Chapter 12
The mediastinum
Felix J.F. Herth
Dept of Pneumology and Critical Care Medicine, Thoraxklinik University of Heidelberg, Heidelberg, Germany.
Correspondence: Felix J.F. Herth, Dept of Pneumology and Critical Care Medicine, Thoraxklinik University of Heidelberg, Röntgenstr. 1, D-69126 Heidelberg, Germany. E-mail: Felix.herth@med.uni-heidelberg.de
The number of patients suffering from lung diseases is increasing worldwide. Lung cancer, for example, is now one of the most common cancers. Many of these diseases having mediastinal lymph node involvement. Accurate diagnosis or possible staging of the mediastinum is important not only to determine the prognosis but to decide the most suitable treatment plan. CT, MRI, PET and PET-CT are used for noninvasive imaging of the mediastinum. If a cytological or histological conformation is required, EBUS-TBNA is the preferred test. This chapter will focus on technique, procedures and limitations.
Cite as: Herth FJF. The mediastinum. In: Laursen CB, Rahman NM, Volpicelli G, eds. Thoracic Ultrasound (ERS Monograph). Sheffield, European Respiratory Society, 2018; pp. 161–171 [https://doi.org/10.1183/2312508X.10007017].
The need to access the mediastinum has increased over the last two decades. Almost 100 years ago, lung cancer was a rare disease. Over time, diagnosis has increased and lung cancer is now the leading cause of cancer death worldwide [1, 2]. To treat the disease, proper staging is required and lymph node staging of the mediastinum is particularly important. Adequate lung cancer staging will also improve: research into the disease; accurate data comparison; and quality control. Mediastinal lymph node staging can be performed with radiological imaging, endoscopically or surgically [3]. CT scans, MRI [4], PET and PET-CT are useful noninvasive imaging techniques for staging lung cancer; however, they are neither sufficiently sensitive nor specific for the determination of mediastinal lymph node involvement [5].
In 1959, CARLENS [6] published his first report about mediastinoscopy. Following the report, the technique became and remained the gold standard for the mediastinum for several years.
The surgical option remained undiscussed, even after the implementation of improved imaging techniques, such as CT and PET. Not all lymph node stations were accessible with this invasive procedure (only 2, 4 and anterior 7); access to the posterior and inferior mediastinum was limited and required extended cervical mediastinoscopy or thoracoscopy. This invasive procedure also required general anaesthesia and clinical admission [7].
In 2004, technical improvements involving the integration of US technology into a flexible bronchoscope offered the possibility of minimally invasive examination of the mediastinum [8]. Using the EBUS scope enables imaging of the lymph nodes, lesions and vessels located beyond the tracheobronchial wall.
Nowadays, EBUS is routinely used and allows visualisation and sampling of mediastinal lymph nodes or masses adjacent to the tracheobronchial system. It is used in the diagnosis of benign diseases (such as sarcoidosis and tuberculosis) and in the inflammatory processes of malignant diseases. Initially used for diagnosing and staging lung cancer, the technology is now the preferred technique in the diagnosis of mediastinal diseases [9].
Equipment
Currently, several companies offer EBUS equipment. The difference in scope size, and the size of the working channels, is minimal. There have been no head-to-head trials and it seems that all scopes can be used with comparable results.
In 2002, Olympus developed the first EBUS bronchoscopes (models BF-UC160F-OL8 or BF-UC260F-OL; Olympus Medical Systems Corp., Tokyo, Japan), which were 6.7 mm wide and had a 2-mm working channel. A curved linear array ultrasonic transducer sits on the distal end (figure 1) and can be used either with direct contact to the mucosal surface or via an inflatable balloon, which can be attached at the tip. This allows a conventional endoscopic picture side-by-side with the ultrasonic view.
Figure 1. Tip of an EBUS scope. The US processor is visible (in red).
Since 2009, Pentax has offered a quite similar EBUS scope (EBUSpro; Pentax, Tokyo, Japan). The CCD chip of the EBUSpro is integrated directly into the endoscope’s distal end, the scope is 6.3 mm wide and has a 2 mm instrument channel, and a slightly degreed side-viewing optic. Later, Fujifilm launched their EBUS scope with a distal end outer diameter of 6.7 mm (EB-530US; Fujifilm, Tokyo, Japan).
Depending on the scope used, US scanning is performed with frequencies of 5–12.5 MHz and a tissue penetration of 20–50 mm. To see the parabronchial structure via US, the scope must be connected to an US processor. The newest series of these processors allows additional US features during the procedure, such as elastography or harmonic imaging.
Regarding evidence, it must be stated that most of the literature published up to this point was obtained using the Olympus scope.
Procedure and needles
TBNA is performed by direct contact of the transducer with the wall of the trachea or bronchus [10]. When a lesion is visible with US (figure 2), a needle can be moved through the working channel into the lymph nodes under real-time US visualisation (figure 3). While the procedure is taking place, the time colour Doppler can be used to identify the surrounding vascular structures (figure 4). After insertion of the needle into the lesion, suction is applied with a syringe, and the needle is moved back and forth though the node. Normally, physicians stab the target 10–15 times. Prior to retracting the needle into the needle sheath, suction must be removed to minimise sample loss into the syringe.
Figure 2. A lymph node in position 10 r (light grey). Directly below, the primary is visible (darker grey).
Figure 3. Needle inserted in a lymph node.
Figure 4. The vessel (in orange) within the lymph node in the Doppler mode.
After collection, there are several possible ways to handle the material. The most frequently used method is air flushing the material on a glass slide or sending the material for cell-block analysis [11]. The local handling should be discussed with the pathologist in order to provide the tissue-reader with the material in his preferred way.
Several needle sizes are available ranging 19–25 gauge. Most of the published data have used the 21 or 22 gauge needle [12–16]. Larger and smaller needles have benefits for different special indications, but a real evidence regarding this question does not, as yet, exist.
Results
The success story of EBUS-TBNA began in 2003 in a Thorax article by KRASNIK et al. [12]. The article provided the first description of the principle of EBUS-TBNA. 3 years later, in the same journal, HERTH et al. [13] presented their study of 502 patients showing that EBUS-TBNA resulted in a 93% diagnostic yield, a sensitivity of 94%, a specificity of 100% and an accuracy of 94%, with a positive predictive value (PPV) of 100% and a negative predictive value (NPV) of 11%. A further interesting outcome of the study was that no significant difference between US diagnosis under local and general anaesthesia was identified. Since then, several studies have been published that show comparable data relating to the staging and diagnosis of lung cancer with the help of EBUS-TBNA [13].
In 2006, a guide to performing EBUS-TBNA was published, which offered a detailed description of local lymph node positions and orientation within the mediastinum [14]. It was the first comprehensive reference tool for the growing number of EBUS-TBNA users. This detailed description was updated in 2009 by TOURNOY et al. [15] using the TNM (tumour, node, metastasis) map.
In 2010, ANNEMA et al. [16] presented the results of ASTER (Assessment of Surgical Staging versus Endosonographic Ultrasound in Lung Cancer: a Randomized Clinical Trial) in JAMA. This large randomised controlled multicentre trial conducted between February 2007 and April 2009 enrolled 241 patients with resectable, suspected nonsmall cell lung cancer. Where the need for mediastinal staging was indicated via CT or PET, patients were randomised to either surgical staging or endosonography followed by surgical staging in cases were no nodal metastases were found at endosonography. The group were able to show that nodal metastases were found in 41 patients via surgical staging and 56 patients via endosonography, with sensitivities of 79% versus 85%. Thoracotomy was unnecessary in 21 patients in the mediastinoscopy group versus nine in the endosonography group. The complication rate was similar in both groups.
All of this data changed the gold standard for staging the mediastinum. In the lung cancer guidelines of the American College of Chest Physicians (ACCP), EBUS-TBNA was for the first time clearly recommended as the initial method for examining the mediastinum [5], superseding classical mediastinoscopy. Several other international societies then followed this recommendation [17, 18].
Beyond its use for staging and diagnosing lung cancer, Dr Yasufuku’s group are dedicated to evaluating the benefits of EBUS-TBNA samples for immunohistochemical analysis and have reported encouraging results with cell cycle-related proteins in chemotherapy patients [19]. A year earlier, the same group presented a study that showed that epidermal growth factor receptor mutation can be easily detected in metastatic lymph node samples using EBUS-TBNA [20]. The group had also previously reported chemosensitivity-related aberrant methylation profiling in samples obtained using EBUS-TBNA [21]. In short, the group proved that samples obtained using EBUS-TBNA allowed genetic evaluation of tumour cells from lymph nodes, e.g. in EML4-ALK testing [22]. Recently, the first studies have been published that show that the material obtained with EBUS-TBNA can also be used to confirm the PDL-1 situation of the patients [23, 24].
In 2007, WONG et al. [25] presented the first study that evaluated EBUS-TBNA use beyond mediastinal staging. This large-scale case study included 65 patients. EBUS-TBNA was shown to be a safe method that allowed a high diagnostic yield for sarcoidosis. EBUS-TBNA was diagnostic in 85–91.8% of patients with a final diagnosis of sarcoidosis.
ANNEMA et al. [26] also conducted a prospective randomised trial that compared the use of EBUS-TBNA with transbronchial biopsies in patients with suspected sarcoidosis. They found a diagnostic yield for the detection of granulomas of 80% with EBUS-TBNA and only 53% with transbronchial biopsies. As a result, EBUS-TBNA is now the recommended technique in these patients and should be used instead of mediastinoscopy.
The use of EBUS-TBNA in suspected tuberculosis (TB) is less well established. However, one multicentre case series describing 156 patients with a diagnosis of TB-associated lymphadenopathy reported a sensitivity and accuracy of 94% [27].
Several studies have evaluated the usefulness of EBUS-TBNA in the diagnosis of lymphoma [28, 29]. In a retrospective analysis that included 25 patients with mediastinal adenopathies and suspected lymphomas [28], EBUS-TBNA enabled a sample of lymph tissue to be obtained in 96% of the patients (24 out of 25). The sensitivity, specificity, PPV and NPV observed were 90%, 100%, 100% and 92.6%, respectively. Larger specimens are often required to identify the exact subtype and grading when diagnosing lymphoma. A study by STEINFORT et al. [29] evaluated EBUS-TBNA in lymphoma and found a diagnostic sensitivity of 76%; however, some of these patients required further biopsy reducing the sensitivity to 57%. EBUS-TBNA should not be considered the first diagnostic approach in suspected lymphoma but can identify a proportion of patients with the disease. This may change in the future with use of better and, in particular, larger needles.
If additional tissue specimens are required for histological analysis, it is possible to insert 1.15-mm mini-forceps through the EBUS scope and past the airway wall via a needle puncture. This method allows the pulmonologist to obtain real-time guided forceps biopsies of mediastinal lymph nodes [30]. EBUS-TBNA has also been used to successfully obtain biopsy specimens in centrally located paratracheal and peribronchial tumours with a diagnostic sensitivity of 82–94% [31, 32]. Furthermore, real-time guided EBUS-TBNA has been used therapeutically to drain mediastinal, as well as bronchogenic cysts and consequently relieve central airway obstruction [33, 34].
Complications
Reading the available literature and the reviews, it is undisputable that EBUS-TBNA is a very safe procedure and that complications are very rare. In a meta-analysis by GU et al. [35], an overall complication rate of 0.15% was reported. Infectious complications were found when puncturing cysts. As with regular bronchoscopy, other complications (including infection, transient fever and haemoptysis) can also occur [36].
Damage to the bronchoscope caused by untrained use of the needle and other handling mistakes can be a very expensive complication. It is recommended that the whole team involved with handling this device receives regular training from the bronchoscope provider. In a Japanese survey, damage to the working channel was found to be the most frequent complication [37, 38].
Learning curve
EBUS-TBNA takes time to learn under experienced guidance in order to fully understand the ultrasonographic pictures not seen in regular bronchoscopy. There is currently no consensus in the national and international guidelines regarding how best to train physicians in this technique. Learning curves have been examined and one Australian study reported that the diagnostic yield improved significantly by 20 procedures but that 50 procedures was required to reach the peak diagnostic yield [38, 39]. Another trial recommended performing 13 procedures under supervision in order to be able to perform the procedure safely and obtain an adequate yield [40]. Different learning models have been proposed and KONGE and co-workers [39, 40] published data on the use of virtual reality bronchoscopy simulators to assess performance in order to set a standard for training prior to “real-life” bronchoscopy. Skills learned on simulators appear to be transferrable to a clinical setting. Larger trials are necessary here but more training opportunities need to be implemented since the technique is becoming ubiquitous.
A structured, modular training course is provided by the European Respiratory Society (ERS). In the first module, following web-based self-directed assessment, participants take part in a structured, physical course where EBUS equipment is introduced and demonstrations of live procedures are performed. The second module involves intensive simulation-based training and active clinical observation. In the third module, participants perform procedures in a supervised environment. After successfully completing the training programme, the participant can then obtain an ERS training certificate [41].
Combining EBUS and EUS
According to the guidelines, minimally invasive methods such as EBUS-TBNA and EUS-FNA are preferable to more invasive procedures such as mediastinoscopy and video-assisted thoracoscopic surgery [17]. EUS-FNA and EBUS-TBNA have been shown to prevent mediastinoscopies to a large extent [16, 17]. EBUS-TBNA and EUS-FNA have a complementary reach in analysing mediastinal nodes whereby EBUS has access to the paratracheal, subcarinal and hilar regions and EUS has access to the lower mediastinum and the aortopulmonary window [16].
EUS and EBUS provide access to different areas of the mediastinum. By combining these techniques, most lymph node stations, as well as the left adrenal gland, can be reached (with the exception of stations 5 and 6). In a meta-analysis, the accuracy of EUS-FNA and EBUS-TBNA used in combination for the diagnosis of mediastinal cancer was 95% [42].
The use of EBUS and EUS-B have also been examined. HWANGBO et al. [43] and HERTH et al. [44] showed similar results compared to those seen with the use of EBUS and EUS. Therefore, combining oesophageal and bronchoscopic endoscopic staging with a single linear US bronchoscope is effective when performed by an experienced endoscopist.
EBUS via the oesophagus
An almost complete evaluation of the mediastinal and hilar nodes can be achieved in a single staging procedure by inserting the EBUS scope into the oesophagus (EUS-B) following an endobronchial evaluation [45]. Following an EBUS procedure via the airways, the EBUS scope is retracted from the trachea and positioned into the oesophagus while the patient remains in a supine position. Para-oesophageal nodes should be investigated systematically, particularly those located in the lower mediastinum, and the subcarinal and paratracheal space. The advantages and results of this procedure are presented in a systematic review by J.T. Annema’s group [42].
Future developments
The next generation of EBUS scopes are now entering the market. The newest model of Olympus scope (prototype TCP-EBUS, BF-Y0046; Olympus Medical Systems Corp.) has been tested and will be launched in 2018. The scope is slimmer, has a thinner tip (5.9 mm), and has a larger bending angle (170 degrees upward) than the current convex EBUS probe (6.9-mm tip, 120 degrees upwards, and 35 degrees downwards). Accessibility, operability and TBNA capability of the slim EBUS scope were examined by K. Yasufukus’ group [46, 47]. They were able to show that both the endoscopic visibility range (14.7 mm) and the maximum reach (16.0 mm) were greater than in the current EBUS scope. The TCP-EBUS was able to visualise 1–3 distal bifurcations further. Adequate sampling from the lobar and segmental lymph nodes was possible using the aspiration needle.
New needle options have also been launched. With the current 21- and 22-gauge needles, the procedure can be limited by the degree of flexibility in the needle and the size of the lumen in tissue acquisition. A new flexible 19-guage EBUS-TBNA (Flex 19G; Olympus Respiratory America, Redmond, WA, USA) needle has therefore been developed, and has been evaluated at three centres [48]. In 47 patients with enlarged hilar and/or mediastinal lymphadenopathy, the diagnostic yield of the Flex 19G needle was 89% (42 out of 47 patients). The diagnostic yield with the 19G needle was: malignancy in 24 (89%) out of 27 patients, sarcoidosis in 13 (93%) out of 14 patients and reactive lymph node hyperplasia in five (83%) out of six patients. There were no complications. All 13 patients diagnosed with adenocarcinoma using the 19G needle had sufficient tissue for genetic testing. It seems the 19G needle can be feasibly and safely used, with promising diagnostic yield, while providing a greater degree of flexion with the EBUS scope. As always, further clinical evaluation is required.
From a technical perspective, the additional features are also under investigation. Elastography in particular seems to be helpful (figure 5). Elastography is an ultrasonic method in which the stiffness of tissues can be seen in real-time as a colour map. This technology has been used in the EUS field for several years; the European Federation of Societies for Ultrasound in Medicine and Biology (EFSUMB) has published guidelines and recommendations describing the technology [49, 50]. Several groups have also used it in EBUS [51–53]. Up to now, the largest published trial was performed by SUN et al. [54]. The group used the technology in 68 lymph nodes (33 benign and 35 malignant) from 56 patients, in order to examine the ability of EBUS elastography to differentiate between benign and malignant lymph nodes. The EBUS characteristics on B-mode, vascular patterns and elastography were recorded in all patients. The elastographic patterns and the mean grey value inside the region of interest were also analysed. All methods were compared with the definitive diagnosis of the lymph nodes. SUN et al. [54] were able to show that elastography is able to differentiate between benign and malignant lymph nodes with high sensitivity, specificity, PPV and NPV, and accuracy. It seems that EBUS elastography is potentially capable of further differentiating between benign and malignant lymph nodes. Elastography therefore seems to support nonmalignant needle puncture results and help the bronchoscopist believe in those results. Again, however, additional evidence is required here.
