Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) has been recommended for diagnosis of suspected sarcoidosis or suspected tuberculosis with adenopathy and may be used as an initial diagnostic test for suspected lymphoma, according to guidelines issued by CHEST (the American College of Chest Physicians).
The guidelines, which are primarily focused on technical aspects of EBUS-TBNA, also advise obtaining additional samples for the purpose of molecular analysis in patients who undergo the procedure for the diagnosis or staging of non–small cell lung cancer.
The guidelines are based on a systematic review and critical analysis of the literature by an expert panel chaired by Dr. Momen M. Wahidi of Duke University Medical Center, Durham, N.C. Of the 12 guideline statements by the panel, 7 were graded evidence-based recommendations and 5 were ungraded consensus-based statements.
The guideline ( Chest. 2016 Mar;149:816-35 ) has been endorsed by the American Association of Bronchology and Interventional Pulmonology, American Association for Thoracic Surgery, Canadian Thoracic Society, European Association for Bronchology and Interventional Pulmonology, and Society of Thoracic Surgeons.
Use of EBUS-TBNA for diagnosis in patients with suspected sarcoidosis with mediastinal and/or hilar adenopathy is ranked Grade 1C (strong recommendation, low-quality evidence, benefits outweigh risks). The guideline writers concluded that EBUS-TBNA provides safe and minimally invasive access to the mediastinal and hilar lymph nodes with a pooled diagnostic accuracy of 79.1%. They qualified, however, that it may be difficult to use EBUS-TBNA to obtain adequate tissue from fibrotic lymph nodes and conventional bronchoscopic techniques such as transbronchial lung biopsy and endobronchial biopsy may be needed in selected patients.
One systematic review and meta-analysis study included 15 studies with a total of 553 patients with sarcoidosis. The diagnostic yield of EBUS-TBNA ranged from 54% to 93%, with the pooled diagnostic accuracy of 79% (95% confidence interval, 71-86). Ten additional studies including 573 combined patients were identified through updated searches of the systematic review, and led to a pooled diagnostic accuracy of 78.2%.
Similarly, a Grade 1C recommendation was made for using EBUS-TBNA for diagnosis in when other modalities are not diagnostic in patients with suspected tuberculosis with mediastinal and/or hilar adenopathy who require lymph node sampling. However, “it must be noted that no single study assessed the role of EBUS-TBNA for the diagnosis of TB [tuberculosis] as the primary outcome measure,” they wrote. Various techniques are available for the diagnosis of TB and should be incorporated during the diagnostic evaluation.
In patients with suspected lymphoma, EBUS-TBNA is an acceptable initial, minimally invasive diagnostic test, the guideline writers said in an Ungraded Consensus-Based Statement.
In some conditions, minimally invasive EBUS-TBNA may be preferred over surgical intervention. Repeat mediastinoscopy or surgical biopsy after treatment for relapsed lymphoma can be challenging, for example, with a lower diagnostic yield and higher complication rate.
Because treatment regimens for both non-Hodgkin and Hodgkin lymphoma depend on the specific subtype and histologic grade, a definitive diagnosis of lymphoma requires the evaluation of cell morphology, immunophenotype, and the overall architecture of the tissue. Reed-Sternberg cells, diagnostic of Hodgkin lymphoma, are usually scarce in cytologic aspirates, and it often is impossible to evaluate the overall background architecture. Currently available EBUS-TBNA needles provide only cytologic specimens, with reported high discordance between cytologic specimens and histologic specimens.
In five retrospective studies with a total of 212 patients undergoing EBUS-TBNA for suspected lymphoma, the pooled diagnostic accuracy was 68.7%, and there was heterogeneity across studies in the proportion of patients with de novo lymphoma and relapsed lymphoma. Higher diagnostic yield was noted for relapsed lymphoma, compared with de novo lymphoma. Also, the two studies with the highest yield included cases as diagnostic, even when additional tissue sampling was necessary to subclassify the lymphoma for clinical management.
The panel gave a weak recommendation (Grade 2C) based on low-quality evidence to their conclusion that moderate or deep sedation is acceptable for EBUS-TBNA, based on three studies. Moderate sedation allows patients to respond purposefully to verbal commands while maintaining a functional airway, spontaneous ventilation, and cardiovascular function. In deep sedation, patients cannot be easily aroused but respond purposefully to repeated or painful stimulation and may have compromised airway function and spontaneous ventilation; cardiovascular function usually is maintained.
In one retrospective multivariable analysis of 309 patients at two centers, deep sedation had a statistically significant benefit on diagnostic yield. In a prospective randomized, controlled study of 149 patients at a single center with a single operator, there was no difference in diagnostic yield for moderate and deep sedation. However, fewer patients in the moderate sedation group were able to complete the procedure, compared with the deep-sedation group. Patient comfort and satisfaction were similar for the two sedation groups, and no patients had major complications or needed escalation of care.
In terms of diagnostic yield, there was insufficient evidence to recommend for or against using an artificial airway when inserting the EBUS bronchoscope, the authors said. Reported practice is scattered and is largely based on expert opinion, operator comfort, sedation type, and institutional standards.
The placement of the endotracheal tube may block the ultrasonographic view of the higher paratracheal lymph nodes (lymph node stations 1, 2R, 2L, and 3P) and should be avoided if one of these lymph nodes is the sampling target of the procedure, they advised.
If using a transoral artificial airway, a bite block should be considered to protect the bronchoscope from bite damage; this approach is recommended independent of the depth of sedation. A minimum size of 8.0 should be used if placing an endotracheal tube for EBUS-TBNA to accommodate the scope diameter and leave room for gas exchange.
In an Ungraded Consensus-Based Statement, the guideline authors said that ultrasonographic features, such as size, shape, border, heterogeneity, central hilar structure, and necrosis can be used to predict malignant and benign diagnoses, but tissue samples still should be obtained to confirm a diagnosis.
Nine studies provided analysis of the characteristics of lymph nodes that predict malignancy during EBUS; however, the ultrasonographic features assessed were not the same in each study or they had varying definitions of what constituted “abnormal.” As a result, the ultrasonographic predictors of malignancy in lymph nodes are not reliable enough to forgo biopsy to obtain a definitive tissue diagnosis. However, the ultrasound features can be useful to guide sampling from lymph nodes most likely to be malignant.
A round shape, distinct margins, heterogeneous echogenicity, and a central necrosis sign were independently predictive of malignancy in one multivariate analysis that included more than 1,000 lymph nodes in nearly 500 patients. Furthermore, when all four factors were absent, 96% of the lymph nodes were benign.
In three additional studies, size criteria had conflicting results; one found size was not a reliable indicator, two others found that larger lymph nodes are more likely to harbor metastases. These studies also confirmed that round-shaped lymph nodes were more likely malignant than were triangular or draping lymph nodes. The measures used to define size may have caused the inconsistencies.
In a study that examined vascular image patterns within lymph nodes as a way to predict malignancy, nodes were considered malignant if vessel involvement increased in the node to rich flow with more than four vessels (grades 2 and 3) with a sensitivity of 87.7% and a specificity of 69.6%, suggesting that increased vascularity assessed by using power/color Doppler mode ultrasound is useful in predicting malignancy.
Two studies have assessed ultrasound features of lymph nodes in patients with sarcoidosis. In the first, lymph nodes with homogeneous echogenicity and a germinal center were more likely to indicate sarcoidosis than lung cancer. In the second, coagulation necrosis and heterogeneous echogenicity within lymph nodes were more likely to be present in tuberculosis as opposed to sarcoidosis.
In another Ungraded Consensus-Based Statement, the guideline authors said tissue sampling may be performed either with or without suction. In cases in which EBUS-TBNA is being performed with suction and the samples obtained are bloody, operators should consider switching to the use of no suction at the same sampling site. If intranodal blood vessels are visualized on EBUS imaging with or without Doppler imaging, operators should also consider obtaining samples without suction.
Needle choice should be determined by the operator, and the use of either a 21- or 22-gauge needle are acceptable options based on five trials comparing needle sizes, the authors said in a Grade 1C recommendation. No data are available on the use of 25-gauge needles.
“Future studies should investigate if … smaller or more flexible needles would improve sampling at station 4L (known for its slightly angulated location) or if smaller needles would result in less blood contamination when sampling more vascular nodes. Studies should also examine if a particular needle size should be used depending on how the specimens are being processed (histopathology vs. cytopathology) and if needle size can affect the diagnosis of diseases that are more difficult to diagnose by EBUS-TBNA, such as lymphoma,” the authors wrote.
In the absence of rapid on-site evaluation (ROSE), the authors advised a minimum of three separate needle passes per sampling site in patients suspected of having lung cancer. The recommendation is an Ungraded Consensus-Based Statement.
Just one study of 102 patients with potentially operable non–small cell lung cancer and mediastinal adenopathy has examined the number of needle passes per sampling site. The results indicated optimal diagnostic values are reached after three passes. Each pass typically includes 5-15 agitations of the needle within the target site.
Sample adequacy was 90.1% after the first pass, 98.1% after two passes, and reached 100% after three passes. The sensitivity for differentiating malignant from benign lymph node stations was 69.8%, 83.7%, 95.3%, and 95.3% for one, two, three, and four passes, respectively.
No data exist regarding the number of needle passes required to obtain a sufficient diagnostic yield for lymphoma or nonmalignant diseases of the mediastinum.
In a Grade 1C recommendation, the authors said that tissue sampling can be performed with or without rapid on-site evaluation. ROSE does not affect the diagnostic yield in EBUS-TBNA procedures, but it may decrease the number of punctures and reduce the need for additional staging and diagnostic procedures. ROSE may be beneficial in judging the quantity of available malignant cells when testing for molecular markers is planned in patients with advanced adenocarcinoma of the lung.
In another Grade 1C recommendation, the authors said that patients undergoing EBUS-TBNA for the diagnosis or staging of suspected or known non–small cell lung cancer should have additional samples obtained for molecular analysis.
Molecular marker testing is necessary for tailoring chemotherapy to the cancer characteristics of each individual patient. The current data are insufficient to identify the number of passes needed to obtain adequate tissue for molecular marker testing, but it is strongly suggested that additional samples, over the proposed diagnostic threshold of three passes are recommended.
The guideline authors found insufficient quality of evidence to support any route of bronchoscope insertion for EBUS-TBNA over another. Translating the experience and literature from conventional flexible bronchoscopy given the size and rigidity of the EBUS bronchoscope distal tip, as well as the limited bronchoscopic view, is difficult, according to the guideline writers.
They noted that no studies were found that addressed the use of saline-filled balloons to overcome poor contact between the ultrasound probe and the bronchial wall. Although the saline-filled balloon can enhance image acquisition, it is unclear whether that translates into a better diagnostic yield, thus, no recommendations or suggestions can be made.
In a Grade 2C recommendation, they advised that low-fidelity inanimate mechanical airway models and high-fidelity computer-based electronic simulation be incorporated into training. In the three studies that compared conventional EBUS-TBNA training and simulation-based training incorporating either a low- or high-fidelity simulation tool, the same level of training could be acquired via conventional or simulation-based training; however, simulation-based training minimizes novice operators’ practice on patients.
In an Ungraded Consensus-Based Statement, the guideline authors advised that validated EBUS skills assessment tests be used to objectively assess skill level, but added that “none of the included simulation studies examined whether the skills demonstrated on a simulation assessment are transferred to an improvement in clinical skills as performed in patients.”
On Twitter @maryjodales