Jessica Johnson, MD

As of 2019, lung cancer remained the leading cause of cancer death in the United States. In March 2021, the USPSTF updated the guidelines for lung cancer screening, increasing the number of eligible patients in order to identify malignancies in the early stages when more treatment options exist. With the growth of lung cancer screening, increasingly smaller pulmonary nodules are being identified in more peripheral locations previously thought to be unreachable with bronchoscopy. While bronchoscopy has been utilized for over a century for therapeutic interventions, the development of the fiberoptic bronchoscope in 1967 ushered in an era of evolving diagnostic functions. From the initial endobronchial and transbronchial biopsy techniques, to the introduction of endobronchial ultrasound, and now the latest navigational and robotic modalities, these advances have opened a new realm of interventions available in our diagnostic approach to lung cancer.

Bronchoscopy has become essential in the diagnosis of thoracic malignancies, providing both diagnostic and staging information in one procedural setting. By first assessing the mediastinal and hilar lymph nodes with endobronchial ultrasound and transbronchial needle aspiration, involved lymph nodes can give both diagnosis and staging information required to guide treatment. This is particularly important in the case of non-small cell lung cancer, which utilizes the TNM staging system. Through the use of convex probe endobronchial ultrasound (CP-EBUS), combined with rapid on-site evaluation (ROSE) by pathologic condition, we can more accurately target the individual lymph nodes for biopsy without the need for any additional procedures that are often more complex and invasive, such as mediastinoscopy. It is important to note the role of CP-EBUS extends beyond the lymph node assessment and can also be utilized for the evaluation of other mediastinal lesions, such as central parenchymal masses. These would otherwise be difficult to access due to the lack of a clear airway to the lesion (Argento and Puchalski. Respir Med. 2016;116:55-8).

While EBUS has improved the sampling of lymph nodes, advanced imaging technologies and subsequent increases in lung cancer screening have increased the number of lung malignancies identified in earlier stages before extension to the lymph nodes occurs. This scenario requires a direct biopsy of the primary nodule or lung mass. While CP-EBUS can be utilized for some central parenchymal lesions, peripheral nodules pose a greater challenge to the bronchoscopist as they cannot be directly visualized with the conventional bronchoscope. These lesions are amenable to traditional sampling techniques such as bronchial brushings and washings in addition to transbronchial needle aspiration and transbronchial biopsy. However, the yield for peripheral lesions is less than that for central tumors and depends on lesion size, distance from hilum, spatial positioning from bronchus, and operator experience. To help localize peripheral lesions, a separate form of endobronchial ultrasound is available that can be used in combination with fluoroscopy to target a lesion. Radial probe endobronchial ultrasound (RP-EBUS) utilizes a rotating ultrasound transducer that can be advanced either through the working channel of the bronchoscope or through a guide sheath to extend to airways beyond what the conventional bronchoscope can reach. This assists the bronchoscopist with locating the correct airway and, therefore, increases the yield of sampling techniques. The use of RP-EBUS has reported diagnostic yields of almost 85% if the ultrasound is located within the lesion, but less than 50% if adjacent to the lesion (Chen et al. Ann Am Thorac Soc. 2014;11[4]:578-82). While this improves the yield beyond that achieved with conventional bronchoscopy alone, it continues to challenge the bronchoscopist to locate an accessible airway from a series of branching bronchi that are beyond the level of direct visualization.

Due to the historical difficulty in accurately reaching peripheral lesions, alternative technologies for sampling these lesions, such as image-guided biopsies or surgical resection, were employed. While CT scan-guided biopsies traditionally have high diagnostic yields, they also carry a higher rate of complications, including pneumothorax and bleeding. This has led to a significant increase over the past 2 decades in new bronchoscopic technologies targeting safer and more accurate sampling of increasingly smaller, peripheral lesions.

Traditionally, any new technologies created were intended to be used alongside flexible fiberoptic bronchoscopy. The more recently introduced technologies, however, aim to provide a safer, more accurate procedure through virtual bronchoscopy. By obtaining CT scan images prior to the procedure, a 3D visualization is constructed of the tracheobronchial tree, allowing for directed guidance of endobronchial accessories to more distal airways. Where the bronchoscopist was previously limited in navigating the bronchial tree to the subsegmental bronchi, virtual bronchoscopy can depict the airways up to the 7th order subdivision. This is a significant improvement in airway visualization – however, only when partnered with guidance technologies can the model be accurately navigated.

One modality that is often coupled with virtual bronchoscopy to accurately reach peripheral lesions is electromagnetic navigation bronchoscopy (ENB). Multiple ENB software systems have been created and continue to be highly utilized by bronchoscopists to target peripheral lesions, as it has often been likened to a GPS for the lungs. With the addition of specific hardware components, a magnetic field is created around the patient where the sensor position can be elicited to within 1-mm accuracy. When overlaid with the CT scan images, the bronchoscopist can have real-time positioning of the probe in all three planes and guide the necessary sampling tools to the lesion of interest. The reported yields for ENB vary but have been shown to increase in the presence of specific image findings such as a positive bronchus sign – an air-filled bronchus leading into the lesion. The presence of this finding can increase the yield up to almost 75% from just under 50% in the absence of a positive bronchus sign. (Ali et al. Ann Am Thorac Soc. 2018;15[8]:978-87). However, regardless of this finding, the overall diagnostic yields for ENB continue to fall below that seen with other image-guided biopsy techniques. The procedural complications, however, are significantly less and, therefore, many physicians continue to advocate for ENB as the initial procedure in attempt to decrease risk for the patient.

The newest technology to be introduced to target peripheral lung lesions and to improve upon the shortcomings of other techniques is robotic-assisted bronchoscopy. While surgical specialties have seen success with robotic techniques over many years, the first robotic bronchoscopy system was not introduced until 2018. At present, there are two systems available: the Monarch^® system by Auris Health and the Ion Endoluminal^® System by Intuitive Surgical. These systems allow for increased bronchoscope stability, improved visualization, adjustable angulation of biopsy tools, and an improved ability to make even subtle turns in the airways. Early studies on both systems were cadaver based, but an increasing number of patient trials are now being reported or actively enrolling. Both systems have shown high rates of lesion localization, greater than 85%, with varying diagnostic yields from 69-79%. Some cadaver studies that utilized artificial tumors reported higher diagnostic yields – over 90% – but this was not seen in initial patient-based studies. (Agrawal et al. J Thorac Dis. 2020;12[6]:3279-86) More data related to the robotic-assisted bronchoscopy systems can be expected in the future as more experience is gained, but initial results are promising in the system’s ability to diagnose early lung cancers safely and accurately.

With increasing technologies and equipment available, bronchoscopy has quickly become an essential step in the diagnosis of lung cancer. While other techniques exist beyond those described here, these are some of the more widely used options currently available. It is not possible at this time to define one technology as the best tool for the diagnosis of lung cancer, as patient factors will always have to be taken into consideration to ensure safety and accuracy. However, with constantly changing technologies, the bronchoscopist now has a variety of tools available to help target previously “unreachable” lesions as we aim to decrease the historically high mortality rates of lung cancer.

Dr. Jewani and Dr. Johnson are from Loyola University Medical Center, Department of Pulmonary and Critical Care Medicine, Maywood, Illinois.


1. Agrawal, Abhinav et al. “Robotic bronchoscopy for pulmonary lesions: a review of existing technologies and clinical data.” Journal of thoracic disease vol. 12,6 (2020): 3279-3286. doi:10.21037/jtd.2020.03.35

2. Ali MS, Sethi J, Taneja A, Musani A, Maldonado F. Computed Tomography Bronchus Sign and the Diagnostic Yield of Guided Bronchoscopy for Peripheral Pulmonary Lesions. A Systematic Review and Meta-Analysis. Ann Am Thorac Soc. 2018 Aug;15(8):978-987. doi: 10.1513/AnnalsATS.201711-856OC. PMID: 29877715.

3. Argento AC, Puchalski J. Convex probe EBUS for centrally located parenchymal lesions without a bronchus sign. Respir Med. 2016 Jul;116:55-8. doi: 10.1016/j.rmed.2016.04.012. Epub 2016 Apr 29. PMID: 27296821.

4. Chen A, Chenna P, Loiselle A, Massoni J, Mayse M, Misselhorn D. Radial probe endobronchial ultrasound for peripheral pulmonary lesions. A 5-year institutional experience. Ann Am Thorac Soc. 2014 May;11(4):578-82. doi: 10.1513/AnnalsATS.201311-384OC. PMID: 24635641.
 

As of 2019, lung cancer remained the leading cause of cancer death in the United States. In March 2021, the USPSTF updated the guidelines for lung cancer screening, increasing the number of eligible patients in order to identify malignancies in the early stages when more treatment options exist. With the growth of lung cancer screening, increasingly smaller pulmonary nodules are being identified in more peripheral locations previously thought to be unreachable with bronchoscopy. While bronchoscopy has been utilized for over a century for therapeutic interventions, the development of the fiberoptic bronchoscope in 1967 ushered in an era of evolving diagnostic functions. From the initial endobronchial and transbronchial biopsy techniques, to the introduction of endobronchial ultrasound, and now the latest navigational and robotic modalities, these advances have opened a new realm of interventions available in our diagnostic approach to lung cancer.

Bronchoscopy has become essential in the diagnosis of thoracic malignancies, providing both diagnostic and staging information in one procedural setting. By first assessing the mediastinal and hilar lymph nodes with endobronchial ultrasound and transbronchial needle aspiration, involved lymph nodes can give both diagnosis and staging information required to guide treatment. This is particularly important in the case of non-small cell lung cancer, which utilizes the TNM staging system. Through the use of convex probe endobronchial ultrasound (CP-EBUS), combined with rapid on-site evaluation (ROSE) by pathologic condition, we can more accurately target the individual lymph nodes for biopsy without the need for any additional procedures that are often more complex and invasive, such as mediastinoscopy. It is important to note the role of CP-EBUS extends beyond the lymph node assessment and can also be utilized for the evaluation of other mediastinal lesions, such as central parenchymal masses. These would otherwise be difficult to access due to the lack of a clear airway to the lesion (Argento and Puchalski. Respir Med. 2016;116:55-8).

While EBUS has improved the sampling of lymph nodes, advanced imaging technologies and subsequent increases in lung cancer screening have increased the number of lung malignancies identified in earlier stages before extension to the lymph nodes occurs. This scenario requires a direct biopsy of the primary nodule or lung mass. While CP-EBUS can be utilized for some central parenchymal lesions, peripheral nodules pose a greater challenge to the bronchoscopist as they cannot be directly visualized with the conventional bronchoscope. These lesions are amenable to traditional sampling techniques such as bronchial brushings and washings in addition to transbronchial needle aspiration and transbronchial biopsy. However, the yield for peripheral lesions is less than that for central tumors and depends on lesion size, distance from hilum, spatial positioning from bronchus, and operator experience. To help localize peripheral lesions, a separate form of endobronchial ultrasound is available that can be used in combination with fluoroscopy to target a lesion. Radial probe endobronchial ultrasound (RP-EBUS) utilizes a rotating ultrasound transducer that can be advanced either through the working channel of the bronchoscope or through a guide sheath to extend to airways beyond what the conventional bronchoscope can reach. This assists the bronchoscopist with locating the correct airway and, therefore, increases the yield of sampling techniques. The use of RP-EBUS has reported diagnostic yields of almost 85% if the ultrasound is located within the lesion, but less than 50% if adjacent to the lesion (Chen et al. Ann Am Thorac Soc. 2014;11[4]:578-82). While this improves the yield beyond that achieved with conventional bronchoscopy alone, it continues to challenge the bronchoscopist to locate an accessible airway from a series of branching bronchi that are beyond the level of direct visualization.

Due to the historical difficulty in accurately reaching peripheral lesions, alternative technologies for sampling these lesions, such as image-guided biopsies or surgical resection, were employed. While CT scan-guided biopsies traditionally have high diagnostic yields, they also carry a higher rate of complications, including pneumothorax and bleeding. This has led to a significant increase over the past 2 decades in new bronchoscopic technologies targeting safer and more accurate sampling of increasingly smaller, peripheral lesions.

Traditionally, any new technologies created were intended to be used alongside flexible fiberoptic bronchoscopy. The more recently introduced technologies, however, aim to provide a safer, more accurate procedure through virtual bronchoscopy. By obtaining CT scan images prior to the procedure, a 3D visualization is constructed of the tracheobronchial tree, allowing for directed guidance of endobronchial accessories to more distal airways. Where the bronchoscopist was previously limited in navigating the bronchial tree to the subsegmental bronchi, virtual bronchoscopy can depict the airways up to the 7th order subdivision. This is a significant improvement in airway visualization – however, only when partnered with guidance technologies can the model be accurately navigated.

One modality that is often coupled with virtual bronchoscopy to accurately reach peripheral lesions is electromagnetic navigation bronchoscopy (ENB). Multiple ENB software systems have been created and continue to be highly utilized by bronchoscopists to target peripheral lesions, as it has often been likened to a GPS for the lungs. With the addition of specific hardware components, a magnetic field is created around the patient where the sensor position can be elicited to within 1-mm accuracy. When overlaid with the CT scan images, the bronchoscopist can have real-time positioning of the probe in all three planes and guide the necessary sampling tools to the lesion of interest. The reported yields for ENB vary but have been shown to increase in the presence of specific image findings such as a positive bronchus sign – an air-filled bronchus leading into the lesion. The presence of this finding can increase the yield up to almost 75% from just under 50% in the absence of a positive bronchus sign. (Ali et al. Ann Am Thorac Soc. 2018;15[8]:978-87). However, regardless of this finding, the overall diagnostic yields for ENB continue to fall below that seen with other image-guided biopsy techniques. The procedural complications, however, are significantly less and, therefore, many physicians continue to advocate for ENB as the initial procedure in attempt to decrease risk for the patient.

The newest technology to be introduced to target peripheral lung lesions and to improve upon the shortcomings of other techniques is robotic-assisted bronchoscopy. While surgical specialties have seen success with robotic techniques over many years, the first robotic bronchoscopy system was not introduced until 2018. At present, there are two systems available: the Monarch^® system by Auris Health and the Ion Endoluminal^® System by Intuitive Surgical. These systems allow for increased bronchoscope stability, improved visualization, adjustable angulation of biopsy tools, and an improved ability to make even subtle turns in the airways. Early studies on both systems were cadaver based, but an increasing number of patient trials are now being reported or actively enrolling. Both systems have shown high rates of lesion localization, greater than 85%, with varying diagnostic yields from 69-79%. Some cadaver studies that utilized artificial tumors reported higher diagnostic yields – over 90% – but this was not seen in initial patient-based studies. (Agrawal et al. J Thorac Dis. 2020;12[6]:3279-86) More data related to the robotic-assisted bronchoscopy systems can be expected in the future as more experience is gained, but initial results are promising in the system’s ability to diagnose early lung cancers safely and accurately.

With increasing technologies and equipment available, bronchoscopy has quickly become an essential step in the diagnosis of lung cancer. While other techniques exist beyond those described here, these are some of the more widely used options currently available. It is not possible at this time to define one technology as the best tool for the diagnosis of lung cancer, as patient factors will always have to be taken into consideration to ensure safety and accuracy. However, with constantly changing technologies, the bronchoscopist now has a variety of tools available to help target previously “unreachable” lesions as we aim to decrease the historically high mortality rates of lung cancer.

Dr. Jewani and Dr. Johnson are from Loyola University Medical Center, Department of Pulmonary and Critical Care Medicine, Maywood, Illinois.


1. Agrawal, Abhinav et al. “Robotic bronchoscopy for pulmonary lesions: a review of existing technologies and clinical data.” Journal of thoracic disease vol. 12,6 (2020): 3279-3286. doi:10.21037/jtd.2020.03.35

2. Ali MS, Sethi J, Taneja A, Musani A, Maldonado F. Computed Tomography Bronchus Sign and the Diagnostic Yield of Guided Bronchoscopy for Peripheral Pulmonary Lesions. A Systematic Review and Meta-Analysis. Ann Am Thorac Soc. 2018 Aug;15(8):978-987. doi: 10.1513/AnnalsATS.201711-856OC. PMID: 29877715.

3. Argento AC, Puchalski J. Convex probe EBUS for centrally located parenchymal lesions without a bronchus sign. Respir Med. 2016 Jul;116:55-8. doi: 10.1016/j.rmed.2016.04.012. Epub 2016 Apr 29. PMID: 27296821.

4. Chen A, Chenna P, Loiselle A, Massoni J, Mayse M, Misselhorn D. Radial probe endobronchial ultrasound for peripheral pulmonary lesions. A 5-year institutional experience. Ann Am Thorac Soc. 2014 May;11(4):578-82. doi: 10.1513/AnnalsATS.201311-384OC. PMID: 24635641.
 

As of 2019, lung cancer remained the leading cause of cancer death in the United States. In March 2021, the USPSTF updated the guidelines for lung cancer screening, increasing the number of eligible patients in order to identify malignancies in the early stages when more treatment options exist. With the growth of lung cancer screening, increasingly smaller pulmonary nodules are being identified in more peripheral locations previously thought to be unreachable with bronchoscopy. While bronchoscopy has been utilized for over a century for therapeutic interventions, the development of the fiberoptic bronchoscope in 1967 ushered in an era of evolving diagnostic functions. From the initial endobronchial and transbronchial biopsy techniques, to the introduction of endobronchial ultrasound, and now the latest navigational and robotic modalities, these advances have opened a new realm of interventions available in our diagnostic approach to lung cancer.

Bronchoscopy has become essential in the diagnosis of thoracic malignancies, providing both diagnostic and staging information in one procedural setting. By first assessing the mediastinal and hilar lymph nodes with endobronchial ultrasound and transbronchial needle aspiration, involved lymph nodes can give both diagnosis and staging information required to guide treatment. This is particularly important in the case of non-small cell lung cancer, which utilizes the TNM staging system. Through the use of convex probe endobronchial ultrasound (CP-EBUS), combined with rapid on-site evaluation (ROSE) by pathologic condition, we can more accurately target the individual lymph nodes for biopsy without the need for any additional procedures that are often more complex and invasive, such as mediastinoscopy. It is important to note the role of CP-EBUS extends beyond the lymph node assessment and can also be utilized for the evaluation of other mediastinal lesions, such as central parenchymal masses. These would otherwise be difficult to access due to the lack of a clear airway to the lesion (Argento and Puchalski. Respir Med. 2016;116:55-8).

While EBUS has improved the sampling of lymph nodes, advanced imaging technologies and subsequent increases in lung cancer screening have increased the number of lung malignancies identified in earlier stages before extension to the lymph nodes occurs. This scenario requires a direct biopsy of the primary nodule or lung mass. While CP-EBUS can be utilized for some central parenchymal lesions, peripheral nodules pose a greater challenge to the bronchoscopist as they cannot be directly visualized with the conventional bronchoscope. These lesions are amenable to traditional sampling techniques such as bronchial brushings and washings in addition to transbronchial needle aspiration and transbronchial biopsy. However, the yield for peripheral lesions is less than that for central tumors and depends on lesion size, distance from hilum, spatial positioning from bronchus, and operator experience. To help localize peripheral lesions, a separate form of endobronchial ultrasound is available that can be used in combination with fluoroscopy to target a lesion. Radial probe endobronchial ultrasound (RP-EBUS) utilizes a rotating ultrasound transducer that can be advanced either through the working channel of the bronchoscope or through a guide sheath to extend to airways beyond what the conventional bronchoscope can reach. This assists the bronchoscopist with locating the correct airway and, therefore, increases the yield of sampling techniques. The use of RP-EBUS has reported diagnostic yields of almost 85% if the ultrasound is located within the lesion, but less than 50% if adjacent to the lesion (Chen et al. Ann Am Thorac Soc. 2014;11[4]:578-82). While this improves the yield beyond that achieved with conventional bronchoscopy alone, it continues to challenge the bronchoscopist to locate an accessible airway from a series of branching bronchi that are beyond the level of direct visualization.

Due to the historical difficulty in accurately reaching peripheral lesions, alternative technologies for sampling these lesions, such as image-guided biopsies or surgical resection, were employed. While CT scan-guided biopsies traditionally have high diagnostic yields, they also carry a higher rate of complications, including pneumothorax and bleeding. This has led to a significant increase over the past 2 decades in new bronchoscopic technologies targeting safer and more accurate sampling of increasingly smaller, peripheral lesions.

Traditionally, any new technologies created were intended to be used alongside flexible fiberoptic bronchoscopy. The more recently introduced technologies, however, aim to provide a safer, more accurate procedure through virtual bronchoscopy. By obtaining CT scan images prior to the procedure, a 3D visualization is constructed of the tracheobronchial tree, allowing for directed guidance of endobronchial accessories to more distal airways. Where the bronchoscopist was previously limited in navigating the bronchial tree to the subsegmental bronchi, virtual bronchoscopy can depict the airways up to the 7th order subdivision. This is a significant improvement in airway visualization – however, only when partnered with guidance technologies can the model be accurately navigated.

One modality that is often coupled with virtual bronchoscopy to accurately reach peripheral lesions is electromagnetic navigation bronchoscopy (ENB). Multiple ENB software systems have been created and continue to be highly utilized by bronchoscopists to target peripheral lesions, as it has often been likened to a GPS for the lungs. With the addition of specific hardware components, a magnetic field is created around the patient where the sensor position can be elicited to within 1-mm accuracy. When overlaid with the CT scan images, the bronchoscopist can have real-time positioning of the probe in all three planes and guide the necessary sampling tools to the lesion of interest. The reported yields for ENB vary but have been shown to increase in the presence of specific image findings such as a positive bronchus sign – an air-filled bronchus leading into the lesion. The presence of this finding can increase the yield up to almost 75% from just under 50% in the absence of a positive bronchus sign. (Ali et al. Ann Am Thorac Soc. 2018;15[8]:978-87). However, regardless of this finding, the overall diagnostic yields for ENB continue to fall below that seen with other image-guided biopsy techniques. The procedural complications, however, are significantly less and, therefore, many physicians continue to advocate for ENB as the initial procedure in attempt to decrease risk for the patient.

The newest technology to be introduced to target peripheral lung lesions and to improve upon the shortcomings of other techniques is robotic-assisted bronchoscopy. While surgical specialties have seen success with robotic techniques over many years, the first robotic bronchoscopy system was not introduced until 2018. At present, there are two systems available: the Monarch^® system by Auris Health and the Ion Endoluminal^® System by Intuitive Surgical. These systems allow for increased bronchoscope stability, improved visualization, adjustable angulation of biopsy tools, and an improved ability to make even subtle turns in the airways. Early studies on both systems were cadaver based, but an increasing number of patient trials are now being reported or actively enrolling. Both systems have shown high rates of lesion localization, greater than 85%, with varying diagnostic yields from 69-79%. Some cadaver studies that utilized artificial tumors reported higher diagnostic yields – over 90% – but this was not seen in initial patient-based studies. (Agrawal et al. J Thorac Dis. 2020;12[6]:3279-86) More data related to the robotic-assisted bronchoscopy systems can be expected in the future as more experience is gained, but initial results are promising in the system’s ability to diagnose early lung cancers safely and accurately.

With increasing technologies and equipment available, bronchoscopy has quickly become an essential step in the diagnosis of lung cancer. While other techniques exist beyond those described here, these are some of the more widely used options currently available. It is not possible at this time to define one technology as the best tool for the diagnosis of lung cancer, as patient factors will always have to be taken into consideration to ensure safety and accuracy. However, with constantly changing technologies, the bronchoscopist now has a variety of tools available to help target previously “unreachable” lesions as we aim to decrease the historically high mortality rates of lung cancer.

Dr. Jewani and Dr. Johnson are from Loyola University Medical Center, Department of Pulmonary and Critical Care Medicine, Maywood, Illinois.


1. Agrawal, Abhinav et al. “Robotic bronchoscopy for pulmonary lesions: a review of existing technologies and clinical data.” Journal of thoracic disease vol. 12,6 (2020): 3279-3286. doi:10.21037/jtd.2020.03.35

2. Ali MS, Sethi J, Taneja A, Musani A, Maldonado F. Computed Tomography Bronchus Sign and the Diagnostic Yield of Guided Bronchoscopy for Peripheral Pulmonary Lesions. A Systematic Review and Meta-Analysis. Ann Am Thorac Soc. 2018 Aug;15(8):978-987. doi: 10.1513/AnnalsATS.201711-856OC. PMID: 29877715.

3. Argento AC, Puchalski J. Convex probe EBUS for centrally located parenchymal lesions without a bronchus sign. Respir Med. 2016 Jul;116:55-8. doi: 10.1016/j.rmed.2016.04.012. Epub 2016 Apr 29. PMID: 27296821.

4. Chen A, Chenna P, Loiselle A, Massoni J, Mayse M, Misselhorn D. Radial probe endobronchial ultrasound for peripheral pulmonary lesions. A 5-year institutional experience. Ann Am Thorac Soc. 2014 May;11(4):578-82. doi: 10.1513/AnnalsATS.201311-384OC. PMID: 24635641.