Robotic bronchoscopy 2022

Over the last several years, hundreds of millions of dollars have been spent on robotic bronchoscopy systems in the United States. The release of robotic scopes was made to great fanfare, translating into the market being infiltrated with these systems. With base costs in the hundreds of thousands of dollars, robotic bronchoscope systems are easily the most expensive singular capital investment in the bronchoscopy suite. I frequently get asked questions from those who have not yet made that purchase: “Should I buy a robot?” “How could I justify a new robot purchase to my hospital?” “Is the hype real?” These are complex questions to answer. Before one can answer, I think it’s best to look back on the last 2 decades of bronchoscopy for peripheral lung nodules to get a better understanding of the value proposition robotic bronchoscopes may offer.

Guided bronchoscopy for lung nodules has significantly evolved over the past 2 decades, shifting diagnostic procedures from interventional radiologist to the pulmonologist. Some of these advances were based in redesigns of the bronchoscope (ultrathin bronchoscopy) or application of technology to the bronchoscope (radial EBUS, virtual bronchoscopy); but, these were not broadly applicable to the pulmonology community at large. It was not until the development of electromagnetic navigational bronchoscopy (ENB) that widespread adoption of bronchoscopy for lung nodules occurred. By and large, ENB fueled a rapid expansion of nodule bronchoscopy, mainly due to its ease of use and novel approach. Initial studies of ENB had impressive results; however, studies were criticized for having small numbers, inadequate follow-up, spurious definitions of yield, and that they were being done at highly specialized centers. The NAVIGATE trial was launched to address these criticisms among “real world” conditions. Sponsored by Medtronic, it studied ENB (superDimension platform, v6.0 or higher) across 29 academic and medical centers in the United States, enrolling over 1,000 patients (Folch EE, et al. J Thorac Oncol. 2019;14[3]:445-58. Epub 2018 Nov 23), and reported a diagnostic yield of 73%.

This led to a drive to improve upon yield, resulting in development of new technologies specifically designed to address some of the factors thought associated with diminished yield, and, out of this, robotic bronchoscopy was born. These factors included CT scan-body registration divergence, deflection of the extended working channel (EWC) by rigid biopsy tools, and inability to accurately “aim” the EWC-biopsy tool at the nodule; these were especially problematic in nodules not associated with airways. Robotic scopes were specifically designed to reach into the peripheral lung airways similar to an EWC, but with better structural integrity and steerability. This tip integrity would resist tool-related displacement, and steerability would allow for improved targeting of nodules during the biopsy.

There are two robots approved by the FDA at the time of this writing (Auris Monarch, Intuitive Ion), with a third awaiting FDA clearance (Noah Galaxy). In general, though the engineering of the robotic scopes to improve structural tip integrity are similar, the approach to navigation and targeting vary significantly. The Monarch platform uses electromagnetic guidance, similar to other traditional ENB platforms. The Ion platform does not use ENB; instead, it uses fiberoptic shape sensing technology, which analyzes the shape and orientation of the scope to provide location information. There are potential advantages to shape sensing, the most notable being the absence of electromagnetics; this allows for use of fluoroscopy during the procedure, which otherwise would have interfered with ENB-based navigation. There are other subtle differences between the two robots. The Monarch uses a scope-in-scope design, with a robotic scope contained within a robotic sheath; the Ion uses a single robotic scope. The Ion scope diameter is 3.5 mm, whereas the Monarch diameter is 4.4 mm; this may be a potential advantage when having to navigate through smaller airways.

So, which robot is better suited to reach peripheral nodules more consistently and accurately? I get asked this question a lot, since I have both platforms at my institution. But, answering with my own opinion based on my institution’s anecdotal experience would be irresponsible. I’m more of a “what does the data show?” person. Luckily, we do have clinical trials in both robot technologies. It should be noted here that there will likely never be a head-to-head randomized trial, so evaluating published studies with each platform is going to be the best method we have for comparison going forward, albeit an imperfect one. It should also be noted that many of the early robotic bronchoscopy trials have to be looked at with caution, as yield definitions tended not to be conservative and/or the follow-up of non-malignant was not robust. With that in mind, let’s review representative high-quality studies for each platform.

The best study to date using the Ion platform came out of Memorial Sloan Kettering Cancer Center (Kalchiem-Dekel O, et al. Chest. 2022;161[2];572-82). This single-site study reported on 159 nodule biopsies, with the primary outcome being diagnostic yield. The patients had 1 year of follow-up, and the definition of yield was conservative. The average lesion size was 18 mm, and nodule locations and characteristics were representative of real-world conditions. Overall diagnostic yield was 81.7%; however, it dropped to under 70% for nodules under 20 mm in size.

The largest study to date using the Monarch platform was also a single center study, this from the University of Chicago (Agrawal, et al. Ann Thorac Surg. 2022 Jan 17;S0003-4975(22)00042-X. Online ahead of print). This study included 124 nodules with at least 12 months of follow-up; diagnostic yield definition was conservative. Median nodule size was 20.5 mm, with distribution and characteristics representative of real-world conditions. Overall accuracy was 77%, and, similar to the Ion study, dropped to under 70% when nodule size was smaller than 20 mm.

Overall, both robot studies seemed to show a modest improvement in diagnostic yield when compared with ENB, and their outcomes were overall similar. It is important to remember that these were studies of each center’s first experiences with early versions of each technology; over time, the technology will continue to improve, as will operator skill and experience, and with that, perhaps improvements in yield will be seen, as well.

Interestingly, both studies evaluated target localization using radial EBUS (rEBUS), which also allowed for airway-nodule relationships to be reported. In Kalchiem-Dekel’s study, 85% of cases used rEBUS to determine localization, and, of these, 91.2% of cases showed accurate localization. In Agrawal’s study, rEBUS was used in all cases with a reported localization of 94%. In both, yield did not seem to be affected by airway-nodule relationships, perhaps explained by more robust tip control of the robotic scope. However, localization did not equate to yield in all cases, which brings up a very important question: Can the yield of robotic bronchoscopy be further improved with better real-time on-board imaging, such as CBCT scanning or C-arm based tomography? Currently, there is a study using 3D technology (Cios 3D Mobile Spin) in conjunction with the Ion platform to evaluate this.

So, let’s circle back to where we started. I think if you look at the totality of the data, it is clear that the robotic platforms currently offer a modest improvement in diagnostic yield over traditional ENB, with individual performances that are somewhat equivalent despite differences in design and operation. But does this improvement in yield justify the cost? Individual hospitals will have to make that decision. The capital cost and per-use price of the scope is significant, which has to be balanced against each center’s current performance with non-robotic bronchoscopy.

To date, there have been over 25,000 robotic procedures performed in the United States, so enthusiasm across diverse centers is being maintained. Whether this enthusiasm is driven by yield or novelty, or both, I’m not sure. With other nonrobotic platforms having reached, or soon to reach, the market, this is a good time to be in the business of bronchoscopy.

Dr. Cicenia is in the Section of Bronchoscopy at Cleveland Clinic’s Respiratory Institute, Cleveland, Ohio.

Over the last several years, hundreds of millions of dollars have been spent on robotic bronchoscopy systems in the United States. The release of robotic scopes was made to great fanfare, translating into the market being infiltrated with these systems. With base costs in the hundreds of thousands of dollars, robotic bronchoscope systems are easily the most expensive singular capital investment in the bronchoscopy suite. I frequently get asked questions from those who have not yet made that purchase: “Should I buy a robot?” “How could I justify a new robot purchase to my hospital?” “Is the hype real?” These are complex questions to answer. Before one can answer, I think it’s best to look back on the last 2 decades of bronchoscopy for peripheral lung nodules to get a better understanding of the value proposition robotic bronchoscopes may offer.

Guided bronchoscopy for lung nodules has significantly evolved over the past 2 decades, shifting diagnostic procedures from interventional radiologist to the pulmonologist. Some of these advances were based in redesigns of the bronchoscope (ultrathin bronchoscopy) or application of technology to the bronchoscope (radial EBUS, virtual bronchoscopy); but, these were not broadly applicable to the pulmonology community at large. It was not until the development of electromagnetic navigational bronchoscopy (ENB) that widespread adoption of bronchoscopy for lung nodules occurred. By and large, ENB fueled a rapid expansion of nodule bronchoscopy, mainly due to its ease of use and novel approach. Initial studies of ENB had impressive results; however, studies were criticized for having small numbers, inadequate follow-up, spurious definitions of yield, and that they were being done at highly specialized centers. The NAVIGATE trial was launched to address these criticisms among “real world” conditions. Sponsored by Medtronic, it studied ENB (superDimension platform, v6.0 or higher) across 29 academic and medical centers in the United States, enrolling over 1,000 patients (Folch EE, et al. J Thorac Oncol. 2019;14[3]:445-58. Epub 2018 Nov 23), and reported a diagnostic yield of 73%.

This led to a drive to improve upon yield, resulting in development of new technologies specifically designed to address some of the factors thought associated with diminished yield, and, out of this, robotic bronchoscopy was born. These factors included CT scan-body registration divergence, deflection of the extended working channel (EWC) by rigid biopsy tools, and inability to accurately “aim” the EWC-biopsy tool at the nodule; these were especially problematic in nodules not associated with airways. Robotic scopes were specifically designed to reach into the peripheral lung airways similar to an EWC, but with better structural integrity and steerability. This tip integrity would resist tool-related displacement, and steerability would allow for improved targeting of nodules during the biopsy.

There are two robots approved by the FDA at the time of this writing (Auris Monarch, Intuitive Ion), with a third awaiting FDA clearance (Noah Galaxy). In general, though the engineering of the robotic scopes to improve structural tip integrity are similar, the approach to navigation and targeting vary significantly. The Monarch platform uses electromagnetic guidance, similar to other traditional ENB platforms. The Ion platform does not use ENB; instead, it uses fiberoptic shape sensing technology, which analyzes the shape and orientation of the scope to provide location information. There are potential advantages to shape sensing, the most notable being the absence of electromagnetics; this allows for use of fluoroscopy during the procedure, which otherwise would have interfered with ENB-based navigation. There are other subtle differences between the two robots. The Monarch uses a scope-in-scope design, with a robotic scope contained within a robotic sheath; the Ion uses a single robotic scope. The Ion scope diameter is 3.5 mm, whereas the Monarch diameter is 4.4 mm; this may be a potential advantage when having to navigate through smaller airways.

So, which robot is better suited to reach peripheral nodules more consistently and accurately? I get asked this question a lot, since I have both platforms at my institution. But, answering with my own opinion based on my institution’s anecdotal experience would be irresponsible. I’m more of a “what does the data show?” person. Luckily, we do have clinical trials in both robot technologies. It should be noted here that there will likely never be a head-to-head randomized trial, so evaluating published studies with each platform is going to be the best method we have for comparison going forward, albeit an imperfect one. It should also be noted that many of the early robotic bronchoscopy trials have to be looked at with caution, as yield definitions tended not to be conservative and/or the follow-up of non-malignant was not robust. With that in mind, let’s review representative high-quality studies for each platform.

The best study to date using the Ion platform came out of Memorial Sloan Kettering Cancer Center (Kalchiem-Dekel O, et al. Chest. 2022;161[2];572-82). This single-site study reported on 159 nodule biopsies, with the primary outcome being diagnostic yield. The patients had 1 year of follow-up, and the definition of yield was conservative. The average lesion size was 18 mm, and nodule locations and characteristics were representative of real-world conditions. Overall diagnostic yield was 81.7%; however, it dropped to under 70% for nodules under 20 mm in size.

The largest study to date using the Monarch platform was also a single center study, this from the University of Chicago (Agrawal, et al. Ann Thorac Surg. 2022 Jan 17;S0003-4975(22)00042-X. Online ahead of print). This study included 124 nodules with at least 12 months of follow-up; diagnostic yield definition was conservative. Median nodule size was 20.5 mm, with distribution and characteristics representative of real-world conditions. Overall accuracy was 77%, and, similar to the Ion study, dropped to under 70% when nodule size was smaller than 20 mm.

Overall, both robot studies seemed to show a modest improvement in diagnostic yield when compared with ENB, and their outcomes were overall similar. It is important to remember that these were studies of each center’s first experiences with early versions of each technology; over time, the technology will continue to improve, as will operator skill and experience, and with that, perhaps improvements in yield will be seen, as well.

Interestingly, both studies evaluated target localization using radial EBUS (rEBUS), which also allowed for airway-nodule relationships to be reported. In Kalchiem-Dekel’s study, 85% of cases used rEBUS to determine localization, and, of these, 91.2% of cases showed accurate localization. In Agrawal’s study, rEBUS was used in all cases with a reported localization of 94%. In both, yield did not seem to be affected by airway-nodule relationships, perhaps explained by more robust tip control of the robotic scope. However, localization did not equate to yield in all cases, which brings up a very important question: Can the yield of robotic bronchoscopy be further improved with better real-time on-board imaging, such as CBCT scanning or C-arm based tomography? Currently, there is a study using 3D technology (Cios 3D Mobile Spin) in conjunction with the Ion platform to evaluate this.

So, let’s circle back to where we started. I think if you look at the totality of the data, it is clear that the robotic platforms currently offer a modest improvement in diagnostic yield over traditional ENB, with individual performances that are somewhat equivalent despite differences in design and operation. But does this improvement in yield justify the cost? Individual hospitals will have to make that decision. The capital cost and per-use price of the scope is significant, which has to be balanced against each center’s current performance with non-robotic bronchoscopy.

To date, there have been over 25,000 robotic procedures performed in the United States, so enthusiasm across diverse centers is being maintained. Whether this enthusiasm is driven by yield or novelty, or both, I’m not sure. With other nonrobotic platforms having reached, or soon to reach, the market, this is a good time to be in the business of bronchoscopy.

Dr. Cicenia is in the Section of Bronchoscopy at Cleveland Clinic’s Respiratory Institute, Cleveland, Ohio.

Over the last several years, hundreds of millions of dollars have been spent on robotic bronchoscopy systems in the United States. The release of robotic scopes was made to great fanfare, translating into the market being infiltrated with these systems. With base costs in the hundreds of thousands of dollars, robotic bronchoscope systems are easily the most expensive singular capital investment in the bronchoscopy suite. I frequently get asked questions from those who have not yet made that purchase: “Should I buy a robot?” “How could I justify a new robot purchase to my hospital?” “Is the hype real?” These are complex questions to answer. Before one can answer, I think it’s best to look back on the last 2 decades of bronchoscopy for peripheral lung nodules to get a better understanding of the value proposition robotic bronchoscopes may offer.

Guided bronchoscopy for lung nodules has significantly evolved over the past 2 decades, shifting diagnostic procedures from interventional radiologist to the pulmonologist. Some of these advances were based in redesigns of the bronchoscope (ultrathin bronchoscopy) or application of technology to the bronchoscope (radial EBUS, virtual bronchoscopy); but, these were not broadly applicable to the pulmonology community at large. It was not until the development of electromagnetic navigational bronchoscopy (ENB) that widespread adoption of bronchoscopy for lung nodules occurred. By and large, ENB fueled a rapid expansion of nodule bronchoscopy, mainly due to its ease of use and novel approach. Initial studies of ENB had impressive results; however, studies were criticized for having small numbers, inadequate follow-up, spurious definitions of yield, and that they were being done at highly specialized centers. The NAVIGATE trial was launched to address these criticisms among “real world” conditions. Sponsored by Medtronic, it studied ENB (superDimension platform, v6.0 or higher) across 29 academic and medical centers in the United States, enrolling over 1,000 patients (Folch EE, et al. J Thorac Oncol. 2019;14[3]:445-58. Epub 2018 Nov 23), and reported a diagnostic yield of 73%.

This led to a drive to improve upon yield, resulting in development of new technologies specifically designed to address some of the factors thought associated with diminished yield, and, out of this, robotic bronchoscopy was born. These factors included CT scan-body registration divergence, deflection of the extended working channel (EWC) by rigid biopsy tools, and inability to accurately “aim” the EWC-biopsy tool at the nodule; these were especially problematic in nodules not associated with airways. Robotic scopes were specifically designed to reach into the peripheral lung airways similar to an EWC, but with better structural integrity and steerability. This tip integrity would resist tool-related displacement, and steerability would allow for improved targeting of nodules during the biopsy.

There are two robots approved by the FDA at the time of this writing (Auris Monarch, Intuitive Ion), with a third awaiting FDA clearance (Noah Galaxy). In general, though the engineering of the robotic scopes to improve structural tip integrity are similar, the approach to navigation and targeting vary significantly. The Monarch platform uses electromagnetic guidance, similar to other traditional ENB platforms. The Ion platform does not use ENB; instead, it uses fiberoptic shape sensing technology, which analyzes the shape and orientation of the scope to provide location information. There are potential advantages to shape sensing, the most notable being the absence of electromagnetics; this allows for use of fluoroscopy during the procedure, which otherwise would have interfered with ENB-based navigation. There are other subtle differences between the two robots. The Monarch uses a scope-in-scope design, with a robotic scope contained within a robotic sheath; the Ion uses a single robotic scope. The Ion scope diameter is 3.5 mm, whereas the Monarch diameter is 4.4 mm; this may be a potential advantage when having to navigate through smaller airways.

So, which robot is better suited to reach peripheral nodules more consistently and accurately? I get asked this question a lot, since I have both platforms at my institution. But, answering with my own opinion based on my institution’s anecdotal experience would be irresponsible. I’m more of a “what does the data show?” person. Luckily, we do have clinical trials in both robot technologies. It should be noted here that there will likely never be a head-to-head randomized trial, so evaluating published studies with each platform is going to be the best method we have for comparison going forward, albeit an imperfect one. It should also be noted that many of the early robotic bronchoscopy trials have to be looked at with caution, as yield definitions tended not to be conservative and/or the follow-up of non-malignant was not robust. With that in mind, let’s review representative high-quality studies for each platform.

The best study to date using the Ion platform came out of Memorial Sloan Kettering Cancer Center (Kalchiem-Dekel O, et al. Chest. 2022;161[2];572-82). This single-site study reported on 159 nodule biopsies, with the primary outcome being diagnostic yield. The patients had 1 year of follow-up, and the definition of yield was conservative. The average lesion size was 18 mm, and nodule locations and characteristics were representative of real-world conditions. Overall diagnostic yield was 81.7%; however, it dropped to under 70% for nodules under 20 mm in size.

The largest study to date using the Monarch platform was also a single center study, this from the University of Chicago (Agrawal, et al. Ann Thorac Surg. 2022 Jan 17;S0003-4975(22)00042-X. Online ahead of print). This study included 124 nodules with at least 12 months of follow-up; diagnostic yield definition was conservative. Median nodule size was 20.5 mm, with distribution and characteristics representative of real-world conditions. Overall accuracy was 77%, and, similar to the Ion study, dropped to under 70% when nodule size was smaller than 20 mm.

Overall, both robot studies seemed to show a modest improvement in diagnostic yield when compared with ENB, and their outcomes were overall similar. It is important to remember that these were studies of each center’s first experiences with early versions of each technology; over time, the technology will continue to improve, as will operator skill and experience, and with that, perhaps improvements in yield will be seen, as well.

Interestingly, both studies evaluated target localization using radial EBUS (rEBUS), which also allowed for airway-nodule relationships to be reported. In Kalchiem-Dekel’s study, 85% of cases used rEBUS to determine localization, and, of these, 91.2% of cases showed accurate localization. In Agrawal’s study, rEBUS was used in all cases with a reported localization of 94%. In both, yield did not seem to be affected by airway-nodule relationships, perhaps explained by more robust tip control of the robotic scope. However, localization did not equate to yield in all cases, which brings up a very important question: Can the yield of robotic bronchoscopy be further improved with better real-time on-board imaging, such as CBCT scanning or C-arm based tomography? Currently, there is a study using 3D technology (Cios 3D Mobile Spin) in conjunction with the Ion platform to evaluate this.

So, let’s circle back to where we started. I think if you look at the totality of the data, it is clear that the robotic platforms currently offer a modest improvement in diagnostic yield over traditional ENB, with individual performances that are somewhat equivalent despite differences in design and operation. But does this improvement in yield justify the cost? Individual hospitals will have to make that decision. The capital cost and per-use price of the scope is significant, which has to be balanced against each center’s current performance with non-robotic bronchoscopy.

To date, there have been over 25,000 robotic procedures performed in the United States, so enthusiasm across diverse centers is being maintained. Whether this enthusiasm is driven by yield or novelty, or both, I’m not sure. With other nonrobotic platforms having reached, or soon to reach, the market, this is a good time to be in the business of bronchoscopy.

Dr. Cicenia is in the Section of Bronchoscopy at Cleveland Clinic’s Respiratory Institute, Cleveland, Ohio.