Review of the role of EBUS-TBNA for the pulmonologist, including lung cancer staging

Abstract

This review focuses on the role of endobronchial ultrasound-guided transbronchial needle aspiration in day-to-day pulmonology practice. Case examples are given of the common indications for endobronchial ultrasound-guided transbronchial needle aspiration which are: (i) lung cancer staging; (ii) confirming a diagnosis of malignancy in thoracic lymph nodes; (iii) diagnosing central pulmonary masses; (iv) sarcoidosis; and (v) inflammatory/benign thoracic lymph nodes. The technique is widely used, and after appropriate training by experienced bronchoscopists can be easily integrated into a bronchoscopy service.

Introduction

Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is now an established procedure for pulmonologists.^1,2 Initially bronchoscopy was most often used to diagnose endobronchial pathology. Now we have tools which expand our reach beyond the bronchus. Blind transbronchial needle aspiration (TBNA) using flexible needles via a fiber optic bronchoscope has been available since the early 1980s.^3,4 Despite proving to be safe and effective, there has only been partial acceptance by pulmonologists, with most surveys showing 10–30% uptake of the technique.⁵ Usually lack of success and fear of vascular penetration are the cited reasons, even though serious bleeding has never been reported with this technique in over 30 years. EBUS-TBNA directly addresses these difficulties. In most hands, compared to standard TBNA, the endobronchial ultrasound (EBUS) method allows better access through the bronchial wall (due to the forward angulation of the TBNA needle as it exits the biopsy channel), imaging of confirmed location of the needle within the node, and visualization of adjacent vessels, particularly in vascular regions such as the hilum and lower paratracheal regions. The learning curve is relatively short and persistently good cytological yields can be obtained after 20–30 cases or less.⁶

There are a number of excellent published reviews on EBUS-TBNA covering methods, results and indications.^7,8 This review will present a series of five cases which represent the most common indications for EBUS-TBNA and discuss the relevant literature. Four of the cases were seen within one month. In our unit we perform between 20–25 EBUS-TBNA per month, and have experience of approximately 600 cases.

EBUS-TBNA anatomy and method

The new International Association for the Study of Lung Cancer lung cancer staging guidelines give excellent clarification of nodal positions and this is an important starting point.⁹ It is useful to become familiar with the positions of a few key structures as they relate to the bronchus. On the right they are the superior vena cava, the azygos vein, the right pulmonary artery and the right lower lobe pulmonary artery. On the left they are the descending aorta, the left main pulmonary artery, the left lower lobe pulmonary artery and the esophagus. A useful way to learn this is to scroll up and down on a digital computerized tomography (CT), seeing where these structures contact the bronchus (Fig 1). It can further help to flip the CT images horizontally to give a view as if standing behind the patient as with a bronchoscope. It is also very helpful to read Wang's papers on the bronchoscopic surface landmarks of node positions^3,4 and Ko's paper on CT locations of nodes.¹⁰ Above all a pattern recognition approach for each station is best.⁷ We suggest starting with the subcarinal position then moving to other positions as more experience is gained.

Figure 1. Main relevant anatomical structures for transbronchial needle aspiration on computerized tomography at the (a) carina and (b) lower hilar regions. SVC, Superior vena cava.

A number of excellent publications outline the EBUS-TBNA method.^11–14 The procedure can be successfully done either under local anesthetic with conscious sedation or with general anesthetic and laryngeal mask airway. The latter is preferred by those new to the technique. Briefly, a dedicated EBUS-TBNA bronchoscope is used along with dedicated EBUS-TBNA needles. After preparing the needle and applying the small balloon to the ultrasound tip, the bronchoscope is passed in the usual way through the bronchial tree to a point adjacent to the node of interest. A small amount of saline is used to inflate the balloon to improve ultrasound image quality. Once the node is observed and seen to be separate from adjacent vessels the assistant holds the bronchoscope in place while the TBNA needle is passed down the biopsy channel. The needle is passed gently under direct ultrasound control into the node. A stylet is used to push any unwanted material out of the needle tip and suction is applied; the sample is obtained by passing the needle in and out through various parts of the node. After removing the needle the sample is extruded onto slides or into saline for processing. Usually we have an onsite cytology technician to comment on specimen quality and whether there is adequate diagnostic tissue.

Case presentations and literature review

Case 1: lung cancer staging

A 78-year-old man was referred for staging of a known 4 cm left lower lobe adenocarcinoma, diagnosed by cytology of bronchial washings. Chest CT showed enlarged lymph nodes (10–13 mm diameter) at the subcarinal (7L), left paratracheal (4L) and right paratracheal (4R) regions. Positron emission tomography (PET) scan showed each of these nodal stations was positive (Fig 2a). At EBUS-TBNA the 4R node was sampled and onsite cytology confirmed diagnostic material from the first pass (Fig 2b,c), therefore N3 disease was confirmed. Total procedure time was 20 min.

Figure 2. Case 1. (a) Positron emission tomography scan. (b) Endobronchial ultrasound appearance of 4R lymph node. (c) Cytology smear of transbronchial needle aspiration material showing a clump of malignant cells among erythrocytes (Pap), 10x. A.

Staging of lung cancer is one of the most well known indications for EBUS-TBNA..¹⁵ EBUS-TBNA can reach to the hilar and interlobar regions, whereas mediastinoscopy can only reach paratracheal regions and occasionally subcarinal regions.⁷ Both PET scans and CT scans have false positive and false negative results with respect to lymph nodes and tissue is needed if a true diagnosis of the node will change management, particularly if surgery is to be denied on the basis of a single enlarged lymph node.

Yasufuku's paper¹⁶ was a milestone in the development of EBUS-TBNA for lung cancer staging. One hundred and two resectable lung cancer patients all had CT, PET and EBUS-TBNA before formal surgical staging at the time of thoracotomy. The relative sensitivity and specificity of the three methods were 77%, 80%, and 92% respectively. The diagnostic accuracies were respectively 61%, 73% and 98%. As well as its broad access to nodes, EBUS-TBNA allowed sampling of nodes as small as 5 mm. A meta-analysis of EBUS-TBNA staging of lung cancer showed that in 1299 patients EBUS-TBNA had a pooled sensitivity of 93% and specificity of 100%.¹⁷ Rintoul et al. published a paper on EBUS-TBNA for the clarification of PET positive lymph nodes.¹⁸ Of 109 patients who had EBUS-TBNA and surgical staging a final diagnosis of malignant nodes was made in 77 cases. There were seven cases where the TBNA was negative but the surgical biopsy showed malignancy (false negative TBNA), this was due to sampling error in four cases and detection error in three cases. They therefore recommended negative EBUS-TBNA from PET positive nodes should be followed by surgical biopsy, usually at the time of resection.

Herth et al. presented data on staging of radiologically and PET normal mediastinum with EBUS-TBNA.¹⁹ In 97 resectable patients 156 lymph nodes (5–10 mm) were sampled. Malignancy was detected in nine patients but missed in one patient (as proven by histology of resected lymph nodes). Despite these findings, at the present time most centers would generally accept that if a CT of a resectable lung cancer patient showed no mediastinal node enlargement and there was a negative PET scan then EBUS-TBNA staging was not required prior to surgical resection.

In 2005 Rintoul et al. and Vilmann et al. reported on combined EBUS and endoscopic ultrasound (EUS) staging of the mediastinum.^20,21 In the Vilmann et al. study 33 patients had staging with both EUS and EBUS. EUS sampled 59 nodes and EBUS sampled 60 nodes. Malignancy was found in 26 cases by EUS and 28 cases by EBUS. Eleven additional malignant nodes were diagnosed by EBUS which were not obtained by EUS fine needle aspiration (FNA) (stations 1, anterior trachea, 2R, 4L, 4R, 7 and 10L). Conversely EUS found 12 malignant nodes which EBUS did not (2R, 4R, 4L, 7 and L adrenal). Note that in terms of access EBUS exclusively reached anterior trachea, and stations 10 and 11. EUS exclusively reached stations 8 (2 cases), 9 (1 case) and adrenal. They concluded that the two methods have a combined role. Exactly how this is translated into clinical practice is not clear, however, a number of centers worldwide combine the two methods. The key question is despite the increased number of individual nodes detected, is there actually a change of stage in an individual patient? Larger patient numbers will be required to answer this question more definitively.

In a staging procedure lymph nodes are sampled starting at N3, then distal N2, then proximal N2, then N1. It is best to do this with onsite cytology so that the procedure can be stopped once positive cytology is confirmed. If this is not available in clinical studies of staging a different needle is used at each site because washing the needle with saline is not sufficient to rule out contamination from the previous site. It is acknowledged that in day-to-day practice this would be expensive and as long as there is strict order of sampling, a patient would not be up-staged.

Case 2: Lymph node biopsy to diagnose lung cancer

A 76-year-old man presented with right sided chest pain. He had previous occupational exposure to asbestos and a 35 pack year history of smoking. Five years previously he had a nephrectomy for renal cell carcinoma. CT scan showed a poorly defined pleural based mass in the right lower lobe, associated with pleural effusion (Fig 2a). In addition there were enlarged lymph nodes at the lower R hilar position (R11i) and subcarinal position (7R). Previous bronchoscopy by the referring doctor was negative on bronchial washings and there was no endobronchial disease. PET scan showed surprisingly low uptake in the mass but strong uptake in the R hilar and subcarinal nodes (Fig 3). We performed EBUS-TBNA and sampled the right 11i node. Onsite cytology showed cells consistent with non-small cell carcinoma on the first pass. Subsequent immunohistochemistry was positive for thyroid transcription factor 1 (TTF1). We did not sample the subcarinal node as the primary tumor was inoperable. The total procedure time was 15 min.

Figure 3. Case 2 positron emission tomography scan.

In the diagnosis of enlarged mediastinal lymph nodes in general, a number of prospective studies of EBUS-TBNA have shown high sensitivity (85 to 96%), specificity (100%), and diagnostic accuracy (89 to 97%).¹ In this case there was strong suspicion of an N1 and N2 node both in easily accessible positions, so this was sampled rather than the more difficult to access peripheral mass. Had the mass been metastatic renal cell carcinoma, bleeding from a transbronchial lung biopsy (TBLB) would be likely. Also, if it was mesothelioma it would not have been accessible by TBLB. We published a small series which considers EBUS-TBNA as the first test in patients with peripheral mass that is very likely to be lung cancer.²² It could be argued that because the peripheral mass is not sampled it may be possible the node cells could have a different origin, however, we believe this to be unlikely where the two shadows (the mass and the node) are the only radiological abnormalities. This is particularly true if a prebronchoscopy PET scan shows no obvious uptake in other parts of the body. We believe this approach could obviate other procedures such as CT FNA.

Case 3: Biopsy of the primary lung cancer mass

A 40-year-old woman presented with upper central thoracic back pain. She had a 40 pack year history of smoking. CT scan showed a mass arising high in the left upper lobe immediately adjacent to the posterolateral left trachea (Fig 4a). There was no regional lymphadenopathy. At bronchoscopy there was mild extrinsic compression of the trachea 3 cm below the vocal cords, but no endobronchial disease. EBUS-TBNA was performed with the needle entering the mass at the 7 o'clock position 3 cm below the vocal cords (Fig 4b). Onsite cytology showed diagnostic material from the first pass which was later confirmed as non-small cell carcinoma. Total procedure time was 10 min.

Figure 4. Case 3. (a) Chest computerized tomography. (b) Endobronchial ultrasound appearance showing the large mass occupying most of the ultrasound view with endobronchial ultrasound-guided transbronchial needle aspiration needle inserted for sampling.

Authors have pointed to the advantage of using EBUS-TBNA to diagnose a primary pulmonary mass where there is no endobronchial disease.^23,24 Where the mass is located against a large airway it obviates the need for CT FNA or more difficult procedures, such as TBLB or surgical anterior mediastinotomy. Where a mass is adjacent to segmental bronchi, if the EBUS-TBNA scope can reach the mass, it is much easier than a TBLB. The limiting factor for this method is of course the size of the candidate bronchus. The EBUS-TBNA scope has a maximum outer diameter of 6.9 mm. It is generally accepted that if CT scan shows the bronchus leading to the mass has a caliber of approximately 4 mm then this will allow the scope to be passed close enough to allow EBUS-TBNA. It may be difficult to do this type of biopsy in the upper lobes because of the more acute angles required for access to these bronchi with this large scope.

Case 4: Sarcoidosis

A 48-year-old man presented for investigation of persistent lymphadenopathy and mild persistent chest wall pain. He had commenced prednisone 3 months previously for presumed sarcoidosis without a tissue diagnosis. Chest CT had only shown bilateral hilar and mediastinal lymphadenopathy without any evidence of lung parenchymal disease. The CT changes persisted despite prednisone. EBUS-TBNA was performed with subcarinal and low R hilar (11i) nodes sampled (Fig 5a). The EBUS appearance showed numerous straight septations through the node which contained small vessels on Doppler imaging. Onsite cytology and histology on aspirates showed non-caseating granulomas with negative stains for infecting organisms. An example of pathology from another sarcoidosis case is shown (Fig 5b,c). Endobronchial biopsies were negative. The interpretation was that he in fact had sarcoidosis and treatment has continued with good symptomatic response.

Figure 5. Case 4. (a) Endobronchial ultrasound appearance of sarcoid lymph node showing typical septations. (b) Cytology smear of transbronchial needle aspiration material showing a granuloma (Pap), 10x. (c) Histology of transbronchial needle aspiration material detail of granuloma showing histiocytes (hematoxylin-eosin 40x).

Subsequent to this case we have routinely requested onsite pathology in suspected sarcoidosis cases, as smears can show granulomas and allow the procedure to be stopped. In addition we usually send aspirate material on small squares of filter paper placed in formalin which can be sectioned for histology. Positive results can be obtained with 22-gauge TBNA needles but now, with a larger needle becoming available (21-gauge), yield should improve.

Granulomas are nodular collections of epithelioid histiocytes.²⁵ The cells have round to oval nuclei with some elongated forms with features resembling foot prints or the shape of the sole of a shoe. The nuclei, although pleomorphic, have smooth nuclear membranes and contain fine granular chromatin. The cytoplasm is pale, abundant and has indistinct cell borders. Unlike the granulomas associated with tuberculosis the background in these slides are clean, with blood contamination but no necrotic debris present.²⁶

The EBUS appearance of sarcoidosis lymph nodes has been commented on by a number of practitioners as having strong characteristics, node septae and vessels are often straight and this can support a diagnosis.* The nodal vessels in malignancy, on the other hand, are coiled and tortuous.

Numerous authors have reported the efficacy of EBUS-TBNA in the diagnosis of sarcoidosis with typical yields of granulomas of 80–90%.^27–29 Most of these cases were in stage 1 sarcoidosis, as in the prospective study of Nakajima et al. where patients had both EBUS-TBNA and TBLB, EBUS-TBNA was shown to be significantly more sensitive than TBLB. It remains to be seen whether EBUS-TBNA will replace TBLB in stage 2 or 3 sarcoidosis, where it would be more likely to obtain granulomas. Randomized studies are under way to answer this question.

Tremblay et al. reported a study comparing EBUS-TBNA (22-gauge) with blind TBNA (19-gauge) for sarcoidosis lymph nodes; the respective sensitivities were 96% and 73%, which was statistically significant in favor of EBUS-TBNA.³⁰ These authors questioned the need to continue with TBLB in stage 1 sarcoid patients given the increased risks of pneumothorax and bleeding with TBLB.

Case 5: Biopsy of Benign lymph nodes

A 65-year-old man presented with an incidentally found 3 × 2 cm lesion in the left lower lobe (Fig 6a). Sputum cytology showed squamous cell carcinoma. A chest X-ray had been performed during follow up for hepatitis C. He had a 90 pack year smoking history but had normal lung function. The chest CT also showed a number of small lymph nodes in subcarinal and both hilar regions. He therefore had a PET scan, this showed that the lesion in the left lower lobe was PET-avid, consistent with primary lung cancer (Fig 5b). The PET also showed multiple symmetrical lymph nodes with moderate PET-avidity in hilar and mediastinal regions. The PET node pattern suggested benign disease, but clearly the nodes needed tissue diagnosis before referring for a lobectomy. At EBUS-TBNA we sampled 11L, 7L and 4 R lymph nodes. All nodes were less than 1 cm in maximum dimension. All specimens showed only carbon laden macrophages and some lymphocytes, but no malignant cells. The patient subsequently underwent lobectomy which confirmed a 3 cm squamous cell carcinoma. Hilar and number 9 lymph nodes were sampled, showing no evidence of malignancy; the nodes showed evidence of anthracosis and fibrosis.

Figure 6. Case 5. (a) Chest computerised tomography. Red arrows point to small mediastinal and hilar nodes which show low grade uptake on PET scan. (b) Positron emission tomography scan showing peripheral mass as well as multiple symmetrical hilar and mediastinal lymph node positivity.

EBUS-TBNA offers great advantage in this type of case. With blind TBNA if a sample is negative for malignancy this may be because the needle was not in the lymph node. However, imaging can confirm the needle location with EBUS-TBNA which greatly reduces uncertainty. Then, if there are moderate to abundant lymphocytes, it is reasonably definitive that the node is benign. The overall false negative rate for malignancy is 4–8%, being slightly higher in staging cases as the nodes tend to be smaller.^7,11 Sometimes, as in this case, confirmation of true negative status is made at the time of surgical resection. Other factors which support a benign diagnosis are abundant carbonaceous material or silica. If PET is available it can be helpful to see low- to intermediate-grade node positivity in a bilateral symmetrical pattern which includes the hilar. We often see this in patients with occupational exposure to dusts and silica who are referred for a tissue diagnosis of abnormal hilar or mediastinal nodes. In these cases if surgical confirmation of benign disease is not made, it is necessary to observe with repeat CT scans for 6 to 12 months depending on circumstances.

Complications

EBUS-TBNA is a safe technique and no serious adverse effects have been reported. As with blind TBNA the main adverse effect is damage to the biopsy channel of the bronchoscope. This occurs when the sheath of the dedicated needle is not outside the biopsy channel when the needle is advanced. Learning to do this correctly is very important. Recently there have been two reports of infection arising in cystic structures (bronchogenic cysts or thyroid cysts) aspirated by EBUS-TBNA.^31,32 If fluid is left after the aspiration, infection can occur; it is best to commence antibiotics after the procedure in such cases. Infection does not seem to be a problem when the fluid is completely drained.³³ Alternatively, it may be best not to aspirate such nodes or masses if the EBUS image is distinctly fluid. Another adverse effect was found by the author early in his experience; a 40-year-old man had multiple lymph nodes positive on PET scan and aspirates were performed at 4R and subcarinal positions. The case was done under local anesthetic and the patient was coughing excessively. During one pass performed using the piggyback technique, the patient coughed heavily causing the TBNA needle to break and remain lodged in the bronchial wall. The needle tip was immediately completely removed with bronchoscope forceps and there were no lasting effects. We have since changed our procedure to avoid the piggyback method and, of course, curtail the procedure if there is excessive coughing. There is one reported case of pneumothorax after EBUS-TBNA.¹⁷

Conclusion

EBUS-TBNA has a wide range of applications. It has become an essential part of our bronchoscopy technique. It has gained an important place in lung cancer staging, often replacing mediastinoscopy. After a training period of 20–30 cases the technique is suitable for all experienced bronchoscopists.

Acknowledgements

We gratefully acknowledge the assistance of other scientists and pathologists from the Anatomical Pathology Department Royal Brisbane and Women's Hospital, including Alan Nott, Kylie Coleman and Stephen Donaldson.

Footnotes

• * 

  Kurimoto N, World Congress Bronchology, Tokyo March 2008.