The Authors: Kazuhiro Yasufuku, MD, PhD, is a Thoracic Surgeon with specific expertise in minimally invasive thoracic surgery and diagnostic procedures. He is Assistant Professor of Surgery at the University of Toronto, Toronto General Hospital. He is also the Director of the Interventional Thoracic Surgery Program at the University Health Network. Kasia Czarnecka, MD, completed her Internal Medicine residency at the University of Toronto where she is currently training in Adult Respirology. She is also training in the Interventional Thoracic Surgery Program with Dr. Yasufuku and the Thoracic Surgery Team. Her research interests include interventional bronchology and pulmonology.
Conflict of Interest statement: K.Y. has received educational and research grants from Olympus Medical Systems Corp.
SERIES EDITORS: JOHN E HEFFNER AND DAVID CL LAM
Kazuhiro Yasufuku, Division of Thoracic Surgery, Toronto General Hospital, University Health Network, 200 Elizabeth Street, 9N-957, Toronto, ON M5G2C4, Canada. Email: firstname.lastname@example.org
Interventional pulmonology (IP) allows comprehensive assessment of patients with benign and malignant airway, lung parenchymal and pleural disease. This relatively new branch of pulmonary medicine utilizes advanced diagnostic and therapeutic techniques to treat patients with pulmonary diseases. Endobronchial ultrasound revolutionized assessment of pulmonary nodules, mediastinal lymphadenopathy and lung cancer staging allowing minimally invasive, highly accurate assessment of lung parenchymal and mediastinal disease, with both macro- and microscopic tissue characterization including molecular signature analysis. High-spatial resolution, new endobronchial imaging techniques including autofluorescence bronchoscopy, narrow-band imaging, optical coherence tomography and confocal microscopy enable detailed evaluation of airways with increasing role in detection and treatment of malignancies arising in central airways. Precision in peripheral lesion localization has been increased through innovative navigational techniques including navigational bronchoscopy and electromagnetic navigation. Pleural diseases can be assessed with the use of non-invasive pleural ultrasonography, with high sensitivity and specificity for malignant disease detection. Medical pleuroscopy is a minimally invasive technique improving diagnostic safety and precision of pleural disease and pleural effusion assessment. In this review, we discuss the newest advances in diagnostic modalities utilized in IP, indications for their use, their diagnostic accuracy, efficacy, safety and challenges in application of these technologies in assessment of thoracic diseases.
Interventional pulmonology (IP) is a dynamically evolving field within pulmonary medicine focusing on comprehensive, minimally invasive approach to the diagnosis and management of malignant and benign disease of the thorax, including lung cancer, mediastinal and hilar lymphadenopathy, lung nodule, central airway obstruction, pleural and obstructive airway diseases. Lung cancer has become the most common cancer worldwide since 1985, both in terms of mortality and incidence. Its progressing epidemic, with many patients presenting with bulky endobronchial tumours and malignant airway obstruction requiring endobronchial therapy, was an initial impetus for development and later evolution of IP.1,2 Now, IP includes a variety of not only therapeutic interventions but also advanced diagnostic techniques aiding in assessment of intrathoracic pathology, specifically revolutionizing diagnostic assessment of lung cancer. Early diagnosis of lung cancer is imperative, the current poor 5-year survival of only 15.6% (US data)2 is mostly due to most patients coming to medical attention at an advanced stage and a better survival has been observed in less advanced disease. Traditional modalities available to assess pulmonary nodules, including flexible bronchoscopy even with assistance of bronchoalveolar lavage, cytology brushing, transbronchial needle aspiration (TBNA) and transbronchial biopsy, offer unacceptably low sensitivity, especially in setting of smaller, peripheral lesions (ranging from 14% to 50%).3,4 For larger (>4 cm), malignant, peripheral nodules, percutaneous approach via transthoracic fine needle aspiration under the computed tomography (CT) guidance has high sensitivity (90%) and specificity (97%). However, the procedure carries high risk of pneumothorax (ranging from 17% to 33%) and risk of seeding the needle tract with tumour cells as well as fatal arterial air embolism. Yield of percutaneous fine needle aspiration drops significantly to 74.4% in lesions smaller than 1.5 cm.5 Similarly, mediastinoscopy—the gold standard procedure used for mediastinal staging in patients with confirmed or suspected lung cancer—despite its high yield of >90%, is an invasive procedure carrying low but significant risks of complications including injury of major vascular, peripheral nerve, tracheobronchial tree structures and oesophagus (<0.5%), pneumothorax and infection (up to 2.5%).6
Revolution in imaging technology has challenged the major roadblocks (diagnostic yield and complication rate) limiting the traditional thoracic pathology diagnostic techniques and has allowed the interventional pulmonologist to assume an increasingly more important role in the diagnostic assessment of thoracic diseases. The introduction of endobronchial ultrasound (EBUS) allowed advanced assessment of mediastinal pathology including staging of lung cancer and offered a convenient, precise and safe means of assessment of peripheral lung nodules. Electromagnetic navigational bronchoscopy (ENB) and virtual bronchoscopic navigation (VBN) allow for quick and precise localization of peripheral pulmonary lesions. Advanced airway assessment techniques, such as autofluorescence bronchoscopy, narrow-band imaging (NBI) and optical coherence tomography, have opened an avenue for early endobronchial malignancy detection and surveillance. Medical pleuroscopy (MP) increasingly performed by pulmonologists is a safe tool in diagnosis and management of pleural effusions and abnormalities. This review will provide an overview of the current status of diagnostic modalities utilized in IP.
Endobronchial ultrasound (EBUS)
The development of the radial probe EBUS in the early 1990s has taken the diagnostic flexible bronchoscopy to the next level by allowing the bronchoscopist to see beyond the airway, extending visualization into peribronchial structures and lung nodules, making it a useful modality in the evaluation of airway abnormalities, diagnosis of lung nodules and mediastinal lymphadenopathy, as well as in staging of non-small cell lung cancer (NSCLC).7,8
There are two types of EBUS: the radial probe EBUS and the convex probe EBUS (CP-EBUS). The radial probe EBUS is a flexible rotating mechanical probe introduced through the channel of a flexible bronchoscope, allowing direct visualization of bronchial wall and surrounding structures. It allows assessment of the tumour invasion depth and can therefore offer guidance with respect to therapeutic intervention choices. It also enhances yield of TBNA for lymph node (LN) cancer staging.9,10 The CP-EBUS is a flexible bronchoscope integrated with a convex-type transducer on the tip, which allows real-time EBUS-guided TBNA (EBUS-TBNA) of mediastinal LN and hilar LN.7
Radial probe EBUS
The radial probe EBUS is a 20-MHz rotating mechanical probe capable of producing high-resolution (up to less than 1 mm) 360° images. Twenty- and 30-MHz probes are also available. However, the 20-MHz probe is the most commonly used and the only probe commercially available.7 It has penetration depth of 5 cm and is inserted through the working channel of a flexible bronchoscope. The central-type radial probe EBUS (UM-BS20-26R) is fitted with a 2.6-mm balloon sheath carrying a water-filled balloon at the tip. Using this probe, detailed imaging of the central airway bronchial wall and peribronchial structures, such as LN and pulmonary vasculatures, can be obtained.
A recently developed peripheral-type radial probe EBUS has drawn more attention from IP and is being used for the assessment of peripheral nodules (Fig. 1). Different size probes are used with or without a guide sheath (GS) for the peripheral EBUS. A large channel flexible bronchoscope with at least a 2.8-mm working channel is needed for the use of the central-type radial probe EBUS and the larger peripheral-type EBUS.
Diagnostically, after its initial description in 1992, the radial probe EBUS was first used in evaluation of central airways.8 The cartilaginous portion of the bronchial wall has been described as a five- to seven-layer structure.11,12 Given the penetration depth of the radial probe EBUS (5 cm), its use allows assessment of depth of tumour infiltration into the airways. After assessing the laminar structure of the bronchial wall in 45 normal specimens, Kurimoto et al. assessed the depth of tumour invasion in 24 confirmed cancer subjects using radial probe EBUS, documenting a 95.8% (23/24 subjects) correlation with the histopathological findings.11 A 93% diagnostic accuracy in assessment of bronchial wall tumour invasion was reported by Tanaka et al. based on comparison to preoperative EBUS assessments in patients with intrathoracic malignancy and postoperative pathological examination in 15 out of 35 patients who underwent surgical management of their disease.12
Photodynamic therapy is a minimally invasive management option for patients with early central airway lung cancer.13 In one study, 18 patients with biopsy-confirmed squamous cell carcinoma of the lung were evaluated with EBUS. Tumour extent was assessed: nine patients had extracartilagenous lesions and were considered for other therapeutic modalities; the other nine patients had intracartilagenous lesions. These nine patients were treated with photodynamic therapy and demonstrated 100% remission with no recurrence at the mean 32-month follow-up.14 EBUS also improved assessment specificity (50–90%) of autofluorescence detected but invisible by white-light bronchoscopy lung cancers.13
Clinical practice guidelines recommend invasive mediastinal tissue sampling for lung cancer staging given the rather low sensitivity of non-invasive modalities (CT, sensitivity of 60.8%, positron emission tomography scan 72.5%).15 The radial probe EBUS has been shown to be helpful in mediastinal node sampling in workup of lung cancer, in particular, to guide TBNA. TBNA is a minimally invasive method of sampling mediastinal LN. However, given its blind nature, it often lacks precision even when combined with CT of the thorax. The technique has also been demonstrated to be operator dependent. Other predictors of positive biopsy outcome include the following: lesion size, location (with the right paratracheal LN and subcarinal LN being more often positive), presence of small cell lung cancer and use of histology needle. One study designed to assess the diagnostic utility of TBNA in conjunction with CT of the thorax in 360 patients with confirmed primary lung cancer showed that TBNA provided exclusive diagnosis in 18% of the patients (65/360) and precluded surgery in 29% of the patients (104/360). However, its diagnostic yield, ranging from as low as 53% to 70% (if all 607 biopsied lesions were considered) even when combined with CT of the thorax, and other limitations make the TBNA a widely underused diagnostic modality.16 In contrast, in a study of 242 patients with CT-proven mediastinal lymphadenopathy, radial EBUS-TBNA gave a diagnostic yield of 72%. Surprisingly, this was independent of node size and location.9 In randomized trial of 200 patients with mediastinal LN comparing the two techniques, the overall diagnostic yield of EBUS-TBNA was 80% and was significantly higher than the overall yield of the conventional TBNA (71%) performed in this study.10
The peripheral-type radial probe EBUS is utilized to improve yield in transbronchial biopsies of parenchymal lung lesions. Flexible bronchoscopy has a variable success in sampling peripheral lung nodules. Sensitivity of this procedure in detecting malignancy in solitary pulmonary nodule depends on the site of the nodule, its proximity to the tracheobronchial tree and prevalence of cancer in the studied population.17 Yield of flexible bronchoscopy in true solitary pulmonary nodules is quite low. According to the American College of Chest Physicians lung cancer guidelines, yield in lesions of less than 2 cm in size is only 34% (based on 383 patients evaluated in the studies reviewed). For lesions of more than 2 cm, the sensitivity was 63% (984 patients).18,19 Fluoroscopic guidance improves diagnostic yield but nodules of <2 cm often cannot be visualized. For these nodules, diagnostic yield was reported in a meta-analysis by Schreiber and McCorory19 to be only 33% (5–76%). CT-guided transthoracic needle aspiration offers a 74–96% diagnostic yield but requires radiation and is often complicated by a pneumothorax (rates of 15–44%).20 Radial probe EBUS can improve diagnostic yield in sampling of the solitary pulmonary nodule, giving a yield of 80%, similar to that of fluoroscopy-guided biopsies (40/50 patients investigated in one study).21 For smaller lesions, less than 2 and 3 cm, EBUS diagnostic yield is lower than 71% (47–95% at 95% confidence interval (CI), n = 25) and 75% (60–90% at 95% CI, n = 47), respectively, but significantly better than with transbronchial biopsy, 23% (3–43% at 95% CI, n = 31) and 31% (16.3–45.3% at 95% CI), respectively.22 For lesions smaller than 2 cm, the reported yield is lower at 46% (46/100 consecutive patients evaluated in one study).23 Even without fluoroscopic guidance, radial probe EBUS proves to be a successful, safe and effective diagnostic modality in the evaluation of parenchymal lung nodules, especially when combined with electromagnetic navigation (EMN) with a diagnostic yield of 88% (35/40 patients in one study).24
More recently, the introduction of the GS has made the use of the radial probe EBUS-guided transbronchial biopsy easier for physicians.25–27 The ultra-miniature 20-MHz radial probe (UM-S20-17S, Olympus Medical Systems, Tokyo, Japan) is placed into a radio opaque GS and manipulated through the working channel of a regular size flexible bronchoscope (with a 2.0-mm working channel, the probe external diameter is 1.4 mm). The probe is advanced into the site of the lesion where it can locate the lesion. Subsequently, the probe is removed from the airway leaving the GS behind, allowing precise sampling of the lesion through the bronchial brushings and biopsy for pathological and cytological characterization.7 The use of GS has increased the diagnostic yield of transbronchial biopsy, allowing for successful sampling of lesions of less than 10 mm in diameter and fluoroscopically invisible lesions.25–27
Radial probe EBUS has also been useful in the diagnostic evaluation of non-malignant pulmonary disease. Soja et al. demonstrated equivalent results between high-resolution CT and radial probe EBUS measurement of the bronchial wall thickness in patients with asthma. High positive correlation of bronchial wall thickness, as assessed by radial EBUS, with asthma severity parameters, like forced expiratory volume in 1 s, suggests a possible future use of this modality in assessment of asthma severity. This can be especially useful in the non-perceiver patients or severe asthmatics with severe airflow obstruction on spirometry and no bronchodilator reversibility.28 Radial probe EBUS has also been used successfully in placement of fiducial markers in preparation for stereotactic radiosurgery in patients with central and peripheral lung cancers.29
The CP-EBUS is a flexible bronchoscope integrated with a convex transducer on the tip, which scans parallel to the insertion direction of the bronchoscope. The outer diameter of the insertion tube of the flexible bronchoscope is 6.2 mm. The viewing angle is 90° and the direction of view is 35° oblique. Visualization can be achieved either via direct EBUS contact with the bronchial wall or via an inflated water-filled balloon at the tip of the probe. The latter method allows for a superior image quality. New CP-EBUS (BF-UC180F, Olympus Medical Systems) can now be connected to ultrasound scanner (EU-ME1, Olympus Medical Systems), a universal endoscopic ultrasound scanner for excellent image quality. Power and colour Doppler modes are available for precise characterization of examined structures. Twenty-one- and 22-G needles with an internal stylet for tip clearance once the needle has penetrated through the bronchial wall are used for tissue aspiration during a real-time EBUS-TBNA (Fig. 2).7 Once the lesion of interest is visualized, Doppler is used to confirm the localization of the adjacent structures and differentiate the lesion from surrounding vessels. On-site cytopathological assessment is possible in some centres. This allows for prompt assessment of sample adequacy and in, many cases, diagnosis. Papanicolaou staining and light microscopy are performed at the same time. Samples can also be processed using immunohistochemistry for more specific assessment.7
CP-EBUS can conveniently access highest mediastinal (#1), upper paratracheal (#2R, #2L), lower paratracheal (#4R, #4L), subcarinal (#7), hilar (#10), interlobar (#11) and lobar (#12) nodal stations. CP-EBUS cannot access prevascular and retrotracheal (#3), sub-aortic (#5), para-aortic (#6), paraoesophageal (#8) and pulmonary ligament (#9) nodes. Endoscopic ultrasound can be used complimentarily with EBUS for station 3, 8 and 9 access and can improve diagnostic yield from station 4L, sampling of which can be challenging with CP- EBUS because of the node angle.7
After the first introduction of the CP-EBUS in 2004,30,31 EBUS-TBNA has significantly changed the approach to the mediastinum and the hilum. Over the years, improved CP-EBUS has become an important modality in evaluation of the mediastinal pathology including mediastinal masses, lymphadenopathy and lung cancer staging.
CP-EBUS for LN staging of lung cancer
Appropriate mediastinal assessment is a crucial component of primary lung cancer staging given the importance of mediastinal involvement on prognostication and subsequent therapeutic interventions for both surgical and non-surgical treatment. There have been multiple studies demonstrating high sensitivity, specificity and diagnostic accuracy of the CP-EBUS in diagnosis of mediastinal lymphadenopathy in lung cancer patients. The first study to prospectively evaluate mediastinal LN staging in lung cancer patients demonstrated a sensitivity of 94.6%, specificity of 100%, positive predictive value of 100%, negative predictive value of 89.5% and diagnostic accuracy of 96.3%. Out of the 108 studied patients, in 20 suspected cancer subjects, mediastinal nodal sampling was also used for cancer diagnosis. Consequently, EBUS-TBNA prevented 29 mediastinoscopies, eight thoracotomies, four thoracoscopies and nine CT-guided percutaneous nodal biopsies.32 The largest multicentre, prospective study evaluated 502 lung cancer patients with documented mediastinal and hilar lymphadenopathy. The sensitivity of the EBUS-TBNA for malignancy detection was 94% after 572 nodal biopsies and positive results in 535 nodes. However, given the high malignancy prevalence in the studied population (98.2%), the negative predictive value was only 11%.33
Non-invasive staging with CT and positron emission tomography is not accurate enough for patients with discrete mediastinal lymphadenopathy.15 Several studies have compared non-invasive mediastinal staging with CT and/or with positron emission tomography with EBUS-TBNA in patients with and without mediastinal lymphadenopathy, demonstrating the superiority of EBUS-TNBA in terms of specificity, sensitivity and accuracy.34–36 Diagnostic accuracies reported by Yasufuku et al. were the following: CT 60.2%, positron emission tomography 72.5% and EBUS-TBNA 98.0% in 102 surgical candidates, with either biopsy-proven or suspected primary lung malignancy, enrolled in the study.34 This finding was confirmed by Herth et al. who showed high EBUS-TBNA sensitivity, specificity and negative predictive value of mediastinal nodal staging in patients with primary lung cancer who had previously documented negative CT and positron emission tomography investigations.36
Meta-analysis and systematic reviews have demonstrated equivalent EBUS-TBNA sensitivity as compared with that of mediastinoscopy for NSCLC mediastinal staging with sensitivity ranging between 85% and 93%, specificity of 100%, but quite variable negative predictive value (11–97.4%).37–39
Comparative studies evaluating EBUS-TBNA versus mediastinoscopy showed the overall excellent diagnostic accuracy of EBUS-TBNA. Recent randomized controlled, cross-over study comparing diagnostic yield between EBUS-TBNA and mediastinoscopy showed a diagnostic accuracy of 91% versus 78% (biopsied LN number yield 109/120 vs 94/120, P = 0.007), sensitivity of 87% versus 68% and negative predictive value of 78% versus 59% in favour of EBUS-TBNA in a per-node analysis in patients with suspected lung primary malignancy. In a per patient analysis, EBUS-TBNA diagnostic yield remained high at 89% but equivalent to that of mediastinoscopy (79%) (n = 66).40 In another study, Annema et al. compared the diagnostic accuracy of surgical staging alone versus endosonography (EBUS-TBNA combined with endoscopic ultrasound fine needle aspiration) in patients suspected of having NSCLC and showed that the bimodality approach had a similar sensitivity to mediastinoscopy (85% (56/66 patients; 95% CI 74–95%) sensitivity vs 79% (41/52; 95% CI 66–88%) (P = 0.047), respectively). Triple diagnostic intervention of EBUS-TBNA with endoscopic ultrasound fine needle aspiration and followed by mediastinoscopy showed an even greater sensitivity of 94% (62/66; 95% CI 85–98%).41 Recently, the first head-to-head comparison of EBUS-TBNA and mediastinoscopy performance for mediastinal staging of patients with potentially resectable lung cancer was published. Both procedures were performed sequentially by different surgeons on 153 patients suspected of having NSCLC. The EBUS-TBNA and mediastinoscopy sensitivity, negative predictive value and diagnostic accuracy for meditational LN staging were 81%, 91%, 93% and 79%, 90%, 93%, respectively. Specificity and positive predictive value for both methods was 100%.42 Mediastinoscopy has had a long-standing role for the definitive pathological exclusion of N2 or N3 disease. This study also established that EBUS-TBNA can accurately distinguish N0/N1 disease from N2 and N3 disease (an important distinction impacting in the nature of therapeutic interventions in lung cancer patients) to date provided by mediastinoscopy alone. These studies demonstrate that in certain patient populations where lung cancer prevalence is low (low lung cancer pretest probability), EBUS-TBNA could potentially be used instead of mediastinoscopy for mediastinal nodal staging without compromising diagnostic yield. However, in patients with high-pretest probability of lung cancer, negative EBUS-TBNA may not be sufficient to safely exclude N2/N3 disease (lower negative predictive value as compared with mediastinoscopy in such setting), obviating the need for surgical mediastinal disease staging.
Overall, the advantages of EBUS-TBNA over mediastinoscopy include the following: (i) minimally invasive nature with only minor (cough, blood at the puncture site and agitation) or no complications43 reported. In contrast, cervical mediastinoscopy carries significant mortality ranging between 0.08% and 0.2% and morbidity rate of 2–2.5%, including injury of major mediastinal structures (trachea, oesophagus, major vessels and peripheral nerve structures), which can lead to life-threatening bleeding, left recurrent laryngeal nerve injury and infection;15 (ii) EBUS, if used with endoscopic ultrasound, offers a wider access to mediastinal LN (stations 8 and 9) and alone to hilar LN (station 10), lobar LN (station 11) and interlobar LN (station 12), all inaccessible to mediastinoscopy; (iii) EBUS-TBNA can be performed as an outpatient procedure requiring only local anaesthesia and conscious sedation without the need for general anaesthesia (necessary for mediastinoscopy) and subsequent hospital admission, making it a quicker and more-cost effective procedure;44 (iv) EBUS has been shown to be useful in evaluation of postoperative mediastinal lymphadenopathy45 and re-staging of lung cancer46 following medical therapy to monitor response to treatment and to guide further therapy. In contrast, mediastinoscopy leaves significant mediastinal scarring that makes it technically very challenging to repeat it for such purposes; and (v) finally, EBUS-TBNA is a procedure shared by both pulmonologists and thoracic surgeons which, at the centres with access to and expertise in both procedures, could make it a more accessible one of the two procedures.
EBUS has some disadvantages, however. (i) It requires trained personnel including the supporting staff. It is a highly operator-dependent procedure with diagnostic yield and accuracy strongly dependent on the experience of the operator, especially because no general guidelines exist on EBUS training and no standardization of mediastinal nodal staging procedure has been suggested by American College of Chest Physicians. (ii) As mentioned above, due to its lower negative predictive value than mediastinoscopy, negative EBUS does not rule out micrometastatic mediastinal disease in high-risk patients, suggesting that it may not totally replace mediastinoscopy for mediastinal staging. (iii) Aside from issues with accessibility to a high-quality training of professionals interested in performing EBUS procedures, implementation of EBUS by clinical centres can be hindered by initial cost relating to equipment purchase, personnel training and, subsequently, relatively costly equipment maintenance and repair fees that need to be budgeted into the service operation fees. Based on Canadian data collected by Hergott et al., costs of maintenance and equipment repair in one Interventional Pulmonology Program in Canada have been estimated to be on average US$100.80 per procedure with equipment damage shown to be unavoidable long term even in the hands of experienced bronchologist, raising wear and tear issue as a major culprit behind the bronchoscope damage.47 Cost may be higher in centres with less expertise in equipment handling.
CP-EBUS for diagnosis of other diseases of the mediastinum
Mediastinal lymphadenopathy predicts a poor outcome in patients with metastatic lung tumours. EBUS has been used to evaluate mediastinal and hilar lymphadenopathy in metastatic malignant lung disease. In one study, EBUS showed sensitivity, specificity and diagnostic accuracy rates of 92%, 100% and 95.3%, respectively, in 106 patients with metastatic lung tumours who underwent EBUS prior to planned metastectomy.48 This demonstrates that EBUS can be a useful, minimally invasive and accurate modality in assessment of mediastinal and hilar disease in patients with metastatic lung tumours. EBUS has also been evaluated in diagnosis of primary LN disorders of the mediastinum including lymphomas, infections and non-infectious granulomatous disease.49 Owing to the small and structurally distorted nature of the EBUS-TBNA sample, it has been thought that EBUS-TBNA could not be used reliably in diagnosis of mediastinal/hilar lymphoma. One retrospective study reported 91% (24/25 patients under study) EBUS-TBNA diagnostic sensitivity for diagnosis of mediastinal lymphoma.50 Another study by Ko et al. showed excellent EBUS-TBNA diagnostic yield in lymphoma in patients with mediastinal lymphadenopathy.51 The study used EBUS-TBNA with on-site assessment and triage of samples for multiple ancillary techniques for the diagnosis and subclassification of lymphomas and non-neoplastic diseases. In 120 patients enrolled, adequate sample for appropriate analysis was obtained in 95 subjects. A combination of Papanicolaou stained direct smears, immunohistochemistry, immunophenotyping and fluorescence in situ hyberdization allowed for diagnosis and subclassification of three Hodgkin's lymphomas and seven non-Hodgkin lymphomas (one small lymphocytic, one small lymphocytic with scattered Reed-Sternberg cells, one marginal zone lymphoma and four large B cell lymphomas), proving that EBUS-TBNA is a useful diagnostic modality for mediastinal lymphoproliferative disorders providing adequate sample for cytomorphological assessment and other ancillary tests, permitting accurate diagnosis of lymphoproliferative disorders leading to management decision without the need for invasive core biopsy.
Diagnostic yield of modalities currently available for diagnosis of sarcoidosis (including bronchoalveolar lavage, transbronchial lung biopsy and mediastinoscopy) varies between 40% and 90%. Tissue diagnosis is crucial to rule out malignant disease, tuberculosis or histoplasma infection especially if steroid treatment is planned.52 Case-series and cohort studies have shown that EBUS-TBNA can also be used successfully in diagnosis of sarcoidosis, with high diagnostic yield ranging between 90% and 96% and sensitivities between 71% and up to 93%.53–57
EBUS-TBNA has also been shown to be useful in diagnosis of primary and metastatic lung tumours with a high diagnostic yield if the lesion is within the reach of the CP-EBUS.58 One retrospective chart review showed high sensitivity (96.4%), specificity (100%) and diagnostic accuracy (97.2%) of EBUS-TBNA in the diagnosis of small cell lung cancer, with significant impact on patient management and survival (n = 40)59 (77.8% 5-year survival in EBUS-TBNA assessed patients vs 33% and 21% same interval survival in historical control as reported for stage IA and IB small cell lung cancer, respectively).
Molecular analysis of LN sampled by CP-EBUS
EBUS-TBNA can sample metastatic mediastinal LN and hilar LN repeatedly without significant complications. Although the current CP-EBUS is limited to the use of 21- and 22-G needles, samples obtained by EBUS-TBNA have been used for molecular analysis of metastatic LN in lung cancer patients. An early report in patients with N2 or N3 lung cancer showed that DNA extracted from paraffin embedded samples is usable for the detection of epidermal growth factor receptor mutation analysis.60 Methylation analysis and extraction of RNA from EBUS-TBNA samples has been demonstrated and has been used for aberrant fusion gene detection (EML4-ALK fusion gene).61–63 The ability to perform biological analysis using non-surgical biopsy samples using EBUS-TBNA will become very important for the management of patients with lung cancer in the era of personalized treatment.
Advanced bronchoscopic diagnostic assessment of the airways
Squamous cell carcinomas, accounting for approximately 25–30% of all lung cancers, arise in central airways. Pathobiologically, progression from normal bronchial epithelium to squamous metaplasia followed by dysplasia, carcinoma in situ and finally invasive carcinoma has been well described.64 Studies have shown that patients with moderate to severe dysplasia progress to develop invasive carcinoma over the course of 3–4 years. Approximately 11% of patients with moderate dysplasia and from 19% to as high as 50% with severe dysplasia will develop invasive carcinoma.65–67 Therefore, prompt detection through screening of high-risk patients (heavy smokers especially) could potentially offer early diagnosis of early pre-invasive or early invasive lesions and allow for prompt therapeutic intervention and improved survival. However, conventional airway imaging modality, white-light bronchoscopy has been shown to be relatively insensitive in inspection of bronchial mucosa with only 30% sensitivity to detect early-stage carcinoma in the central airways.66
New bronchoscopic modalities, with higher-spatial resolution able to take advantage of intrinsic properties of healthy and abnormal tissues to change appearance when illuminated with different wavelengths of light, have been developed to serve the purpose of more advanced central airway imaging for the purpose of abnormal airway diagnosis. Currently available in clinical practice modalities include the following: autofluorescence bronchoscopy68–70 and NBI (Fig. 3).71–73 More precise airway inspection can be obtained with radial probe EBUS (described above) and optical coherence tomography.13
Electromagnetic navigational bronchoscopy (ENB)
Modalities currently available for diagnosis of peripheral lung nodules include percutaneous needle aspiration and thoracoscopic and open lung biopsy. Each of the approaches is associated with high risk of morbidity and mortality. Pneumothorax rate following percutaneous needle aspiration ranges between 20% and 32%. Mortality of 0.5–5.3% has been described in surgical lung biopsy cases.20,74 Flexible bronchoscopy with low complication rate is, however, of limited diagnostic utility given its inability to access and precisely localize peripheral lesions. ENB is a recently developed technology designed to assist in diagnostic evaluation of peripheral lung nodules in conjunction with flexible bronchoscopy.
ENB uses electromagnetic field to track a locatable guide in real time, correlating its position in the tracheobronchial tree to the patient's CT scan. A path to the lesion is planned this way. Locatable guide is advanced to the lesion location using a probe within the working channel of a flexible bronchoscope. Once the lesion location has been reached, the guide is removed leaving the GS in place. Biopsy tools can be passed through the GS. The inReach ENB system is the most widely used one.75
Utility of ENB in diagnosis of peripheral LN has been assessed in several studies (n = 285) with overall yield ranging between 59–77% and 54–75% in nodules up to 3 cm in size.24,75–77 Successful mediastinal LN biopsies have also been described using EMN.77 When combined with the radial probe EBUS, diagnostic yield is even higher (88%).24 No EMN complications have been reported. Pneumothorax, related to nodule biopsy, is the most often reported complication in the studies (1.2–6%).75–77 Limitations of the ENB include the need for a bronchus leading to the lesion. Also no absolute confirmation that the lesion has been reached can be obtained. The bronchoscopist only knows that the probe is within the area of the lesion based on the results of the CT scan that was performed on a different date and at different lung volumes. Therefore, prolonged time between the planning CT and the procedure may result in inaccurate identification of the lesion site. Respiratory movement can also preclude diagnostic accuracy especially for distal, lower lobe lesions. Cost of the equipment is another issue that may exceed that associated with other diagnostic methods. However, the learning curve in performing the ENB is steep, and in centres where technology and expertise are available, ENB offers a safe and useful adjunct to flexible bronchoscopy, increasing diagnostic yield for peripheral lung nodules and mediastinal lymphadenopathy.
Virtual bronchoscopic navigation (VBN)
The yield of flexible bronchoscopy for diagnosis of peripheral lung nodules is exceedingly poor (57% for all lesions and 34% for lesions less than 2 cm).78 The development of ultrathin bronchoscopes (external diameter of 2.8 mm) and thin-cut CT imaging enabled advancements in the three-dimensional visualization of the distal tracheobronchial tree through the development of VBN. Virtual bronchoscopy (VB) creates a three-dimensional display of the border between the bronchial lumen and the bronchial wall, as if observed from the bronchial lumen using a bronchoscope.79 VB can visualize airway beyond stenosis and extramural structures using volume rendering.80 For that reason, VB has been used for evaluation of endobronchial malignancy, tracheobronchial injury, airway foreign body, postoperative bronchial complications, TBNA, stent placement and education of bronchoscopists.81 The use of VB is limited to central bronchi as conventional CT image visualization of airways peripheral to segmental bronchi is inaccurate with less consistency with true anatomic findings. VBN uses real-time bronchoscopy to chart the path to the lesion correlating and aligning the VB images with the real images. The challenge with VBN is that the amount of detail displayed is dependent on the CT quality, which can be affected by the obtained volume data (often differing among different systems).82 Rotation of the bronchoscope at insertion may result in misalignment of the virtual and real image leading to misguidance towards a wrong bronchus. Also, inappropriately chosen thresholds for differentiation between airway and lumen may result in misguided path.81
VBN system (Bf-NAVI; KGT, Olympus Medical Systems) has been developed to overcome these challenges. It produces VB images and automatically aligns them with the real images for bronchoscopic navigation. VNBS corrects for bronchoscope rotation at the time of insertion through the ‘rotation function’.83 Since the arrival at the site of the lesion cannot be confirmed with the VNBS, the technology has been used in conjunction with X-ray fluoroscopy, CT guidance and EBUS.
In one study, VNBS was used in conjunction with fluoroscopy to navigate the path to 96 peripheral pulmonary lesions, 3 cm or less in size, in 94 patients. Diagnostic yield was directly related to lesion size with yield of 35% (7/20), 61.4% (35/57) and 94.7% (18/19) for lesions less than or equal to 10, 11–20 and 21–30 mm, respectively. Seven of eight ground glass nodules were successfully diagnosed. Average examination time was 24.1 min.84
VBN has also been used in conjunction with EBUS. Asahina et al. showed 80% (24/30) visualization and 63.3% (19/30) overall diagnostic yield of peripheral pulmonary lesions of 3 cm and less in diameter using VBN with EBUS-GS assistance.27 Similarly, Asano et al. showed 93.8% (30/32) visualization using EBUS-GS and 84.4% (27/32) peripheral lesion diagnostic yield using VBN and EBUS-GS and thin bronchoscope (4.0 mm outer diameter and 2.0 mm inner channel).85 Usefulness of the VBN system has recently been confirmed in a multicentred randomized controlled trial (virtual navigation in Japan).86 A total of 199 patients were studied. In the VBN-assisted (VBNA) group (n = 102), the bronchoscope was navigated into the site of the lesion using the VBN system. In the non-VBNA group (n = 97), CT-generated image prior to procedure was used to choose the path to the lesion. Arrival at the lesion site and biopsy took place under fluoroscopic guidance in both groups. Diagnostic yield was superior in the VBNA group (80.4% (82/102) vs 67% (65/97), P = 0.032). Thin bronchovideoscope was used in all patients. VBN-assistance also reduced the overall diagnostic procedure time (24.0 vs 26.2 min in the VBNA vs non-VBNA group). Complications were few and minor including small pneumothorax in the Non-VBNA group. Pneumothorax or other complications have not been found in any of the reported EBUS + VBN studies to date.
Overall, all-lesion diagnostic yield of VBN with EBUS-GS ranges from 63.3% to 84.4% and from 44% to 75.9% in lesions less than 2 cm in diameter. VBN increases diagnostic yield in small peripheral pulmonary nodules and decreases overall procedure time. It is usually combined with fluoroscopy, CT or EBUS to confirm the lesion location. At present, manual adjustment of VB images to real images is needed. However, a method to automatically adjust the two images has been developed87 and tested in a phantom study, showing that the image-based guidance using real-time registration of the three-dimensional multidetector CT scan image data and live bronchoscopic information may correct for respiratory motion interference and improve lesion localization precision, resulting in higher diagnostic yield in the peripheral lung nodules even in hands of inexperienced bronchoscopists.88 Randomized studies comparing diagnostic yield of different modalities in conjunction with VBN are needed to further assess VBN utility. However, given its relatively low as compared with EMN cost and no need for sophisticated equipment, its future widespread use is likely.
Medical pleuroscopy (MP)
Pleural disease represents 25% of referrals to pulmonologists,89 hence the growing interest of pulmonologists in acquiring skills in both new diagnostic and therapeutic pleural interventions. Techniques utilized by interventional pulmonologists in diagnosis of pleural disease involve pleural ultrasonography, MP and pleural biopsy.
Undiagnosed pleural effusion can be assessed utrasonographically. Pleural thickening of more than 10 mm, pleural nodularity and diaphragmatic pleural thickening >7 mm are strongly suggestive of malignancy with 73% sensitivity (26/33 patients with confirmed histology/long-term follow up and CT evidence malignant pleural effusion) and 100% specificity reported by Qureshi et al.90 Echogenic swirling pattern with the presence of free-floating echogenic particles within the pleural cavity is suggestive of exudative pleural effusion.91 However, histological assessment of pleural tissue or fluid is a gold standard for diagnosis of malignant pleural effusion of primary or metastatic pleural process. Ultrasound-guided thoracocentesis is becoming a more popular way of diagnostic and therapeutic drainage of pleural effusions. Ultrasound guidance also allows for a safe selection of site for port insertion during MP and tube thoracostomy placement.92 Its use is slowly becoming a standard of care for thoracoscopic interventions. Expertise in thoracic ultrasound is a crucial skill of an IP clinician.
Large volume thorocentesis has an approximately 62% diagnostic yield for malignancy. Second aspiration increases that chance slightly by up to 15% with a third sample being non-contributory.93 Pleural fluid analysis has an even lower sensitivity in diagnosis of mesothelioma (32% (12/37) chance of positive or suspicious for mesothelioma).94 Contrast-enhanced pleural ultrasonography utilizing a contrast agent such as SonoVue, a second-generation contrast agent (sulphur hexafluoride microbubbles), may provide more diagnostic clues to the type of effusion based on differing enhancement patterns of malignant and infectious/inflammatory effusions.95 However, tissue (pleural) sample has highest diagnostic yield for both malignant and benign pleural process. Diagnosis of malignant (metastatic) versus para-malignant pleural effusion (i.e. related to post-obstructive pneumonia) or malignancy-unrelated pleural effusion is crucial for appropriate cancer staging. In case of NSCLC, the presence of malignant pleural effusion has been reclassified in the new seventh edition tumour, node, metastasis NSCLC staging system from T4 to M1 disease, given that the survival of patients with malignant pleural effusion has been observed to be similar to that of patients with other intrathoracic metastasis (5-year survival of approximately 2% vs 15% for T4 disease).96 Interventional pulmonologists can obtain pleural biopsy via blind procedure using Abrams needle or under direct vision via MP. Abrams needle closed parietal pleural biopsy yield for malignancy ranges between 45% and 47%.97,98 The reason for poor diagnostic yield is that most of the malignant pleural deposits occur along the diaphragmatic pleura or midline, places that cannot be accessed with the Abrams needle. Also, Abrams needle closed pleural biopsy may be associated with significant morbidity (hemothorax, pneumothorax and empyema) and sometimes mortality.99 However, Abrams needle pleural biopsy has a much higher yield for diagnosis of pleural TB with the combined yield of histology, tissue culture, pleural fluid smear and culture ranging between 80% and 90%. TB pleural deposits are much more diffusely distributed than the malignant ones resulting in this higher yield. CT-guided pleural biopsy has been reported to give higher diagnostic yields in malignancy than the Abrams needle. Maskell et al. reported 87% sensitivity (13/15 patients) for the CT-guided approach versus 47% (8/17 patients) sensitivity for the Abrams needle biopsy in a sample of 50 consecutive patients with cytology-negative, suspected malignant pleural effusion randomized for either of the two diagnostic approaches. Nevertheless, MP offers superior yield compared with Abrams needle and CT-guided pleural biopsy in malignant pleural disease. Evidence combined from 22 case series showed overall 92.6% (1268/1369; 95% CI 91.1–94%) sensitivity of local anaesthetic thoracoscopy in malignant pleural disease.100 Ninety to 100% MP yield has been reported in pleural TB. MP not only allows for diagnostic but also therapeutic interventions (indwelling pleural catheter placement, lysis of fibrin septations or sclerosing agent insufflation) for malignant disease palliation.99,101
MP is also referred to as medical thoracoscopy, local anaesthetic thoracoscopy or video-assisted thoracoscopy. It differs from the video-assisted thoracoscopic surgery in that it can be performed under conscious sedation, local anaesthesia and with spontaneously breathing patients. Video-assisted thoracoscopic surgery is performed by thoracic surgeons via several ports under general anaesthesia using double lumen endotracheal tube with single lung ventilation.
Two types of thoracoscopes exist, rigid and semi-rigid. Both utilize light source, endoscopic camera, video monitor and an image capture device. Rigid thoracoscope has a trocar of 5–10 mm diameter and a 5-mm rigid forceps. Direct or oblique 7-mm pleuroscopes offering a more panoramic view of the pleura are also available.89 Mini-rigid pleuroscopes with a 3.3-mm telescope and 3-mm biopsy forceps can be used for assessment of small loculated pleural effusions.89 Semi-rigid thoracoscope offers better manoeuvrability. It has a 7-mm outer diameter and 2.8-mm working channel. Both rigid and semi-rigid thoracoscopes offer high diagnostic yield of at least 93%.102,103 However, in circumstances where suspicion of malignancy is high, rigid pleuroscopy is the procedure of choice. This is especially true for diagnosis of suspected mesothelioma in which case pleural deposits may be thickened in which case flexible forceps of the flexible pleuroscope lack the mechanical strength to obtain pleural sample of sufficient depth.104
When performed by a trained operator, MP is a relatively safe procedure with minimal mortality (0.8% and less) and morbidity (ranging between 2% and 6%).105 Most commonly encountered complications include postoperative fever, minor bleeding, empyema, subcutaneous emphysema, persistent air leak, re-expansion pulmonary oedema, arrhythmias, myocardial infarction and seeding of chest wall with malignant cells. Mortality is lower with semi-rigid pleuroscope.103,106
MP can also be used in lung and parietal pleura biopsy. It has been utilized in facilitating diagnosis in diffuse interstitial lung disease with diagnostic yield ranging between 75% and 100%.107 However, the predominating conditions in the study populations were sarcoidosis, pneumoconiosis and lymphangitic carcinomatosis. MP-obtained samples have relative paucity of precapillary arterioles and central pulmonary arteries, which decreases the MP lung biopsy yield for vascular disorders (e.g. vasculitis). In one study, only 54% of the 118 MP-obtained lung biopsy samples had documented the presence of small pre-capillary arterioles and 10% had central pulmonary arteries. Similarly, membranous bronchi representation in peripherally obtained MP lung biopsy samples tends to be relatively low (10%), suggesting lower MP performance in diagnosis of more central lesions such as bronchiolitis obliterans with organizing pneumonia or cryptogenic organizing pneumonia.107 Morbidity and mortality of the procedure are low with air leak being the most commonly reported complication. Complication rate is similar to that of video-assisted thoracoscopic surgery procedure. Given limitations of MP in diagnostic assessment of diffuse interstitial lung disease, the American Thoracic Society and the European Respiratory Society recommend the use of video-assisted thoracoscopic surgery wedge lung biopsy for diagnostic purpose in this indication.108,109
Autofluorescence and NBI have also been utilized in video thoracoscopy. Diminution of natural autofluorescence of the malignant tissue allows for better visualization of malignant pleural deposits through colour changes that abnormal tissue undergoes upon exposure to blue light wavelength. Autofluorescence thoracoscopy has high sensitivity (100%) but low specificity (75%) for diagnosis of malignant disease.110 NBI distinguishes malignant tissue through enhanced visualization of abnormal vasculature at malignant sites.111 At present, both of these advanced imaging techniques are only used in research setting.
Technological advancements over the past two decades have revolutionized the approach to diagnostic assessment of pleuropulmonary pathology, with IP playing an increasingly more important role. Combining different diagnostic modalities in multimodality approaches offers an opportunity to revolutionize the tactic and algorithm to diagnosis of peripheral pulmonary lesions as more programs acquire expertise in use of multimodality approach, which includes a combination of the following: (i) manoeuvrability techniques such as ultrathin bronchoscopy, steerable probe in conjunction with EMN and bihinged curette; (ii) path and roadmap techniques including VB and EMN; and (iii) confirmation of destination techniques including fluoroscopy and radial probe EBUS. Wider implementation of multimodality strategies will allow for precise diagnostic assessments of peripheral pulmonary lesions currently inaccessible to flexible bronchoscopy alone. Future research and development of high-quality tissue sampling techniques hopefully will increase the multimodality diagnostic yield even further, beyond what is currently offered by the biopsy needle or forceps. Even more exciting is the use of flexible bronchoscopy technology and navigational techniques such as EMN for tumour position marking in preparation for tumour excision or radiation treatment of unresectable lesions.112–114 CT imaging-flexible bronchoscopy-guided precise localization of the internally cooled radiofrequency ablation electrode for radiofrequency ablation treatment of pulmonary nodules has recently been reported in humans. This proves that radiofrequency ablation can be a safe and feasible procedure that could become a therapeutic tool for local control in medically inoperable patients with early-stage NSCLC. Moreover, it shows that flexible bronchoscopy can not only assist in diagnostic assessment of the nodules but also precisely localize them, allowing for minimally invasive, potentially curative therapeutic procedure but without the high complication rate associated with traditional, percutaneous radiofrequency ablation.115
Since first report of high yield mediastinal LN biopsy utilizing EBUS technology in 2004,10 EBUS-TBNA has gone a long way. High-quality evidence has recently become available demonstrating equivalent performance of EBUS-TBNA (if performed by skilled clinician with excellent knowledge of the mediastinal anatomy) in mediastinal cancer staging as compared with the gold standard—cervical mediastinoscopy. This could represent a future shift in the mediastinal staging practice as more lung cancer care institutions acquire the EBUS technology and expertise for its use. Also, literature reports that EBUS-TBNA can distinguish N0/N1 disease from N2 and N3 disease and that it may also be used in detection of N1 disease, suggesting expansion of EBUS role to routine mediastinal re-staging practice following surgery, neo- and/or adjuvant chemoradiation treatment.42 Excellent quality tissue sampling obtained with EBUS opens up another avenue towards future evolution of personalized lung cancer care through development of molecular tissue analysis tools to be used in conjunction with EBUS for spot molecular diagnosis and subsequent individualized cancer care.
In the field of MP, use of advanced imaging modalities such as autofluorescence pleuroscopy or use of NBI may become valuable tools in early detection of pleural malignancy. However, more studies comparing diagnostic yield of pleural biopsy using autofluorescence and NBI technology versus regular white-light MP are needed to further assess the efficacy of these advanced pleural imaging techniques and their utility in advancing already very high diagnostic yield of MP pleural biopsy. The improvement in diagnostic yield may not be large enough to warrant the cost associated with acquisition and maintenance of autofluorescence and NBI equipment.
Overall, the rapidly evolving field of IP continues to expand its armamentarium of high quality, and precision, minimally invasive and safe diagnostic modalities with continued focus on improving patient outcomes. Miniaturization of diagnostic modality equipment through development of active bending catheters allowing for introduction of miniaturized endoscopes, enhancing resolution of diagnostic technology (through evolution of endobronchial confocal microscopy, optical coherence tomography and magnetic resonance imaging) and allowing for real-time, in vivo imaging of the functional status of the tissues will generate new insight into intrathoracic pathobiology and suggest development of technology for its treatment.116