Introduction
In 2002, a new bronchoscope was developed by integrating a convex-type ultrasound probe on its tip and introduced into clinical practice. The convex probe endobronchial ultrasound (CP-EBUS), also known as linear EBUS, can be combined with a dedicated biopsy needle for real-time endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) of centrally located peribronchial lung lesions, mediastinal lymph nodes, and hilar lymph nodes. EBUS-TBNA using a linear transducer is a well-established minimally invasive modality for diagnosis and staging of lung cancer. Lung cancer guidelines recommend combined EBUS-TBNA with endoscopic ultrasound-fine-needle aspiration (EUS-FNA, also called EUS-B-FNA if an EBUS bronchoscope is used in the combined procedure) as the best first test for mediastinal nodal staging in lung cancer. Over the past 20 years, the role of this minimally invasive modality has been expanding to include restaging after neoadjuvant therapy and additional sample acquisition for biomarker testing. Advances in ultrasonography image analysis have expanded the capabilities of linear EBUS. As such, EBUS-TBNA has now also become a minimally invasive diagnostic tool for lymphoma, sarcoidosis, tuberculosis, mediastinal cysts, and other intrathoracic malignancies. New biopsy needles will further expand the potential capabilities of EBUS-TBNA in pulmonary medicine. Use of linear EBUS as a therapeutic modality, via transbronchial injection, has likewise seen growing interest and evidence. Linear EBUS continues to play an essential role in disease diagnosis but is taking on novel indications with potentially significant clinical implications.
Preprocedure Preparation
Indications for Linear Endobronchial Ultrasound
The initial indication for linear EBUS is diagnosis and nodal staging of lung cancer. Suspicion for other intrathoracic malignancies, such as lymphoma, sarcoma, mesothelioma, and other mediastinal metastases, as well as benign conditions, such as sarcoidosis, tuberculosis, and mediastinal cysts, can also be considered as indications for biopsy by EBUS-TBNA. EBUS-guided therapeutic interventions are under investigation. Several EBUS bronchoscopes exist. However, in general, the size and flexibility of currently available EBUS bronchoscopes most reliably provide access to central lesions and in many circumstances the mid-lung of the lower lobes. The accessibility of current EBUS bronchoscopes to specific bronchi is more limited than that of regular bronchoscopes, especially when a biopsy needle is inserted into the working channel. Acce ss to the upper lobes, particularly the peripheral upper lobe, can be more challenging. More flexible EBUS bronchoscopes and needles with improved access to the periphery are under development.
Equipment
- •
Linear endobronchial ultrasound bronchoscope
- •
Universal ultrasound processor
- •
EBUS-TBNA needle (19-gauge [G], 21-G, 22-G, and/or 25-G)
Staff
- •
Bronchoscopist
- •
Endoscopy/respiratory technician
- •
Sedation nurse or anesthesia team
- •
Cytopathologist (optional)
- •
Cytopathology technician (optional)
Setting
The procedure can be performed in an endoscopy suite or operating room, with either moderate/conscious sedation or general anesthesia. The linear endobronchial ultrasound bronchoscope may be inserted into the airway via the oral route. An endotracheal tube or laryngeal mask airway can be selected optionally.
Procedural Techniques
General Linear EBUS/EBUS-TBNA Preparation
A dedicated latex balloon is attached to the probe tip of the EBUS bronchoscope using the balloon applicator and inflated with normal saline during EBUS-TBNA. A 20-mL syringe and extension tube filled with saline is connected to the balloon channel. Approximately 0.3 to 0.5 mL of saline is needed to achieve appropriate balloon inflation. Because the balloon is made of latex, it cannot be used in patients with allergy to latex.
EBUS-TBNA can be performed under either local anesthesia with mild conscious sedation or general anesthesia. With local anesthesia, the EBUS scope is inserted orally and 1% lidocaine (a 2-mL bolus dose) is gently administered into the airway through the instrument channel. With general anesthesia, an endotracheal tube (at least 8.0 mm in internal diameter) or a laryngeal mask airway (#4) is generally used. General anesthesia with these airway devices provides some advantages such as easier EBUS scope insertion and reduced coughing. This must be balanced against the logistic and safety considerations of general anesthesia.
After sedation or induction of anesthesia, a regular flexible bronchoscope is first inserted into the airway. The initial diagnostic bronchoscopy facilitates safe EBUS through clearance of secretions, identification of airway lesions, verification of bronchial tree anatomy, and administration of additional local anesthetic, if required. Once complete, the flexible bronchoscope is removed, and EBUS-guided biopsy can begin. Insertion and manipulation of the EBUS bronchoscope can be more challenging than a conventional flexible bronchoscope. The EBUS bronchoscope optical system is limited by the forward oblique angle relative to the scope neutral position and ultrasound probe. Flexing the bronchoscope downward to provide a traditional “end-on” view during EBUS scope advancement can result in inadvertent injury from forceful dragging of the ultrasound probe. Rather, the EBUS bronchoscope should be advanced in a neutral position, with intermittent pausing and downward flexion to confirm position, if needed.
EBUS/EBUS-TBNA of Specific Lesions
If EBUS is being performed to acquire tissue from a specific lung or mediastinal lesion, the EBUS bronchoscope is navigated to the planned area identified on preprocedural imaging review. The ultrasound balloon should be gently inflated and the bronchoscope upward flexed to maximize contact with the bronchial wall. Once the lesion is centered on the ultrasound image, the EBUS needle sheath is advanced beyond the working channel, followed by the biopsy needle. Care should be made to monitor both the white-light and ultrasound image during advancement, as the bronchoscope may move as the needle is pushed forward. Ideally, for mediastinal lesions, the needle should be deployed in the gaps between cartilage rings.
EBUS/EBUS-TBNA of Lymph Nodes for Lung Cancer Staging
Lymph node staging should be performed in a consistent, systematic fashion to promote accurate staging. Lymph nodes are examined by EBUS for documentation of their station, size, and other ultrasound features (see later), in accordance with the American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC) staging systems. EBUS-TBNA has limited access to lymph nodes far from the central airways such as the prevascular nodes (station 3a), subaortic/paraaortic nodes (stations 5 and 6), and paraesophageal/pulmonary ligament nodes (stations 8 and 9). However, both transbronchial and transesophageal endosonographic procedures can be performed with a single EBUS scope, often referred to as EUS-B-FNA, which can facilitate access to stations 8 and 9 as well as alternative access routes for other stations. EUS-B-FNA offers potential logistic advantages. However, if EUS-B-FNA is being considered, we recommend performing the esophageal portion after the bronchoscopic portion of the procedure to avoid contamination of the respiratory system.
A dedicated TBNA needle is inserted through the working channel of the EBUS bronchoscope, and the designated lymph node is punctured under real-time EBUS guidance. The aspirated material can then be submitted for cytologic/pathologic diagnosis. There is a theoretical risk for contamination of the biopsy needle or channel as the bronchoscope is moved from one lymph node to the next, risking over-staging. It is therefore generally recommended that N3 nodes be biopsied first, then N2, then N1. Although not necessary, an on-site cytopathologist may be able to provide immediate feedback on the quality of the biopsy specimen and potentially a preliminary diagnosis. This information may be used to inform decisions on repeating a biopsy during the same procedure.
Complications
EBUS-TBNA with linear EBUS is a safe and well-established minimally invasive modality for sampling centrally located peribronchial lesions. Complication rates are very low, but major complications including bleeding, infection, recurrent nerve paralysis, and mortality have been reported.
Evidence
Lung Cancer
Nodal Staging in Lung Cancer
The prognosis and operability of a lung cancer patient is influenced by the presence of mediastinal lymph node metastases. One meta-analysis calculated a pooled sensitivity of 0.93 (95% confidence interval [CI], 0.91–0.94) and a pooled specificity of 1.00 (95% CI, 0.99–1.00) for detection of mediastinal nodal disease across 11 studies. The sensitivity, specificity, and accuracy of EBUS-TBNA were superior to positron emission tomography (PET) or PET-computed tomography (PET-CT) in two prospective trials. The combination of EBUS-TBNA and EUS-FNA has a higher staging accuracy than either procedure alone for patients with lung cancer, with a sensitivity of 0.86 (95% CI, 0.82–0.90) and a specificity of 1.00 (95% CI, 0.99–1.00) in a meta-analysis covering eight studies. In the ASTER trial, combined staging with upfront EBUS-TBNA plus EUS-FNA followed by surgical staging showed higher diagnostic yield and fewer unnecessary thoracotomies than surgical staging alone. Recently published guidelines for primary mediastinal staging in lung cancer recommend that ultrasonography-guided needle biopsy (EBUS-TBNA and/or EUS-FNA) be the first-choice modality over surgical staging. However, if EBUS/EUS biopsy results are negative, surgical staging via mediastinoscopy or video-assisted mediastinoscopy is recommended.
Ultrasound Image Analysis of Lymph Nodes
During EBUS-TBNA, ultrasonographic features are helpful to differentiate malignant and benign lymph nodes. Several features on B-mode imaging, such as size (short axis), shape (oval vs. round), margin (indistinct vs. distinct), echogenicity (homogeneous vs. heterogeneous), central hilar structure (CHS) (present vs. absent), and coagulation necrosis sign (present vs. absent), have been shown to be good predictive markers for lymph node metastasis in non–small cell lung cancer (NSCLC). Fujiwara et al. reported round shape, distinct margin, heterogeneous echogenicity, and presence of coagulation necrosis sign as independent risk factors for metastasis. Alici et al. integrated grayscale texture (anechoic, hypoechoic, isoechoic, or hyperechoic) with the previous six features to create a modified algorithm. This algorithm’s sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy for detecting metastatic lymph nodes were 100%, 51.2%, 50.6%, 100%, and 67.5%, respectively. Doppler imaging permits assessment of blood flow and nodal vascular patterns. Nakajima et al. classified lymph nodes by Doppler findings: grade 0, no blood flow or small amounts of flow; grade I, a few main vessels running toward the center of the lymph node from the hilum; grade II, a few cuneiforms or rod-shaped flow signals, or a few small vessels found as a long strip of a curve; and grade III, rich flow with more than four vessels of differing diameters and/or twist-/helical-low signal. The sensitivity, specificity, and diagnostic accuracy of this grading system (grade 0/I benign vs. grade II/III malignant) were 87.7%, 69.6%, and 78.0%, respectively. Wang et al. classified Doppler vascular patterns into avascular, hilar, and nonhilar (central, capsular, or mixed); the authors combined these vascular features with the previous six sonographic features to predict benign lymph node status. The sensitivity, specificity, PPV, and NPV for predicting benign lymph nodes were 81.3%, 90.9%, 85.3%, and 88.2%, respectively.
Elastography is a strain imaging technique to assess tissue stiffness, which is displayed as a color overlay on the B-mode ultrasound image. Most systems identify hard, intermediate, and soft tissues as blue, green, and yellow/red, respectively. Izumo et al. categorized elastography image patterns into type 1 (predominantly nonblue), type 2 (part blue, part nonblue), and type 3 (predominantly blue). The sensitivity, specificity, PPV, NPV, and diagnostic accuracy of this classification system (type 1 benign vs. type 3 malignant) were 100.0%, 92.3%, 94.6%, 100.0%, and 96.7%, respectively. Nakajima et al. compared nodes by stiff area ratio (stiff blue area divided by total lymph node area) and found the mean stiffness ratios were significantly greater for metastatic lymph nodes (0.48) than benign lymph nodes (0.22, P = 0.0002). When a cut-off ratio of 0.31 was used, sensitivity and specificity were 81% and 85%, respectively.
A growing area of focus is the application of artificial intelligence technologies to risk-stratify EBUS images by malignant potential. A 2008 study by Tagoya et al. developed an artificial neural network to predict the presence of nodal metastases using linear EBUS B-mode images, which ultimately developed a 91% diagnostic accuracy. The sensitivity, specificity, and accuracy of this system were 87.0%, 82.1%, and 85.4%, respectively. The application of artificial intelligence may enable significant future advances in EBUS image analysis.
Restaging After Neoadjuvant Therapy
At present, the recommended treatment for stage IIIA-cN2 NSCLC is chemoradiotherapy. However, surgical resection after neoadjuvant chemotherapy or chemoradiotherapy may improve the survival of patients with stage IIIA-cN2 disease. Accurate restaging of the mediastinal lymph nodes in these cases is critical to confirm mediastinal down-staging prior to consideration for surgery. Repeat mediastinoscopy may also be considered; however, mediastinoscopy following neoadjuvant therapy can be challenging and the diagnostic yield is reduced due to development of fibrosis and adhesions. A systematic review of five studies calculated the pooled sensitivity, specificity, and false-negative rate of remediastinoscopy after neoadjuvant therapy as 63%, 100%, and 22%, respectively. Transcervical extended mediastinal lymphadenectomy has shown a sensitivity of 96.6% for mediastinal restaging in patients with NSCLC after neoadjuvant therapy. Mortality and morbidity were 0.3% and 6.4%, respectively. Similarly, restaging with EBUS-TBNA after neoadjuvant therapy has been reported to have lower sensitivity compared with EBUS-TBNA used during initial lung cancer staging. A systematic review and meta-analysis including 10 studies found that endosonographic-guided needle biopsy (EBUS-TBNA, EUS-FNA, or combined endoscopic and endobronchial ultrasound [CUS]) for mediastinal restaging has a pooled sensitivity of 67% (95% CI, 56–77) and pooled specificity of 99% (95% CI, 89–100). The discrepancy of diagnostic yields between initial staging and restaging may relate to difficulty obtaining adequate samples from down-staged nodes, which may be smaller, fibrotic, and/or necrotic following neoadjuvant therapy. There is also difficulty differentiating the sonographic appearance of metastases from postinflammatory adhesions and degenerative changes. Combined EBUS-TBNA and EUS-FNA could enable more accurate minimally invasive mediastinal restaging. Current guidelines recommend EBUS-TBNA and/or EUS-FNA for mediastinal restaging after neoadjuvant therapy, avoiding remediastinoscopy.
Molecular Testing Using EBUS-TBNA Samples
As the treatment of advanced NSCLC has shifted toward molecular targeted therapy, biomarker testing has become necessary for determining the optimal treatment of patients newly diagnosed with NSCLC. Sensitizing mutations in the EGFR gene were first described in 2004, serving as the first class of molecular targeted therapy. Since then, anaplastic lymphoma kinase ( ALK ) gene fusion, ROS1 gene rearrangements, and BRAF mutations were identified as potential treatment targets. Combination therapies, including cytotoxic chemotherapy and targeted gene therapy, have improved overall response rates, increased progression-free survival, and may be associated with improved overall survival in advanced NSCLC when compared with cytotoxic chemotherapy alone. The National Comprehensive Cancer Network (NCCN) 2018 Clinical Practice Guidelines for NSCLC recommend concomitant diagnosis, staging, and acquisition of adequate material for molecular profiling to improve care of patients with NSCLC. The importance of obtaining tissue for molecular profiling is clear. A systematic review and meta-analysis including 33 studies (2698 participants in total) found that use of EBUS-TBNA for molecular profiling of EGFR mutation status had a pooled probability of obtaining sufficient tissue of 94.5% (95% CI, 93.2%–96.4%). For identification of ALK mutations, the pooled probability was 94.9% (95% CI, 89.4%–98.8%). There are several emerging molecular targets and therapies in NSCLC, such as PIK3CA mutation, AKT1 KRAS mutation, RET rearrangements, MET exon 14 skipping mutations, and activating HER2 mutations. Therefore, the NCCN 2018 guidelines recommend testing using broad-based genomic sequencing, such as next-generation sequencing (NGS). A study including 54 TBNA/FNA samples showed a 50-gene assay panel was successful in 97.5% and 100% of 22-G and 25-G samples, respectively. A larger 1231-gene panel was successful in 91.3% and 100% of 22-G and 25-G samples, respectively. Another study including 115 samples undergoing a large (341–469 gene) NGS-based panel found EBUS-TBNA obtained sufficient tissue in 86.1% of samples. Rebiopsy by EBUS-TBNA for follow-up molecular profiling can be performed safely after initial treatment. In the era of biomarker-driven management of cancer, the ability to analyze EBUS-TBNA specimens for multiple biomarkers is critical in selecting an optimal, personalized treatment plan for each patient.
Lymphoma
Approximately 10% of lymphomas are first diagnosed in the chest, often as a mediastinal tumor. Subclassification, which guides treatment and prognosis, is based on morphologic, phenotypic, genotypic, and molecular features. Early diagnosis and staging are key to improving patient survival in those diagnosed with lymphoma. When available, EBUS-TBNA is a useful alternative approach for the diagnosis and subclassification of intrathoracic lymphoma compared to “gold standard” approaches of mediastinoscopy, thoracoscopy, and/or thoracotomy. In a systematic review and meta-analysis including 14 studies, the overall sensitivity and specificity of EBUS-TBNA for diagnosis of lymphoma were 66.2% (95% CI, 55%–75.8%) and 99.3% (95% CI, 98.2%–99.7%), respectively. In subgroup analysis, sensitivity and specificity of EBUS-TBNA for the initial diagnosis of lymphoma were 67.1% (95% CI, 54.2%–77.9%) and 99.6% (95% CI, 99.1%–99.8%), respectively. EBUS-TBNA performed slightly better for diagnosing lymphoma recurrence , with a sensitivity of 77.8% (95% CI, 68.1%–85.2%) and specificity of 99.5% (95% CI, 98.9%–99.8%). These diagnostic metrics are comparable to historical data on using mediastinoscopy for the diagnosis of mediastinal lymphoma. For subtyping lymphoma, EBUS-TBNA obtained sufficient samples for ancillary testing (e.g., flow cytometry, fluorescence in situ hybridization) in 63% of histologically positive samples. This suggests that EBUS-TBNA is an appropriate first-choice modality in patients with suspected lymphoma for the diagnosis of both initial and recurrent disease.
Sarcoidosis
The diagnosis of sarcoidosis requires the following criteria be met: a compatible clinical and radiologic presentation, pathologic evidence of noncaseating granulomas, and exclusion of other diseases with similar findings (e.g., infections, malignancy). Conventional transbronchial biopsy (TBB) and TBNA were historically the most common procedures for obtaining pathologic evidence of noncaseating granulomas. The diagnostic yields of TBNA and TBNA + TBB are reported to be 62% and 83%, respectively. EBUS-TBNA is particularly useful for stage I/II sarcoidosis, for which lymphadenopathy is a common feature. A meta-analysis including 15 studies found that EBUS-TBNA had a pooled diagnostic accuracy of 79% (95% CI, 71%–86%). Performance of EBUS-TBNA was superior to TBNA or TBB alone. However, a separate meta-analysis including 16 studies found the diagnostic yield of combined EBUS-TBNA + TBB + endobronchial biopsy (EBB) was 89.7% and more effective than EBUS-TBNA alone (82.7%) for the diagnosis of sarcoidosis. The pooled diagnostic odds ratio for the two groups was 0.55 (95% CI, 0.39–0.78, P = 0.0007). These results suggest EBUS-TBNA, when combined with TBB and/or EBB, can be an effective minimally invasive approach for confirming the diagnosis of sarcoidosis.
Tuberculosis, Mediastinal Cysts, and Other Malignant Diseases
Pulmonary tuberculosis is often associated with mediastinal or hilar lymphadenopathy. The potential utility of EBUS-TBNA for diagnosis of tuberculosis has been previously reported. A recent meta-analysis revealed the pooled sensitivity and specificity of EBUS-TBNA for diagnosis of intrathoracic tuberculosis were 80% (95% CI, 0.74–0.85) and 100% (95% CI, 0.99–1.00), respectively.
A systematic review including 26 studies and 32 cases outlined the utility of diagnostic and therapeutic transbronchial ultrasound approaches for the diagnosis of mediastinal cysts. However, four cases of postprocedural infection were identified after TBNA.
Rice et al. reported a cases series of nodal staging by EBUS-TBNA in malignant pleural mesothelioma, including 38 EBUS-TBNA and 50 mediastinoscopy cases. The sensitivity and NPV were 28% and 49% for mediastinoscopy versus 59% and 57% for EBUS, respectively. Czarnecka-Kujawa et al. likewise published a case series including 48 patients with malignant pleural mesothelioma who underwent EBUS-TBNA for nodal staging. The sensitivity, specificity, PPV, NPV, and diagnostic accuracy were 16.7%, 100%, 100%, 68.8%, and 70.6%, respectively. Although there is no large cohort study investigating the performance of EBUS-TBNA for the diagnosis of sarcoma, several authors have described successful tissue acquisition in small case series.
EBUS Needles
Several EBUS needles are currently available across a range of sizes (25-, 22-, 21-, or 19-G). The size of the needle may affect the quantity of tissue obtained, degree of tissue trauma, amount of aspirated blood (which can affect the quality of the specimen), diagnostic yield, and maximal angulation range of the EBUS bronchoscope ( Fig. 2.1 ). The increasing number of EBUS-TBNA needles has prompted several investigations on their comparative diagnostic performance. The most common needles are 22-G and 21-G needles; however, there are little data supporting the use of one over another for its size. A large cohort of 1299 patients showed no differences in the diagnostic yield of 22-G and 21-G needles for the diagnosis and staging of NSCLC. Adequate samples were obtained in 94.9% of the 22-G needle group and in 94.6% of the 21-G needle group ( P = 0.81). A pathologic diagnosis was obtained in 51.4% of the 22-G group and 51.3% of the 21-G group ( P = 0.98). These results suggest there is little difference when selecting between 22-G and 21-G needles for cytologic evaluation via TBNA.
