the lesion suction is turned off and needle pulled out. The specimen is placed on the glass slides, air dried and stained for cytological examination. If apparent tissue specks are seen, these can be placed in formalin for histopathological evaluation. Just like EUS-FNA, in many institutions cytopathologists prepare the slides themselves at the time of tissue acquisition. Enough passes should be made as per the cytopathologist's request to obtain adequate tissue.

View Large Image Download to PowerPoint

Fig. 6. View of pretracheal lymph node with convex array endobronchial echoendoscope. A needle is seen passing into the lymph node (arrow).

Multiple lymph nodes are typically seen in the mediastinum of NSCLC patients. Determining what site to sample with EUS-FNA is based on clinical picture and the echo-characteristics of the lymph nodes. However, lymph nodes echo texture could be variable. Although the classic four echo features (hypoechoic, round, size >1cm, sharply demarcated) are useful, some malignant lymph nodes have none of these features, and some benign lymph nodes may have many of these features. Liver metastases are most often hypoechoic, but can be hyperechoic or mixed echogenicity. The adrenal gland may be involved by cancer appearing enlarged and round. The site of the tumour is also predictive of the site of lymph node metastases. Left sided tumours, especially the left upper lobe, typically drain into the aortopulmonary window (level 5), whereas either side can drain into the subcarina (level 7), or inferior mediastinum (levels 8 and 9). It is our practice to sample at least one lymph node (typically the largest lymph node) from the subcarina, and aortopulmonary window, and any other accessible sites. The procedure is stopped if any lymph node is found to be malignant, or if all sites are benign after sampling each site with 3–4 FNA passes. In our experience by performing more than 3–4 passes does not increase the diagnostic yield of lymph nodes FNA.

For safe, effective and easier EUS-FNA performance positioning of the echoendoscope with respect to the lesion is very important. Once a lesion is identified, it should be placed in the natural path of the needle. Doppler can be used before the biopsy in order to avoid unintended puncture of vessels between the wall of the oesophagus or bronchi and the lesion.

EUS-FNA not only requires a linear array echoendoscope, a compatible needle system and cytopathology support, but also a trained endosonographer. Training in EUS-FNA technique can be obtained through advanced endoscopy training fellowship, short courses and tutorials, which can be supplemented with published literature. The American Society of Gastrointestinal Endoscopy (ASGE) has published minimal guidelines for competency in EUS, which is 150 cases of supervised EUS cases, of which 75 should include FNA. For EBUS, ACCP has published that trainees should perform at least 50 procedures in a supervised setting to establish basic competency in analyzing anatomic structures and handling the instrument. To maintain competency, dedicated operators should perform at least 20 examinations per year [12]. These guidelines were published in 2003 and are for radial probes through the bronchoscope. Linear EBUS scope was introduced in 2004 and no guidelines have been made available by ACCP with regard to number of FNA with linear scope.

4. EUS and EBUS for lung cancer staging 

Major indication for EBUS and EUS combined approach is the mediastinal lymph node staging for lung cancer. EUS-FNA which has been developed and widely used for staging and tissue sampling of gastrointestinal tumours now has emerged as great tool for mediastinal lymph node staging in lung cancer and has been reported to have PPV greater than 99% and an NPV of 81% [6].

EUS-FNA is well suited to sample lymph nodes in stations which are located immediately adjacent to the oesophagus and in the posterior mediastinum. EUS-FNA can readily access lymph nodes in aortopulmonary window (level 5), subcarina (level 7), as well as inferior mediastinum along the oesophagus (levels 8 and 9). EUS-FNA access to lymph nodes lateral or anterior to the trachea (levels 2R/L and 4R/L) can be very difficult if not impossible; however these stations can be readily accessed by EBUS-FNA. The EUS-FNA technique has the ability to sample nodes as small as 3mm and to identify metastases found in normal-sized lymph nodes. EUS is capable of evaluating the presence of direct tumour invasion into the mediastinum and can also detect metastatic disease to subdiaphragmatic sites like, left adrenal, celiac lymph nodes and liver.

TBNA is performed by “blind” needle puncture. The bronchoscopist by using information derived from CT on the location of lymph nodes can direct the needle for puncture to lymph nodes in pre-, paratracheal, subcarinal areas and to some parts of the aortopulmonary window area. However, limitations of CT especially its inability to identify smaller nodes and also blind nature of TBNA make its accuracy limited [13]. The accuracy of TBNA can be improved by addition of EBUS. In a prospective, controlled trial by Wallace et al. EBUS-FNA was significantly more sensitive than TBNA for detection of mediastinal lymph node metastases [9]. Other studies have confirmed that use of EBUS-guided TBNA results in a high success rate for accessing lymph nodes (86%) and also a higher diagnostic yield (71%) [14]. Vilmann et al. reported their experience in combined use of EUS and EBUS-FNA in the evaluation of mediastinal lesions [15]. Multiple other studies reported EBUS-FNA to be safe with sensitivity of 85–100%, specificity of 100% and accuracy of more than 96% in distinguishing benign from malignant mediastinal lymph nodes in patient with lung cancer [16], [17], [18], [19].

5. EUS for lung cancer staging 

Various studies have been published in the literature regarding the diagnostic utility of EUS-FNA in NSCLC staging [10], [11], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]. In these studies with regard to the detection of malignant mediastinal (N2 or N3) lymph nodes, the overall sensitivity was 84%, and the overall FN rate was 19% (range, 0–61%). The overall specificity was 99.5% and the overall FP rate was 0.4%. Among patients with enlarged lymph nodes on CT, the sensitivity was 87%, FN rate of 22%, specificity of 98% and a FP rate of 2%. In these patients the prevalence of N2 and N3 involvement was 68%. Among patients with normal-sized lymph nodes on CT, the sensitivity was 66% with FN rate of 14%, and the specificity was 100% with a FP rate of 0%. In these patients with normal-sized lymph nodes on CT, the prevalence of N2 or N3 disease was 35%, which is higher than the expected rate (20–25%) based on the CT data for normal-sized mediastinal nodes.

A recent study comparing the utility of preoperative EUS in addition to standard staging (control group) for detection of malignant lymph nodes, found that only 6.8% patients were found to have malignant lymph nodes at surgery in EUS group as compared to 31% in control group thus concluding that EUS can reduce the need of unnecessary surgery in lung cancer patients with advanced disease [32].

6. Combined EUS and EBUS for lung cancer staging 

The major limitation of EUS-FNA in lung cancer staging is its inability to assess lymph nodes in the anterior mediastinum; this challenge can be overcome by transbronchial ultrasound examination. There is emerging data on combination of EUS-FNA and EBUS-FNA for lung cancer staging. These studies show the combination of these two modalities allow near-complete access to all mediastinal lymph node stations. In their study of combined approach with EUS-FNA and EBUS-FNA, Vilmann et al. showed these two modalities to be complementary with a combined accuracy of 100% in 31 patients. Recently, Wallace et al. have shown that combined NPV of EUS and EBUS to be 97%, approaching that of thoracotomy with mediastinal lymph node dissection. They also suggested that if mediastinoscopy had been performed only when results from EUS plus EBUS were negative, this surgical procedure would have been avoided in 28% (39/138) of patients. If EUS plus EBUS had been used to completely replace mediastinoscopy (100% of patients), 97% would have been correctly labelled as negative. They showed that EUS+EBUS also has high sensitivity and NPV in a patient population with suspicious lung tumours in the absence of mediastinal lymphadenopathy on CT and PET, particularly those with centrally located tumours or intraparenchymal lymphadenopathy thus possibly obviating the need of using mediastinoscopy or surgical exploration for mediastinal staging [9].

A key pragmatic question, particularly since EUS is traditionally performed by gastroenterologists, and EBUS and TBNA by thoracic specialists, is whether both tests are needed in every case. In the study by Wallace et al. subgroup analysis looked at the yield of either test alone vs. combined EUS+EBUS in patient with anterior (EBUS suited) or posterior (EUS suited) or neither. The study found the combination of EUS+EBUS outperformed either test alone in all subgroups and thus should be recommended, even if the lymph nodes appear easily accessible to one procedure or the other.

EUS is not only cost effective but also very accurate modality for lung cancer staging. Studies comparing EUS directly to other modalities suggest that EUS is more accurate than CT [23], PET scan [26], [33], and TBNA [34]. Also cost effective studies suggest EUS to be more cost effective than mediastinoscopy [5], [6], [35].

7. Complications and safety 

In the hands of an experienced endosonographer, EUS-FNA is a highly safe procedure. A recent meta-analysis of 18 studies with a total of 1201 patients reported only 10 (0.8%) cases with minor complications [5]. Major safety precaution with regard to FNA are careful visualisation of the entire length of the needle as it passes into the target lesion and use of colour flow Doppler prior to needle puncture to avoid any blood vessels in the needle path.

8. Limitations and future challenges 

The major limitation of EUS-FNA in lung cancer staging is its inability to assess lymph nodes in the anterior mediastinum (areas 2 and 4) as ultrasound cannot penetrate through the air-filled structures (trachea and bronchi). This limitation can be overcome by transbronchial ultrasound examination. Other limitation could be inability to sample some lymph nodes due to intervening large vessels (level 6). Also, like all sampling techniques, EUS-FNA and EBUS-FNA are subject to sampling error.

Lastly, provision of high-quality EUS-FNA and EBUS-FNA testing requires need of training, specialised equipment and combined endoscopic and bronchoscopic approach. There is emerging evidence that EUS and EBUS in combination is better modality than either one alone. This makes implementation of this modality somewhat more challenging since neither gastroenterologist nor pulmonologist routinely perform both procedures. It is not clear at this time if EUS and EBUS for lung cancer should be performed by one personnel trained in both techniques or it should be a multispecialty approach as done at our institution. At our institution, EUS-FNA is performed by the gastroenterologist and EBUS-FNA is performed by pulmonologist, with both procedure performed on the same day followed by each other. In our view, one personnel (either gastroenterologist, pulmonologist or a surgeon) can be trained to perform both procedures which might make implementation of this emerging modality more feasible.

9. EUS-FNA and biomarkers 

The standard histological or cytological evaluation possibly fails to detect early “micrometastases” in mediastinal lymph nodes, as even after presumed curative surgical resection there is only 62% 5-year survival rate for stage I disease (no histological evidence of lymph node metastases) and 42% for stage II disease (histological metastasis of hilar lymph nodes but not mediastinal lymph nodes) [36]. Many of these tumour cells can be detected using molecular analysis (PCR) for lung epithelial cell specific proteins or mRNA in lymph node tissue samples obtained by EUS-FNA and EBUS-FNA. Promising markers for detection of micrometastases to lymph nodes include KS and telomerase [37], [38]. Combining fine needle aspiration and modern genomic methods offer a rich area of further research and clinical applications. Currently, the effect of epidermal growth factor receptor family of genes (EGFR) mutations on response to treatment and survival is under evaluation.

10. Conclusions 

There are multiple modalities for staging of lung cancer. Treatment of lung cancer depends on its stage, so it is crucial to stage it accurately. Different staging modalities include EUS-FNA, EBUS-FNA, combined EUS-FNA and EBUS-FNA, mediastinoscopy/thoracoscopy, CT and PET. In recent years, there has been a greater role of minimally invasive techniques such as EUS-FNA and EBUS-FNA. Existing evidence suggests that EUS-FNA and EBUS-FNA combined approach may be the preferred modality and will likely have great impact on lung cancer staging. To date, studies have shown these procedures to be safe, complementary and potentially able to minimise the need of staging by thoracoscopy and mediastinoscopy.

Further studies are needed to optimise the staging algorithm of lung cancer. Provision of high-quality EUS-FNA and EBUS-FNA testing requires need of training in both procedures.

Conflict of interest statement 

Dr. Wallace has received educational grants from the Olympus and Wilson Cook companies.