Staging for non-small cell lung cancer (NSCLC) requires accurate assessment of the mediastinal lymph nodes which determines treatment and outcome. As radiological staging is limited by its specificity and sensitivity, it is necessary to sample the mediastinal nodes. Traditionally, mediastinoscopy has been used for evaluation of the mediastinum especially when radical treatment is contemplated, although conventional transbronchial needle aspiration (TBNA) has also been used in other situations for staging and diagnostic purposes. Endobronchial ultrasound guided transbronchial needle aspiration (EBUS-TBNA) offers a minimally invasive alternative to mediastinoscopy with additional access to the hilar nodes, a better safety profile, and it removes the costs and hazards of theatre time and general anaesthesia with comparable sensitivity, although the negative predictive value of mediastinoscopy (and sample size) is greater. EBUS-TBNA also obtains larger samples than conventional TBNA, has superior performance and theoretically is safer, allowing real-time sampling under direct vision. It can also have predictive value both in sonographic appearance of the nodes and histological characteristics. EBUS-TBNA is therefore indicated for NSCLC staging, diagnosis of lung cancer when there is no endobronchial lesion, and diagnosis of both benign (especially tuberculosis and sarcoidosis) and malignant mediastinal lesions. The procedure is different than for flexible bronchoscopy, takes longer, and requires more training. EBUS-TBNA is more expensive than conventional TBNA but can save costs by reducing the number of more costly mediastinoscopies. Revenue based tariff systems have been slow to reflect the innovation of techniques such as EBUS-TBNA. In the future, endobronchial ultrasound may have applications in airways disease and pulmonary vascular disease.
- Endobronchial ultrasound
- transbronchial needle aspiration
- lung cancer
- chest imaging
- adult thoracic medicine
- respiratory tract tumours
Statistics from Altmetric.com
- Endobronchial ultrasound
- transbronchial needle aspiration
- lung cancer
- chest imaging
- adult thoracic medicine
- respiratory tract tumours
Current surgical staging for lung cancer with mediastinoscopy is more invasive, requires general anaesthesia, an operating theatre and thoracic surgeon, has rare but serious complications, and is expensive. Linear probe endobronchial ultrasound guided transbronchial needle aspiration (EBUS-TBNA) offers a minimally invasive option for staging the mediastinum in suspect lung cancer but also in the diagnosis of mediastinal lesions accessible from the airway. EBUS-TBNA has a range of defined clinical indications and offers advantages over other mediastinal staging techniques. Its performance in staging is comparable to mediastinoscopy in many respects and superior to conventional transbronchial needle aspiration (TBNA). It also has an important role in diagnosis of inaccessible lung cancers as well as benign conditions such as sarcoidosis and tuberculosis. There are some important practical differences compared to flexible bronchoscopy as well as important training and cost considerations. In particular, EBUS-TBNA tariffs from Payment by Results have been slow to reflect this innovation. This paper will review the indications, advantages and performance of EBUS-TBNA against other lung cancer staging techniques as well as discussing some of the practical, training and financial issues before briefly discussing more specialist/future applications.
Rationale for EBUS-TBNA and brief historical summary of development
For patients with lung cancer, timely diagnosis and accurate staging are pivotal for facilitating the appropriate treatment. Staging algorithms and treatment practices have been guided by mediastinal lymph node metastases which determine outcome.1 Radiological investigations (contrast enhanced computed tomography (CT) of the chest and upper abdomen and positron emission tomography (PET)) are limited by false negative and positive results, and therefore the need for tissue confirmation of abnormal results.2 Factors contributing to the limitations of CT here are interobserver variability,3 the need for an interface to allow differentiation between adjacent soft tissues,4 benign causes of enlarged adenopathy, and water dense structures such as mucoid plugs and secretions which may mimic solid structures.
CT and PET, however, do allow targeting of invasive mediastinal sampling techniques, guide further diagnostic decisions on how best to obtain tissue to confirm the diagnosis and metastatic spread in one visit, or assess suitability for radical treatment. Moreover, flexible bronchoscopy is effective at diagnosis of direct endoluminal tumours, but many lung cancers present in the mediastinum. Conventional TBNA can allow the mediastinal nodes to be sampled, but without direct vision, relying on correlation of CT imaging and anatomical knowledge. Expanding the view of nodes adjacent to the airway by imaging to allow real-time sampling under direct vision is therefore a logical development.
Transthoracic ultrasound is limited by the reflection of ultrasound wave by air prompting the development of endoluminal ultrasound. Oesophageal endoscopic ultrasound guided fine needle aspiration (EUS-FNA) was developed in the 1980s for evaluation of gastrointestinal malignancies, especially the oesophagus. Because of the proximity of the oesophagus to mediastinal structures, EUS-FNA was used in lung cancer to sample accessible lymph nodes.5 However, airway interference prevented better vision of other lymph node stations and some stations were not accessible via the oesophagus, but would be accessible from the airway, hence the development of the closely related EBUS-TBNA in the 1990s.6 7 Modifications included smaller probes to allow adequate ventilation and instruments that would fit down the working channel. Initially, a radial probe was developed and then subsequently the linear probe.8 9
Before minimally invasive endoscopic needle aspiration techniques were available, mediastinal staging would normally be achieved by surgical staging by cervical mediastinoscopy (accessing lymph node stations 1–4 and 7, see figure 1), or an anterior mediastinotomy (accessing stations 5–6). Notably, neither of these techniques provide access to stations 8–9 (accessible via a thoracoscopic approach or by EUS-FNA) or stations 10 and 11 (accessible via EBUS-TBNA). Endoscopic needle aspiration techniques (EBUS-TBNA and EUS-FNA) have allowed real-time sampling of mediastinal nodes and other more distal structures (for EUS-FNA). EBUS-TBNA can be performed at the same time as flexible bronchoscopy. It gives access to the same stations as mediastinoscopy but also the hilar nodes at stations 10 and 11 (see figure 1). EUS-FNA can also be performed at the same visit and give access to stations 4L, 5–9 (see figure 1) as well as the left adrenal, left lobe of liver.
What are the indications for EBUS-TBNA?
The most common indication for EBUS-TBNA would be staging the mediastinum for suspected non-small cell lung cancer (NSCLC) which is important in determining outcome and treatment (discussed in lung cancer staging section). EBUS-TBNA is also used extensively for diagnosis of both NSCLC and small cell lung cancer, as often there is no endoluminal tumour at bronchoscopy and this avoids the need for either a CT guided lung biopsy or mediastinoscopy (which may not be appropriate, especially if the patient is unlikely to be having surgical treatment). EBUS-TBNA is also used for diagnosis of unexplained mediastinal lymphadenopathy accessible to the major airway including benign conditions such as sarcoidosis or tuberculosis. EBUS-TBNA is also used as a research tool for tissue banking samples for later studies.
What are the contraindications for EBUS-TBNA?
EBUS-TBNA is well tolerated, but sampling from the mediastinal nodes should not be performed with patients on warfarin (international normalised ratio (INR) should be <1.4 ideally) or clopidogrel (both should be stopped for a week before the procedure), or known coagulation or platelet function disorders because of bleeding risk in the mediastinum. Otherwise, patients need to be fit enough to have a flexible bronchoscopy10 and to be able to lie flat as oral intubation is performed. EBUS-TBNA should be postponed for at least 6 weeks after myocardial infarction and is contraindicated in the presence of ongoing myocardial ischaemia, arrhythmias or severe hypoxaemia at rest. EBUS-TBNA is not usually clinically appropriate if lymphoma is a likely possibility. Otherwise, as for flexible bronchoscopy, if only active symptom control and palliative care is proposed, then such a procedure is not appropriate simply to confirm a clinical diagnosis.
Why is sampling the mediastinal nodes so important for lung cancer staging?
Mediastinal staging is often the key step to determine resectability in NSCLC as mediastinal metastatic lymph node involvement in NSCLC correlates with extrathoracic metastases.11 Mediastinal nodal status has prognostic importance and influences initial management. As previously discussed, radiological staging of the mediastinum with CT and PET has limited sensitivity and specificity so histology is required.2
Mediastinal lymph nodes are sampled when enlarged on CT short axis (>1 cm) and/or metabolically active on PET or PET/CT (with a high standard uptake value (SUV)). Mediastinal sampling techniques include mediastinoscopy, EBUS-TBNA and conventional TBNA. (EUS-FNA can also be combined with EBUS-TBNA to sample stations 5–9 (see figure 1) and allow access to the entire mediastinum, as some studies have shown that diagnostic yield is increased compared to either procedure alone.12–14) The relative performance of these staging techniques is now reviewed.
EBUS-TBNA compared to radiological staging
Recent studies have compared EBUS-TBNA to radiological staging (CT and PET) in preoperative staging and examined the role of EBUS-TBNA in patients with a normal mediastinum according to radiological staging.15 16 EBUS-TBNA was superior to CT and PET in sensitivity (92.3% vs 76.9% and 80.0%, respectively), specificity (100% vs 55% and 70%, respectively) and accuracy (98% vs 61% and 73%, respectively) in a prospective study of 102 patients with potentially operable lung cancer.16
A further prospective study of 117 patients with enlarged mediastinal nodes of 5–20 mm (nodal metastases prevalence 26%) on PET/CT revealed similar superiority of EBUS-TBNA in sensitivity (90% vs 70%, respectively), specificity (100% vs 60%, respectively) and accuracy (97% vs 62%, respectively).17 The histology of NSCLC influenced the superiority of EBUS-TBNA, with adenocarcinoma favouring EBUS-TBNA (probably related to the higher rate of mediastinal metastases18 and lower PET SUV with this tumour type19), but no clear superiority for either technique with squamous cell carcinoma.
In a further study of 97 NSCLC patients (confirmed at subsequent thoracotomy or mediastinoscopy) with a “normal” radiological mediastinum, EBUS-TBNA was used to sample mediastinal lymph nodes with a mean diameter of only 7.9 mm. Eight patients (8.2%) had mediastinal lymph node metastases confirmed on EBUS-TBNA (nearly all adenocarcinoma) despite ‘normal’ radiological staging. EBUS-TBNA performed with a sensitivity of 89%, specificity of 100% and negative predictive value of 98.9%.15 EBUS-TBNA may therefore have a role in preoperative staging even in stage 1 disease, and it has superior diagnostic sensitivity compared to radiological staging as well as providing a tissue diagnosis. It should also be appreciated that the performance of PET in all of these studies is likely to have been reduced further as they involved nodal diameters down to 5 mm, below the 10 mm cut-off for PET.
EBUS-TBNA versus conventional TBNA
EBUS-TBNA is superior to conventional TBNA (ie, without real-time imaging guidance). In a recent systematic review, the pooled sensitivity for conventional TBNA was 76% (EBUS-TBNA was not included in this analysis).20 In contrast, EBUS-TBNA achieved a sensitivity of 95% in a prospective cohort of 108 patients with suspected NSCLC, with a 90% negative predictive value and 96% accuracy. EBUS-TBNA also avoided 50 more invasive sampling procedures in the same study.21 In a comparative trial, EBUS-TBNA (albeit with a radial probe) was superior to conventional TBNA at stations other than station 7 (84% vs 58% positive) compared to station 7 (86% vs 74% positive, albeit favouring EBUS-TBNA) for 200 patients with suspected NSCLC.22 This is likely to reflect the relative technical ease of station 7 for conventional TBNA compared to other nodal stations.
EBUS-TBNA versus mediastinoscopy
EBUS-TBNA (and EUS-FNA) has an equivalent sensitivity to mediastinoscopy on the basis of systematic reviews and meta-analyses with superiority to conventional TBNA (see table 1).20 23–26 The performance of EBUS-TBNA varies according to the cohort selection criteria; in patients pre-selected with CT or PET positive nodes the sensitivity is higher than in unselected patients (sensitivity 94% vs 76% respectively).24 The lower sensitivity for mediastinoscopy may be partially attributed to the lower disease prevalence in the mediastinoscopy studies included in these reviews. The negative predictive value of mediastinoscopy remains higher than for EBUS-TBNA (or EUS-FNA and conventional TBNA).
There have been very few comparative studies of mediastinoscopy and EBUS-TBNA. Existing data are not consistent. A recent prospective crossover trial of 66 patients compared the two techniques in patients with suspected NSCLC.27 The disease prevalence was high at 89% and the diagnostic yield was significantly in favour of EBUS-TBNA (91% vs 78%). Discordance between the two techniques occurred at station 7. The sensitivities and negative predictive values were also in favour of EBUS-TBNA (87% and 78% vs 68% and 59%, respectively). There were, however, no significant differences in determining true N status (93% vs 82%).
There is another ongoing study (due for final publication in 2009) evaluating EBUS-TBNA versus mediastinoscopy for nodal staging of patients with suspected or confirmed lung cancer.28 Initial data for 33 patients revealed a lower sensitivity and slightly inferior negative predictive value for EBUS-TBNA (77% and 86% vs 85% and 90%, respectively) with similar accuracies (91% vs 94% for EBUS-TBNA and mediastinoscopy, respectively). The prevalence of N2 or N3 disease was lower at 39% and EBUS-TBNA sampled 3.4 nodes per patient (vs 4.0 nodes per patient for mediastinoscopy). Importantly, three patients were upstaged by mediastinoscopy from N0 (on EBUS-TBNA) to N2 indicating it may not completely replace surgical staging. The results of a further randomised multicentre controlled trial of EBUS-TBNA and EUS-FNA versus mediastinoscopy (ASTER trial) should be available in 2012.29
Other non-comparative studies have used surgical staging for EBUS-TBNA negative cases only. Szlubowski et al performed an extensive surgical mediastinal lymphadenectomy on all negative EBUS-TBNA results in a prospective cohort study of 206 patients.30 The sensitivity of EBUS was 89% (disease prevalence 57%) but the negative predictive value was 83.5%. The reason for false negative results was mainly due to small metastases not inaccessibility to EBUS-TBNA. Rintoul et al, in a prospective study of 109 patients with PET positive nodes, detected a sensitivity of 91% of EBUS-TBNA (where pathological confirmation was possible) but a negative predictive value of only 60% (disease prevalence 71%), as seven of the 19 surgical biopsies in EBUS-TBNA negative patients were positive for malignancy.31 These data highlight the need for surgical evaluation of EBUS-TBNA negative nodes where the pre-test probability (and therefore disease prevalence) is high.
Therefore, available data suggest EBUS-TBNA reduces the number of mediastinoscopies but the variability in performance in these two studies may relate to the significant variation in disease prevalence. When disease prevalence is high, existing data favour EBUS-TBNA, but when it is moderate then surgical staging appears superior. Patients with a high pre-test probability of lung cancer with negative EBUS-TBNA should undergo mediastinoscopy if possible to corroborate the EBUS-TBNA.
Guidelines from the American College of Chest Physician guidelines23 advocate EBUS-TBNA instead of mediastinoscopy to stage the mediastinum with discrete N2 or N3 disease or bulky mediastinal disease, utilising mediastinoscopy for situations where radical treatment is intended; this may include single station N2 disease given its better prognosis recognised in the revised lung cancer TNM staging system.32 For staging the hilar nodes, EBUS-TBNA is the investigation of choice as these are not accessible at mediastinoscopy. This is supported by the findings of a recent multicentre study of 213 patients with CT or PET positive hilar nodes demonstrating a 91% sensitivity of EBUS-TBNA for hilar nodes with similar results to the central nodes.33
Diagnosis of mediastinal lymphadenopathy
EBUS-TBNA is increasingly used as a diagnostic tool for unexplained mediastinal and hilar lymphadenopathy due to malignant or benign disease. EBUS-TBNA is commonly the sole method of obtaining tissue for diagnosis in lung cancer patients due to the lack of an endobronchially accessible lesion (22% in a typical cohort of patients with a high pre-test probability, unpublished observations).
Apart from lung cancer, EBUS-TBNA can be used to diagnose benign disease (commonly sarcoidosis and tuberculosis) although the volume of studies is much smaller. A recent small case series demonstrating a diagnostic yield of 78% in a newly established EBUS-TBNA service in patients without accessible airway or parenchymal involvement demonstrates the potential utility of EBUS-TBNA here.34 Performance of EBUS-TBNA in sarcoidosis has varied according to patient selection criteria. In a study of 48 patients with suspected sarcoidosis, EBUS-TBNA had a sensitivity of 85%.35 It should be noted that the nodal diameter in this study was sub-centimetre and all stages of sarcoidosis without necessarily a pre-EBUS-TBNA CT scan, reflecting a real life cohort. In contrast, a greater sensitivity of 92% was obtained by Wong et al in 65 patients with nodes >1 cm, stage 1–2 sarcoidosis who all had a pre-sampling CT scan.36 A similar sensitivity of 93% was obtained in a further study of 15 patients by Oki et al using similar selection criteria.37
EBUS-TBNA can also be used to diagnose lymphoma in certain settings although this is not standard practice at the current time. In a study of 25 patients, mediastinal lymphoma was diagnosed by EBUS-TBNA with 91% sensitivity and 93% predictive value in a population with lymphoma prevalence of 44%.38 It should be noted this was a retrospective review using rapid on-site evaluation for cytopathology (ROSE) and flow cytometry techniques which are not available in all centres.
The sample tissue volume in EBUS-TBNA and TBNA techniques is particularly crucial for diagnosis of lymphoma and benign disease, and also for enabling mutation testing for growth factors which may have prognostic importance.39 A recent study of 75 patients assessing subcarinal (non-lung cancer) lesions showed 1.15 mm mini-forceps biopsies were superior to both 22 gauge and 19 gauge TBNA needle biopsies (sensitivities 88% vs 36% and 49%, respectively), although the standard four biopsies were not taken in this study which accounts for the particularly low TBNA sensitivities here.40 The superior sensitivity was particularly evident in sarcoidosis (88% vs 36% respectively) and lymphoma (81% vs 35% respectively). However, a randomised controlled trial confirmed the superiority of EBUS-TBNA to 19 gauge conventional TBNA needle biopsies in 50 patients with suspected sarcoidosis with hilar and/or mediastinal adenopathy.41 Sensitivities were in favour of EBUS-TBNA (83% vs 61%, respectively). Real-time sampling may therefore be as important a factor as sample size.
EBUS-TBNA: sonographic and histological predictors of malignancy
Sonographic appearances at EBUS-TBNA and non-malignant histological characteristics of biopsies can be helpful in predicting malignancy. Sonographic nodal characteristics at EBUS-TBNA have been studied in a recent Spanish study of 161 patients with 51% disease prevalence.42 Sampling nodes by EBUS-TBNA with a short axis diameter >5 mm and a short to long axis ratio >0.5 allows identification of all malignant nodes. A short axis diameter >2 cm was highly predictive of malignancy (>90% probability) and a spherical shape with a short to long axis ratio of 1 (55% probability).
Granulomatous reactions are also reported in EBUS-TBNA in the presence of NSCLC. A recent prospective study of 50 patients reported granulomatous reactions in 4.3%, none of whom had nodal metastases (confirmed by mediastinoscopy).43 A granulomatous EBUS-TBNA result in the setting of NSCLC may therefore be a useful negative predictor for nodal metastases.
What are the strengths and weaknesses of EBUS-TBNA compared to other staging techniques?
EBUS-TBNA has both strengths and weaknesses which must be fully appreciated to utilise it to its maximum (see table 2 for summary).
Advantages over conventional TBNA
EBUS-TBNA has a superior yield to conventional TBNA at stations other than station 7 (2–4, 10–11) accessible to either technique (see table 1). The target lesion is directly visualised and sampled under real-time imaging, reducing the chances of major vessel puncture. The EBUS needle system is more robust than conventional TBNA needles and a larger tissue core is obtained as the aspirate is longer. EBUS-TBNA also allows the possibility of image capture while sampling the node for quality control measures.
Advantages over mediastinoscopy
EBUS-TBNA accesses the hilar nodal stations, unlike mediastinoscopy, which may be particularly helpful for diagnosis of sarcoidosis and tuberculosis. EBUS-TBNA is safe and less invasive than mediastinoscopy (which leaves a neck scar, has a 1.4–2.3% risk of important complications, and a 0.5% risk of major complications, including death).44 45 It is a day case procedure using conscious sedation without the need for general anaesthesia or overnight admission and can be performed by pulmonologists as well as thoracic surgeons. Mediastinoscopy is a far longer procedure than EBUS-TBNA (2 h 24 min in one tertiary UK centre46 versus an average of 30 min for EBUS-TBNA by experienced operators13).
Disadvantages compared to other staging techniques
EBUS-TBNA uses a small 22 gauge needle which can yield a tissue core but the volume of tissue is less than at mediastinoscopy which may result in false negative EBUS-TBNA results, particularly in lymphoma, and reduce the amount of tissue for growth factor receptor profiling. The negative predictive value of EBUS-TBNA is less than that for mediastinoscopy (see table 1) so patients with a high pre-test probability of lung cancer with a negative EBUS-TBNA should have this corroborated by mediastinoscopy at the current time. EBUS-TBNA is more expensive and technically demanding than conventional TBNA and takes longer. There is the potential for inadvertent needle penetration of the inner sheath causing costly damage. There are also challenges in Payment by Results regarding tariff evolution with innovation.
Practical issues relating to the procedure itself differing from flexible bronchoscopy
EBUS-TBNA allows ultrasound imaging of structures within and adjacent to the airway wall in contrast to conventional TBNA. This allows imaging of lymph nodes, blood vessels, the heart and tumour masses (see figure 2). EBUS-TBNA also allows real-time imaging of the EBUS-TBNA needle in the target node/mass throughout the whole sampling procedure unlike conventional TBNA. Importantly, a separate flexible bronchoscopy is needed to examine and sample the distal parts of the tracheobronchial tree (if necessary) due to the different size and design of the endobronchial ultrasound (EBUS) bronchoscope and the fact that the endoscopic image at EBUS-TBNA is inferior to normal white light flexible bronchoscopy, so subtle mucosal lesions may be missed.
Equipment and technique
The EBUS bronchoscope tip is larger and bulkier than for flexible bronchoscopes (external diameter 6.9 mm vs 5–6 mm, see figure 3A).47 The tip of the EBUS bronchoscope is also stiffer and more fragile than a flexible bronchoscope due to the ultrasound probe. Oral intubation of the supine patient from behind is the only route of choice given the higher external diameter. Sedation practices for EBUS-TBNA vary from conscious sedation with titrated midazolam and fentanyl to general anaesthesia with a laryngeal mask in situ. Monitoring, local anaesthesia and oxygenation are as for standard bronchoscopy. Optimal sedation and monitoring are especially important because the procedure typically takes longer than a flexible bronchoscopy, especially if multiple stations are sampled.
The EBUS bronchoscope tip contains the ultrasound probe and is enveloped by a balloon sheath (see figure 3A). The balloon is inflated to improve ultrasound image quality (if poor) by improving contact between the ultrasound probe and airway wall. In practice, balloon inflation is seldom needed. The linear/convex probe is the most common probe (and the main focus of this review) used in EBUS-TBNA (see figure 3A,B).
The frequency of the probe is 7.5 MHz giving a depth of penetration of 9 cm (much lower frequency and higher penetration than the radial probe). The EBUS bronchoscope system provides two images: an endoscopic image at an obliquely angled view of 30° forward (see figure 3B) and an ultrasound image at an angled forward view of 90° parallel to the EBUS bronchoscope shaft. The likelihood of vascular puncture is reduced by colour flow and power Doppler features which allow differentiation of vascular structures (see figure 3C).
The radial probe is used more selectively but does have more specialist applications in some centres. Radial probes give a 360° view (see figure 4) allowing assessment of airway tumour infiltration which may guide endobronchial therapy. The probe frequencies range from 20–30 MHz giving greater resolution of <1 mm but reduced penetration of 5 cm.48 49 The probes can either fit into a 2.8 mm working channel or even a 2 mm working channel; the latter can be used for peripheral pulmonary nodule imaging and sampling via a guide sheath in subsegmental bronchi. Radial EBUS-TBNA sampling does not have the advantage of real-time sampling but is performed by sequential sampling.
A one-person technique is as feasible and practical with routine assistance from a bronchoscopy nurse as for other flexible bronchoscopic techniques. For NSCLC staging, the higher stage nodes are sampled first to prevent upstaging by sample contamination.
The EBUS needle is a 22 gauge needle housed in a sheath as for conventional TBNA (see figure 3A,B). The technique for performing the EBUS-TBNA when in the node is similar to conventional TBNA but with a few modifications. The system has to be locked with the sheath just visible to prevent fibre damage. An internal stylet is used and needs to be withdrawn and jiggled on entry to the node to get rid of debris, cartilage, and blood clots that may occlude the needle lumen. The stylet is then replaced with the needle system. The distance to the node (target) centre is also measured from the EBUS ultrasound image and the puncture depth assigned on the EBUS needle system (0.5–4 cm typically).
The stylet is reintroduced after sampling to expel the tissue core (the sample size is bigger and longer than for conventional TBNA). Air and then saline is injected to expel sample remnants after removing the stylet. The EBUS needle and stylet are cleaned to remove clot and debris. Two passes per station is sufficient if good visible tissue cores are obtained,50 although many centres will perform four passes per node until building up significant experience.
Safety and complications
EBUS-TBNA is tolerated as well as flexible bronchoscopy with a similar complication risk.22 51 It is therefore a safe, minimally invasive procedure when performed by an experienced operator in an appropriate patient.10 There are extremely rare reports of pneumomediastinum, pneumothorax and haemomediastinum, but a chest radiograph is not done routinely post-procedure.52 53 As real-time imaging is used, major vessel puncture is theoretically less likely than with conventional TBNA. Even if it occurs, major vessel puncture is not a problem during conventional TBNA as described in the aorta, even if there is an unintended haematoma.52 Other EBUS-TBNA operators intentionally traverse the pulmonary artery for left hilar mass sampling without problem.54 Infectious complications have also been rarely reported.55
Sample handling and processing
As for conventional TBNA, many EBUS-TBNA centres use formalin pots (as well as saline pots for tuberculosis culture) and liquid cytology bottles for samples, or smear part of the sample onto glass slides and fix in alcohol. Data from conventional TBNA suggest the smear technique is superior to liquid cytology and results in less cellular distortion which may be especially important in granulomatous disease.56–58 ROSE is the gold standard and is known to improve the yield of needle aspiration techniques by confirming an adequate sample using the smear technique, but is not available in many centres because of limited resources and cost.59
Training issues and competency
With all new technologies, demonstration of competency and being fully trained is a prerequisite to developing a service. There is variation in the statements from national bodies which differ on EBUS-TBNA training recommendations and they only refer to radial probe (and not the more widely used linear probe) EBUS-TBNA. The European Respiratory Society/American Thoracic Society interventional pulmonology statement recommends 40 supervised (radial probe) EBUS-TBNA procedures and 25 per year to maintain skills.60 The American College of Chest Physicians guidelines advise a minimum of 50 supervised (radial probe) EBUS-TBNA procedures and then 5–10 procedures per year to maintain skills.61 Clinical experience suggests the linear probe system demands the learner to be fully trained in flexible bronchoscopy and to conduct at least 40–50 EBUS-TBNA procedures supervised by an experienced EBUS-TBNA bronchoscopist, although there is no current evidence to confirm these observations.22
Theoretically, radial probe EBUS-TBNA may have a longer learning curve as sampling is not done in real-time, and airway wall infiltration ultrasound interpretation at higher frequency requires more complex interpretation of a more detailed ultrasound image compared to linear probe images. Nevertheless, this is not to underplay the demands of linear probe EBUS-TBNA which should not be underestimated. Previous conventional TBNA and ultrasound (especially supraclavicular node ultrasound62) experience is helpful but not an absolute prerequisite. Other learning materials include courses with simulators, models and live animals which are useful as an introduction but, in reality, there is no better substitute than time with an experienced mentor. Gastroenterologists, interventional radiologists or surgeons who are experienced in EUS-FNA may be an additional source of support (discussed further below).
From experience, the procedure is best learnt by dividing it into two parts. The EBUS needle system and the EBUS-TBNA itself takes 20–25 procedures for an operator to become fully dexterous with sampling nodal stations of all difficulties and size while obtaining good tissue cores. The intubation and manipulation of the EBUS bronchoscope, target location and ultrasound image acquisition and interpretation takes at least 40–50 procedures in the hands of experienced flexible bronchoscopists (this statement is based on observation rather than evidence at the current time). As well as finding the target and interpreting, obtaining and maintaining the EBUS ultrasound image, the more immediate challenges are acclimatising to the altered angle of view, handling, weight and fragility of the EBUS bronchoscope.
In the context of lung cancer staging, EBUS-TBNA and EUS-FNA can be regarded as complementary tools, as combined EBUS-TBNA/EUS-FNA can provide access to most areas of the mediastinum as well as access to left lobe of the liver, left adrenal and other areas in the abdomen.12 14 EUS-FNA is currently performed predominantly by gastroenterologists, interventional radiologists and surgeons, but a number of pulmonologists are also doing this procedure sometimes in conjunction with an EUS-FNA operator. A combined EBUS-TBNA/EUS-FNA procedure performed by pulmonologists will be the longer term objective. Accreditation for EUS-FNA for interventional pulmonologists can be a challenge in some health care systems, although cardiologists perform transoesophageal echocardiography without these obstacles.
It is essential with all new technology to consider the cost implications in a resource rationed health care system. The major costs are the capital costs of the EBUS bronchoscope and ultrasound processor±printer. There are, however, significant running costs mainly due to the disposable EBUS needle apparatus which is significantly more expensive than conventional TBNA needles (approximately £150–175 vs £40, respectively) and other accessories such as valves, syringes and balloons. A recent projected cost analysis in a UK centre has calculated the total capital costs of an EBUS-TBNA system to be £10 4465 over a 5 year period.63
Staff costs are higher because the procedure takes longer than a standard flexible bronchoscopy (at least 30 min assuming a number of nodal stations are sampled at least three passes per station), and also sometimes because an additional flexible bronchoscopy has to be performed after or before the EBUS-TBNA for endobronchial sampling (in 32.5% of unselected cases of suspected lung cancer, unpublished observations). If EUS-FNA is performed at the same sitting, this further lengthens the procedure. ROSE is often not locally available but would be a further cost consideration if used. Repair costs of the hybrid EBUS bronchoscope are higher than for a conventional flexible bronchoscope.
The prevalence of nodal metastases is important in approach. A recent economic analysis in a European EUS-FNA centre confirmed EUS-FNA would be cheaper (£11 115 per patient) than combined EBUS-TBNA/EUS-FNA approach (£11 205 per patient) in low prevalence (<33%) cohorts, whereas the combined approach would be more economical in higher prevalence cohorts (more likely in real world practice).64 Both strategies with needle aspiration techniques were cheaper than mediastinoscopy. The principal cost saving potential of EBUS-TBNA±EUS-FNA is by staging and diagnosing lung cancer in one visit and reducing the need for additional staging tests, especially mediastinoscopy, and thereby also increasing thoracic surgical capacity.
For health care systems operating tariff based revenue, two other challenges exist. Firstly, it is imperative that the tariffs are updated in line with innovation—for example, in the UK there is no specific EBUS-TBNA tariff which is not sufficiently distinguished from a standard flexible bronchoscopy tariff despite the increased cost, complexity and time of EBUS-TBNA.65 66 Secondly, correct procedure coding remains pivotal as exemplified by studies and reports showing significant rates of coding error in the UK health care system.67–69
Finally, EBUS-TBNA does involve significantly higher equipment cost than for conventional TBNA, with more expensive needles and a longer procedure time as already discussed. Conventional TBNA has a very good performance, especially for large, centrally positioned (>25 mm) nodes, although the amount of tissue recovered is less. For patients with bulky (>25 mm) central nodal disease, conventional TBNA is a reasonable alternative to EBUS-TBNA and is by far the cheaper option. However, EBUS-TBNA is superior, especially at more distal stations and for all smaller nodes. Therefore, the cost saving potential for EBUS-TBNA by avoiding mediastinoscopies is far greater than conventional TBNA and this can offset the higher initial and running costs. In centres where EBUS-TBNA is not available, conventional TBNA should certainly be utilised as first line for those with central bulky mediastinal disease.
The technology of EBUS has other potential applications in other disease processes, and the indications for EBUS are likely to increase in the future as the potential of this technology becomes better understood. Radial probe EBUS can provide information about the airway wall and EBUS allows real time imaging of the central pulmonary vasculature. This has been exploited in two recent studies.
Aumiller et al reported a prospective pilot study illustrating the potential utility of EBUS to identify central pulmonary emboli, although this was not a blinded study and all pulmonary emboli had been diagnosed by prior CT pulmonary angiography.70 Soja et al measured airway wall thickness (using a radial probe) and correlated this with asthma severity.71
EBUS-TBNA represents a new technology in the field of bronchoscopy. The primary indications for EBUS-TBNA are staging NSCLC and the diagnostic assessment of mediastinal lymphadenopathy. EBUS-TBNA also has a diagnostic role in suspected benign disease, especially sarcoidosis and tuberculosis. It is a minimally invasive option as the first sampling staging procedure in suspected NSCLC with solitary hilar nodes, discrete N2 or N3 disease, or bulky mediastinal disease. Due to the current inferior negative predictive value of EBUS-TBNA, mediastinoscopy is still required for clarification of EBUS-TBNA negative nodes when the pre-test probability of lung cancer is high.
Currently, where radical treatment is contemplated, mediastinoscopy remains the preferred investigation for mediastinal staging, but this may well change to EBUS-TBNA with time if future larger controlled studies support a role for EBUS-TBNA in staging CT and PET negative sub-centimetre nodes. As such it could be used as the primary staging tool in situations where re-staging (with mediastinoscopy) is likely to be needed or as an alternative re-staging procedure itself.
EBUS-TBNA should be performed by those experienced or trained in the technique, and it is important not to underestimate the learning period for even experienced conventional bronchoscopists. The capital and running costs should also be thoroughly considered before setting up a service, although the principal financial argument for EBUS-TBNA will be to save money by avoiding mediastinoscopies. In the future, centres with appropriate throughput, technical expertise and funding, may seek to develop a combined EBUS-TBNA/EUS-FNA service, giving access to all the nodal stations.
Multiple choice questions; answers after the references
Which of the following statements are true or false?
The sensitivity of EBUS-TBNA is superior to conventional TBNA.
For patients for possible surgery, EBUS-TBNA is the first choice mediastinal staging procedure at the current time.
All patients with a negative EBUS-TBNA should undergo mediastinoscopy to validate the result because of the lower negative predictive value of EBUS-TBNA.
EBUS-TBNA can access more nodal stations than mediastinoscopy.
EBUS-TBNA can reduce costs.
Key learning points
Many lung cancers are not directly visible at conventional flexible bronchoscopy. EBUS-TBNA has evolved from EUS-FNA which was driven by the limitations of staging CT, conventional TBNA and transthoracic ultrasound.
Mediastinal nodal metastases in lung cancer determine outcome. Radiological staging has its limitations and tissue is often needed, but all mediastinal staging tools, including EBUS-TBNA, have their limitations.
EBUS-TBNA can offer a minimally invasive, day case option without general anaesthesia, scars or morbidity for patients and access more nodal stations (hilar nodes) and achieve equivalent performance sensitivity to mediastinoscopy for patients with suspected lung cancer and nodal metastases as well as increasing thoracic surgery capacity by avoiding mediastinoscopies.
EBUS-TBNA also has a role in re-staging after down-staging treatment, and may have a role in staging the radiologically normal mediastinum, but its negative predictive value is inferior to mediastinoscopy at the current time. At the moment, patients with a high pre-test probability of lung cancer should undergo mediastinoscopy after a negative EBUS-TBNA.
A combined EBUS-TBNA/EUS-FNA procedure may allow access to the entire mediastinum.
EBUS-TBNA can also diagnose benign conditions such as sarcoidosis and tuberculosis from positive lymph node culture. It is, however, not the investigation of choice for suspected lymphoma. It may have a future role in airway disease and pulmonary vascular disease.
EBUS-TBNA does require significant costs at the outset but is less expensive for health care providers and saves costs by avoiding mediastinoscopies (and reducing demand on thoracic surgery), although tariff based revenue systems in some countries (eg, the UK) need to recognise such innovation.
EBUS-TBNA requires significant training and meticulous technique as the equipment is more fragile and expensive with a different angled view to flexible bronchoscopy, although the exact requirements have not been specified or studied.
Current research questions
Are there sonographic features at EBUS-TBNA which predict malignancy?
Can EBUS-TBNA (with EUS-FNA) totally replace mediastinoscopy for lung cancer staging?
Are granulomatous reactions in NSCLC EBUS-TBNA samples a good negative predictor for nodal metastases?
Does EBUS-TBNA obtain sufficient tissue for growth factor response profiling?
Can EBUS-TBNA (with EUS-FNA) totally replace mediastinoscopy for diagnosis of benign disease (sarcoidosis and tuberculosis)?
What is the exact role of EBUS-TBNA in lymphoma diagnosis?
Does EBUS-TBNA have a role in the diagnosis of airways disease and airway remodelling?
Can EBUS-TBNA offer a diagnostic alternative to patients with central pulmonary emboli?
How many EBUS-TBNA procedures are required to be properly trained and how many procedures per year should be performed to maintain skills?
Is EBUS-TBNA more cost effective than conventional TBNA looking at patients with all variations of mediastinal lymph node metastases, rather than those with central bulky nodal disease alone?
▶ Annema JT, Versteegh MI, Veselic M, et al. Endoscopic ultrasound-guided fine-needle aspiration in the diagnosis and staging of lung cancer and its impact on surgical staging. J Clin Oncol 2005;23:8357–61.
▶ Herth F, Becker HD, Ernst A. Conventional vs endobronchial ultrasound-guided transbronchial needle aspiration: a randomised trial. Chest 2004;125:322–5.
▶ Wallace MB, Pascual JM, Raimondo M, et al. Minimally invasive endoscopic staging of suspected lung cancer. JAMA 2008;299:540–6.
▶ Herth FJ, Eberhardt R, Krasnik M, et al. Endobronchial ultrasound-guided transbronchial needle aspiration of lymph nodes in the radiologically and positron emission tomography-normal mediastinum in patients with lung cancer. Chest 2008;133:887–91.
▶ Detterbeck FC, Jantz MA, Wallace M, et al. Invasive mediastinal staging of lung cancer: ACCP evidence-based clinical practice guidelines, (2nd edition). Chest 2007;132(3 Suppl):202S–20.
Answers to the questions on page 113
True—especially for smaller distal nodal stations, for subcarinal nodes the superiority is less pronounced.
False—until further studies are available, mediastinoscopy is the staging procedure of first choice in most centres, although this may well change with time.
False—although the negative predictive value of mediastinoscopy is higher than EBUS-TBNA at the current time, this is only necessary if the pre-test probability of malignancy is high.
True—EBUS-TBNA can access the hilar nodes as well as all the stations accessible by mediastinoscopy.
True—although this seems counterintuitive as the EBUS bronchoscope, running, staff and repair costs are higher than for conventional bronchoscopy, in the longer term EBUS-TBNA is cost saving by avoiding mediastinoscopies which are even more expensive and therefore also increasing thoracic surgery capacity.
Competing interests none
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.