Best practices in shape-sensing robotic bronchoscopy with mobile cone beam computed tomography guidance: how I do it

Introduction

In the past few years, the advent of robotic bronchoscopy and cone-beam computed tomography (CBCT) guidance has revolutionized the world of bronchoscopy. After decades of frustration due to our inability to consistently obtain diagnostic samples from small peripheral lung lesions, we now have the tools to “reach” (robotic bronchoscopy) and “visualize” (CBCT) these peripheral lung lesions during bronchoscopy. The correct use of these technologies allows us to achieve and confirm tool-in-lesion (TIL) before obtaining samples, and to determine the relationship between our tools and any vital structure. This has resulted in improved diagnostic yield and safety outcomes (1-8). Just like with any new technology, there is a learning curve needed to master the combination of these two techniques in clinical practice. We have adopted these technologies at their earliest stages (research and development) and have since developed a vast amount of experience with them. In this manuscript, we will share some key aspects of our practice that we believe will lead to a more efficient, accurate, and safe procedure.

Initial set up: the Casal’s “A, B, C, D”

We have developed the following step-by-step set up plan that encompasses what needs to be done from the time the patient gets on the bed until the robot is docked, and we are ready to start our procedure. Working in a systematic fashion will help you reduce errors or omissions and will maximize your efficiency and procedural outcomes.

“A”: prevention of atelectasis, artificial airway, airway exam

A strategy to prevent “atelectasis” is key, should be discussed prior to each case with the anesthesia team, and should be tailored to each patient. We will go over this topic in detail further down in this manuscript.

The selection of the “artificial airway” type is also quite relevant and should be discussed a priori with the anesthesia team. While a small endotracheal tube (ETT)—7 or 8—is enough for the purpose of the robotic bronchoscopy (robotic catheters are typically of a small caliber), if the intention of the bronchoscopist is to perform mediastinal staging with endobronchial ultrasound through the same ETT (in addition to the robotic bronchoscopy), a larger ETT—8.5 or 9—will be required for adequate ventilation. Alternatively, robotic bronchoscopy can be performed with a small ETT, which can then be replaced with a laryngeal mask airway (LMA) if mediastinal staging is indicated. The latter approach may be preferable if mediastinal staging involves the sampling of upper paratracheal lymph nodes, and it is our preferred approach.

A bronchoscopic “airway exam” should be performed to suction any secretions and make sure there are no endobronchial lesions that could be missed or inadvertently run into by the robotic bronchoscope. The suction power of most robotic bronchoscopy catheters is weak, and the ability to clean the robotic catheter’s camera is suboptimal. Hence, cleaning up the airways from secretions beforehand to provide a clear view is of the utmost importance for accurate navigation. Having said so, we should avoid excessive suctioning which can be counterproductive by inducing more atelectasis.

“B”: securing patient/lines/cables to the bed, setting bed height

Once the patient has been intubated and the airway has been secured, before bringing in the two-/three-dimensional (2D/3D) C-arm, the patient should be properly positioned and secured to the bed, and the bed height should be adjusted. With regards to the patient’s positioning, whether in supine or lateral decubitus (which we will describe later), the patient should be centered in the bed, and his/her head should reach the top of the bed. Since the C-arm will be rotating 270 degrees around the patient to acquire 3D images, the patient’s arms, and all lines and cables should be tucked-in so they will not be caught by the rotating C-arm. Figure 1 displays the patient and robot set up. It is our practice to utilize a strap around the patient which we position at mid to lower chest level (always over the rib cage, to avoid compressing the abdomen and limiting ventilation) (Figure 1A). This strap secures the patient and both their arms, and any cable or line, preventing these from falling into the orbit of the C-arm. In all our cases the patient lies on a beanbag, which can be used for lateral decubitus positioning if needed.

Figure 1 Patient and robot positioning. (A) Patient strap should be placed over the ribs. (B) Robot instrument arm should be kept at its lowest position (yellow arrow) when adjusting bed’s height. (C) Swivel connector should be 1–2 inches higher that robot docking spar. (D) Robot instrument arm is kept at 45–60-degree angle to avoid collision with C-arm.

The bed should be leveled (flat position), and its height should be adjusted prior to the isocentering of the C-arm (step B should always precede step C), to guarantee that docking the robot will be feasible and smooth. When bed height is not adjusted and isocentering of the C-arm is performed first, then the bed will need to be elevated to reach the robot for docking, and isocentering and collision check will need to be performed once more, prolonging the procedure. To set the “bed height” properly, bring the robot instrument arm to its lowest position, and bring it next to the patient’s head. Then raise the bed until the swivel connector on the ETT is at least 1–2 inches above the docking spar on the instrument (with the instrument arm at its lowest point) (Figure 1B,1C). Having the swivel connector a few inches above the docking spar is key. If the swivel connector is at the same level of the docking spar instead, and the robotic instrument needs to be inclined to avoid collision with the C-arm (common scenario in upper lobe lesions), when tilting the ETT to align the swivel connector with the docking spar, the swivel connector will end up at a lower level than the docking spar, and you will need to raise the bed (because the instrument arm cannot be lowered any further). If that is the case, then you will need to perform the isocentering and collision check a second time, causing a delay.

“C”: C-arm positioning, isocentering, and collision check

Once the bed height is set, the patient is secured to the bed, and all cables and lines are tucked in, the C-arm can be positioned. Depending on the bronchoscopy suite layout, the C-arm can come from the left or right side of the bed (our bronchoscopy suite layout is depicted in Figure 2). Most 2D/3D C arms (mobile-CBCT) will only reconstruct 3D (CT) images of an area of approximately 16–19 cm by 16–19 cm. Thus, the C-arm should be “isocentered” over the region of interest where the target is located. The isocentering should be performed both in frontal and lateral projections. Since most targets will not be visible in 2D fluoroscopy images, interpreting the pre-procedural chest CT scans and predicting where the target would be in frontal and lateral X-ray images is key. Imaging software available in some institutions provides a function named “localizer mode”. To use this function, both the pre-procedural chest CT image (in any axis) as well as the CT scout should be opened. The target should be identified on the CT images, and the “localizer mode” should be activated (right click, then select “localizer mode”). When positioning the cursor over the target lesion on the CT image, the software will show a crosshair both in frontal and lateral views of the CT scout at the exact position of the target (Figure 3). This software function can be of great help, increasing accuracy and expediting isocentering process. We first perform the isocentering on the frontal projection by moving the C-arm on its wheels towards the head or feet of the patient, and towards the right or left. Isocentering in the lateral projection is done next by increasing the height of the C-arm for anterior lesions or decreasing the height of the C-arm for posterior lesions. Keep in mind that we are adjusting the height of the C-arm and not the height of the bed, which was adjusted to facilitate docking during our step “B”. To minimize radiation, we first utilize the laser beam of the C-arm projected on the patient’s body for a “rough” isocentering (without activating fluoroscopy) (Figure 4A,4B). Once we are over the region of interest, we activate the X-rays and corroborate the correct isocentering position with 2D images. Most 2D/3D C-arms will project a crosshair over the 2D fluoroscopy images which represents the center of the area from which the 3D images will be obtained (Figure 4C,4D). While this crosshair does not need to be positioned exactly over the target, it needs to be relatively close because 3D images will only be reconstructed from a small area (approximately 16 cm by 16 cm). The more accurate the isocentering, the greater the image quality of the target in the reconstructed 3D images.

Figure 2 Bronchoscopy suite layout.

Figure 3 Utilizing “localizer mode”. The yellow arrow indicates the target lesion.

Figure 4 Isocentering the C-arm. (A,B) Isocentering using laser beam in frontal (A) and lateral (B) projections. (C,D) Fluoroscopy images with cross-hairs at the center of the region of interest in frontal (C) and lateral (D) projections.

After isocentering the C-arm over the region of interest, a “collision check” is performed, which consists of a spin of the C-arm without radiation, to ensure that neither the patient, the bed, or any lines will be struck by the C-arm. If there is a collision with the bed, it is the C-arm that needs to be adjusted -and not the bed. The C-arm will be moved vertically or to either side of the patient until there is no collision, even if it means we cannot achieve perfect isocentering (as long as the target is still encompassed in both projections).

If we are sampling multiple targets, unless they are close enough to each other, step “C” may need to be repeated after sampling one target and moving to the next one.

“D”: docking of robot

With the base of the robot typically positioned somewhere “north” of the head of the table (north of the C-arm), the robot arm is maneuvered to reach the swivel connector mounted on the ETT. We typically flex the joints of the robot arm to keep the arm as short as possible, leaving the robot base closer to the patient, and having more space for circulation behind the robot. The individual performing the docking will use one hand to bring the docking spar of the robot instrument closer to the swivel connector of the ETT, while firmly keeping the ETT/swivel connector with the other hand in its desired position. Once the docking spar is 2–3 mm away from the connector, the operator lets go of the ETT and lets magnetism attach these 2 pieces. It is imperative to check that there is correct alignment of connector and docking spar so that they will not detach from each other during the procedure, which would require a new registration. The robotic instrument arm is typically kept at a 45–60-degree angle (Figure 1D). The more cephalad the location of the C-arm (i.e., for apical lesions), the more inclined the instrument arm will be (smaller angle) to avoid a collision with the C-arm.

Preventing atelectasis and reducing target motion

The burden of the intraprocedural development of atelectasis during peripheral bronchoscopy has been well characterized by our group (9-11). Atelectasis can worsen the phenomenon of CT to body divergence, obscure targets on CBCT, or cause false-positive radial-probe endobronchial ultrasound images, negatively impacting the procedural outcomes of peripheral bronchoscopy (12). Not surprisingly, atelectasis is directly related to the time under general anesthesia and to body mass index (BMI) (10). In patients with targets located in zones that are at high risk for developing atelectasis, a preventive strategy is mandatory. We have previously delineated the zones at high risk for developing atelectasis (11). For simplicity, if we trace a horizontal line at the most anterior edge of the corresponding vertebral body in axial views of the chest CT, any lesion located posterior to this line (which we call the atelectasis “equator”) will be at a high risk of being obscured by atelectasis (Figure 5).

Figure 5 The atelectasis “equator”. The yellow arrows indicate the dependent areas of the lungs at high risk for atelectasis.

Both ventilatory and positional strategies to prevent atelectasis during bronchoscopy have been described (13-17). A multicenter randomized controlled study utilizing a ventilatory strategy to prevent atelectasis (VESPA) substantially decreased the rate of atelectasis from 84% in the control group to 29% (16). VESPA is performed with an ETT, tidal volume (VT) of 6–8 ml/kg of ideal body weight (IBW), FiO₂ <100%—titrated as low as possible to maintain an oxygen saturation of >94%—and positive end-expiratory pressure (PEEP) of 8–10 cmH₂O. In addition, VESPA includes recruitment maneuver immediately after intubation with 10 consecutive breaths at a plateau pressure of 40 cmH₂O with PEEP of 20 cmH₂O in pressure control mode. Retrospective reports of local experiences with other ventilatory strategies (with VT and PEEP twice as high as those in VESPA, and with prolonged breath-holds) also reduced atelectasis rate to 20–30% range but caused hemodynamic instability in up to 70% of the patients (14,15). Moreover, these strategies utilized a very high VT, which is known to increase target motion, and thus should be avoided. All ventilatory strategies failed to eradicate atelectasis completely, particularly in patients with a higher BMI (>35 kg/m²). Hence, our group developed a positional strategy to prevent atelectasis which would require no ventilatory adjustments and which has been shown to completely prevent atelectasis, the “lateral decubitus strategy” (LADS) (17). Results from our randomized controlled trial of LADS vs. VESPA (NCT 05714033) will soon be available. Based on the data from our prior studies, we have developed an algorithm that allows us to tailor the strategy to prevent atelectasis to each individual patient (Figure 6). Of note, another positional strategy to prevent atelectasis, “prone positioning”, has been described in a small case series (18,19). Given the increased complexity when compared to LADS, and the excellent outcomes with LADS in NCT 05714033, we have not adopted this technique yet.

Figure 6 Selecting the appropriate strategy to prevent atelectasis. BMI, body mass index; LADS, lateral decubitus strategy; RV, residual volume; VESPA, ventilatory strategy to prevent atelectasis.

Positioning the patient in lateral decubitus occurs during step “B”. The patient is turned to the lateral decubitus position with the target lesion side up, similar to a pleuroscopy procedure, as we previously described (16). A bean bag and foam straps are used to hold patients in the lateral decubitus (a bean bag is in place for all our robotic bronchoscopies). The dependent arm is flexed at 90 degrees, and the nondependent arm is padded in neutral position to the patient’s side (Figure 7). We typically utilize a towel to tuck in the dependent arm by placing one end of the towel under the head of the patient, then wrapping it around the patient’s wrist, and placing the other end of the towel again under the head of the patient, using the weight of the head to maintain the arm in this position. Bed height needs to be adjusted, just like we do in supine position. Isocentering and collision check are more challenging in this position, in particular for patients with lesions that are located laterally. With the patient in lateral decubitus, these lateral lesions are high on the bed, and if perfectly isocentered, the C-arm may collide with the bottom of the bed. Adjusting the position of the patient (bringing the spine of the patient towards the C-arm) may be needed, and an imperfect isocentering may have to be tolerated, provided the lesion is visible in both projections. If isocentering cannot be achieved in LADS, a “semi-lateral” decubitus with a wedge pillow elevating the target lung can be performed. This may improve isocentering and still prevent atelectasis from obscuring the target to a certain degree. Docking in LADS is performed with the robot at the head of the table, and the ETT usually needs to be bent towards the ceiling to align with the robotic instrument arm. Other than adjusting the main carina during the phase of registration (it will need to be rotated 90 degrees), the rest of the robotic bronchoscopy procedure is the same as in supine position.

Figure 7 Lateral decubitus positioning. This image is published with the patient’s consent.

If atelectasis occurs in supine position (with or without VESPA), the patient should be rotated to LADS, which has been demonstrated to be effective in eliminating established atelectasis. We typically place them in lateral decubitus and follow that with a recruitment maneuver like the one described in VESPA.

In addition to preventing atelectasis, reducing target motion is key for those targets located near the diaphragm (lower lobes, right middle lobe, or lingula). These lesions can move up to 2.5 cm with diaphragmatic excursion, and when small, they can make their sampling extremely challenging (20). Our strategy is to minimize VTs. This is the reason why when we designed VESPA, we only utilized 6–8 mL/kg of IBW. In cases with small lesions within 2–3 cm from the diaphragm, we may lower this even further (4–5 mL/kg of IBW). Of note, it is the PEEP that prevents atelectasis and not VT. In addition to minimizing VT, for these same lesions, we get as close as we can with the robotic catheter, leaving the smallest gap between the catheter and the lesion, so they will be more likely to move in unison. With the above strategy, ventilatory pause is rarely required.

Planning, registration, navigation, confirmation, and tissue acquisition

Whenever possible, we perform the “planning” the day before the procedure, which gives us more time to solve any unforeseeable technical problem preventing us from downloading the patient’s scan. The planning phase is relatively simple, since most of the work is performed by the software. It is still important to review the computer-generated pathway, since adjustments may need to be made. When there is not an airway leading straight to the target (bronchus sign), and there are a few airways that are close enough, we carefully select the one that will avoid going through vessels or any other vital structure. We also select the exit angle based on the structure immediately beyond the lesion. If the distal border of the lesion is against the pleura/fissure or any vital structure, we choose an angle that will go tangentially and not directly into these structures in case the needle traverses the entire lesion. In cases when we need to exit the airway and traverse the lung parenchyma (the majority of our cases), we also favor pathways that require the needle to exit the airway at a steep angle, since it allows for better penetration of the needle through the airway wall (more likely to exit the airway than to follow the airway, which is not leading to the lesion). It is important to create at least 2 pathways per lesion, so we always have an alternative path in case an airway cannot be cannulated, or it is not clearly visible.

With regards to “registration” and “navigation”, when “driving” the robot, we simultaneously utilize both hands, drive at a steady speed, keep the catheter centered in the airway lumen, and avoid hesitation (moving back and forward, particularly during the registration). After confirmation of the main carina -which should be adjusted in orientation and depth- we start the “Y” registration ipsilaterally to our target tumor. This way we finish the “Y” registration on the opposite site, register the contralateral lobes, move on to the ipsilateral unaffected lobe, and end in the target lobe, minimizing time. If possible, when we register the lobe where the target is located, we follow the planned pathway, and when we accept the registration, we find ourselves next to the virtual target. This can, of course, only be done when the pathways are simple, thus only in a minority of cases. Once registration is over, we drive the robot to the target lobe, and make sure the registration is proper before accepting it. Though not yet our current practice, partial selective registration has been described to improve registration accuracy and decrease registration and navigation times, but larger prospective studies are needed to corroborate these findings (21,22).

It is our preference to “navigate” to our target utilizing the “preview path” (unlocked). We advance this preview pathway a couple of carinas and mimic that movement advancing then the robot, repeating these steps until we reach the exit point. At this moment we exit the preview path and locate the virtual target. We utilize this function because we have found that “live navigation” tends to hesitate and does not always advance through the airways with the robot, creating distraction, and sometimes making us lose our orientation. Once we have reached the virtual target, it is time to scan the patient with the mobile CBCT so that we can correct any degree of CT to body divergence and update the actual target location (23). When the target is against a vital structure (i.e., pleural-based target), to avoid running into this structure, we “park” the robot at least 2 cm proximal to the virtual target and we perform our first CBCT spin. We do this because we may be much closer to the actual target than it seems, or even past the target, depending on the degree of CT to body divergence. We can always get closer to it once we have updated the target based on our first m-CBCT spin.

Once we have navigated and reached our virtual target, we perform a first CBCT spin to correct for CT to body divergence. This information is sent back to the robot so we can utilize the “integration” function which consists of identifying the tip of the robotic catheter and the target in the intraprocedural CBCT and letting the robotic platform update the virtual target location so that it will now match the actual target. We perform this first scan without deploying the needle—particularly for smaller lesions—since we have reported previously a low rate of TIL when using the center of the original virtual target (1,4). After updating the target, we then deploy the needle adjusting its length to reach the center of the target, and we place a “biopsy marker” on the robotic platform. The second CBCT spin is performed for “confirmation”, and we look for TIL in all three CT axis. If this was not achieved, based on our interpretation of this new CBCT scan, we retract the needle and we make final adjustments with the robot utilizing the “biopsy marker” as the starting point, and following the compass around the target which gives us our orientation. Once the new desired catheter angulation is reached, we place a new “biopsy marker”, re-deploy the needle, and perform a new CBCT spin. In small lesions close to the diaphragm –subject to more movement- this procedure may need to be repeated a few times. It is imperative to not attempt to take samples unless TIL has been achieved.

We tailor our “tissue acquisition” strategy to each patient, depending on factors such as the presence of rapid on-site examination (ROSE), the lesion characteristics (ground glass vs. solid, size), the pre-test probability of malignancy, the need for molecular testing, the risk of pneumothorax (i.e., pleural based lesions), and the risk of bleeding (bleeding diathesis, proximity to a large vessel). We are fortunate enough to almost invariably have access to ROSE, but if this was not the case, we would then obtain both transbronchial needle aspirations (TBNA) and cryobiopsies [and bronchoalveolar lavage (BAL) if we suspected an infectious process]. In general, for solid lesions, we start with TBNA with 21 G needles (we actually utilize 2 of these so we can simultaneously take a new sample as the first sample slide is being prepared), obtain three samples, and send these slides to ROSE. If malignancy is confirmed on ROSE, and the slides look cellular enough for molecular testing, in addition to our first 3 slides, we would perform another 2–3 TBNAs and directly place the material in the cell-block container. If malignancy is confirmed but samples have poor cellularity, malignancy is suspected but cannot be confirmed, samples show benign process (i.e., granulomas), or they are simply non-diagnostic (i.e., blood, bronchial cells, macrophages), after no more than 4–5 TBNAs we move on to obtaining biopsies with a 1.1 mm cryoprobe (Erbe, Tuebingen, Germany) (24-26). We generally obtain a new CBCT spin to confirm TIL with the cryoprobe (on its first pass), since it will not always follow the same trajectory of the needle. This is relevant from the diagnostic standpoint, but also from the safety aspect if we are sampling lesions next to the pleura or a vital structure. If we are dealing with a small (<2 cm) ground glass opacity, with suspicion for a well-differentiated adenocarcinoma, we directly obtain biopsies with cryoprobe, because it is difficult in general to establish this diagnosis with cytology samples. In addition, attempting TBNA first may cause bleeding that can obscure the ground glass opacity, making it more challenging to obtain TIL later with the cryoprobe. When an infection is suspected based on ROSE, we obtain TBNA for culture as well as BAL through the robotic catheter (when feasible) or via bronchoscope.

Conclusions

Shape-sensing robotic bronchoscopy with mobile CBCT guidance results in highly promising diagnostic outcomes. Given its more favorable safety profile and the ability to also stage the mediastinum with endobronchial ultrasound when indicated, we envision that the combination of these technologies will likely replace percutaneous CT-guided biopsies of peripheral lung nodules in the years to come. Hands-on training is paramount before adopting this technique. We are honored to be able to share some guidance based on our vast and successful research and clinical experience.

Acknowledgments

None.

Footnote

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1-2507/prf

Funding: None.

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1-2507/coif). R.F.C. serves as an unpaid editorial board member of Journal of Thoracic Disease from April 2024 to June 2026. R.F.C. has received research grants from Siemens, Intuitive, and Johnson and Johnson. He is a paid consultant for Intuitive, Johnson and Johnson, and Canon. The other author has no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bashour SI, Khan A, Song J, et al. Improving Shape-Sensing Robotic-Assisted Bronchoscopy Outcomes with Mobile Cone-Beam Computed Tomography Guidance. Diagnostics (Basel) 2024;14:1955. [Crossref] [PubMed]
2. Fernandez-Bussy S, Yu Lee-Mateus A, Barrios-Ruiz A, et al. Diagnostic performance of shape-sensing robotic-assisted bronchoscopy for pleural-based and fissure-based pulmonary lesions. Thorax 2025;80:150-8. [Crossref] [PubMed]
3. Fernandez-Bussy S, Valdes-Camacho S, Barrios-Ruiz A, et al. Streamlining Lung Cancer Diagnosis: One Procedure for Multi-Site Biopsy Using Shape-Sensing Robotic-Assisted Bronchoscopy. Respiration 2025;104:930-9. [Crossref] [PubMed]
4. Husta BC, Menon A, Bergemann R, et al. The incremental contribution of mobile cone-beam computed tomography to the tool-lesion relationship during shape-sensing robotic-assisted bronchoscopy. ERJ Open Res 2024;10:00993-2023. [Crossref] [PubMed]
5. Reisenauer J, Duke JD, Kern R, et al. Combining Shape-Sensing Robotic Bronchoscopy With Mobile Three-Dimensional Imaging to Verify Tool-in-Lesion and Overcome Divergence: A Pilot Study. Mayo Clin Proc Innov Qual Outcomes 2022;6:177-85. [Crossref] [PubMed]
6. Boster JM, Goertzen SM, Armas Villalba AJ, et al. Cone-Beam Computed Tomography-guided, Robotic-assisted Bronchoscopy: A Novel Transcavitary Lung Biopsy Approach. Ann Am Thorac Soc 2025;22:938-41. [Crossref] [PubMed]
7. Abdelghani R, Espinoza D, Uribe JP, et al. Cone-beam computed tomography-guided shape-sensing robotic bronchoscopy vs. electromagnetic navigation bronchoscopy for pulmonary nodules. J Thorac Dis 2024;16:5529-38. [Crossref] [PubMed]
8. Abia-Trujillo D, Folch EE, Yu Lee-Mateus A, et al. Mobile cone-beam computed tomography complementing shape-sensing robotic-assisted bronchoscopy in the small pulmonary nodule sampling: A multicentre experience. Respirology 2024;29:324-32. [Crossref] [PubMed]
9. Casal RF, Sarkiss M, Jones AK, et al. Cone beam computed tomography-guided thin/ultrathin bronchoscopy for diagnosis of peripheral lung nodules: a prospective pilot study. J Thorac Dis 2018;10:6950-9. [Crossref] [PubMed]
10. Sagar AS, Sabath BF, Eapen GA, et al. Incidence and Location of Atelectasis Developed During Bronchoscopy Under General Anesthesia: The I-LOCATE Trial. Chest 2020;158:2658-66. [Crossref] [PubMed]
11. Khan A, Bashour S, Sabath B, et al. Severity of Atelectasis during Bronchoscopy: Descriptions of a New Grading System (AtelectasisSeverity Scoring System-"ASSESS") and At-Risk-Lung Zones. Diagnostics (Basel) 2024;14:197. [Crossref] [PubMed]
12. Khan A, Bashour SI, Casal RF. Preventing atelectasis during bronchoscopy under general anesthesia. J Thorac Dis 2023;15:3443-52. [Crossref] [PubMed]
13. Pritchett MA, Lau K, Skibo S, et al. Anesthesia considerations to reduce motion and atelectasis during advanced guided bronchoscopy. BMC Pulm Med 2021;21:240. [Crossref] [PubMed]
14. Bhadra K, Setser RM, Condra W, et al. Lung Navigation Ventilation Protocol to Optimize Biopsy of Peripheral Lung Lesions. J Bronchology Interv Pulmonol 2022;29:7-17. [Crossref] [PubMed]
15. Bhadra K, Baleeiro C, Patel S, et al. High Tidal Volume, High Positive End Expiratory Pressure and Apneic Breath Hold Strategies (Lung Navigation Ventilation Protocol) With Cone Beam Computed Tomography Bronchoscopic Biopsy of Peripheral Lung Lesions: Results in 100 Patients. J Bronchology Interv Pulmonol 2024;31:105-16. [Crossref] [PubMed]
16. Salahuddin M, Sarkiss M, Sagar AS, et al. Ventilatory Strategy to Prevent Atelectasis During Bronchoscopy Under General Anesthesia: A Multicenter Randomized Controlled Trial (Ventilatory Strategy to Prevent Atelectasis -VESPA- Trial). Chest 2022;162:1393-401. [Crossref] [PubMed]
17. Lin J, Sabath BF, Sarkiss M, et al. Lateral Decubitus Positioning for Mobile CT-guided Robotic Bronchoscopy: A Novel Technique to Prevent Atelectasis. J Bronchology Interv Pulmonol 2022;29:220-3. [Crossref] [PubMed]
18. Bhadra K, Condra W, Setser RM. Out of the Box Thinking: Prone Bronchoscopy to Reduce Atelectasis. J Bronchology Interv Pulmonol 2022;29:e57-60. [Crossref] [PubMed]
19. Alraiyes AH, Johnson C, Madjer N, et al. Prone positioning in cone beam CT-guided robotic bronchoscopy case series: a strategy to minimize atelectasis and improve access to posteromedial lower lobe nodules. J Thorac Dis 2025;17:9275-86. [Crossref] [PubMed]
20. Chen A, Pastis N, Furukawa B, et al. The effect of respiratory motion on pulmonary nodule location during electromagnetic navigation bronchoscopy. Chest 2015;147:1275-81. [Crossref] [PubMed]
21. Chrissian AA, Khosa J, Duchman B, et al. Airway Registration Approach and Other Determinants of Procedure Efficiency During Shape-Sensing Robotic-Assisted Bronchoscopy. J Bronchology Interv Pulmonol 2025;32:e01028. [Crossref] [PubMed]
22. Duchman B, Khosa JK, Xu L, et al. Investigating Partial Registration for Optimizing Robotic-Assisted Bronchoscopy Evaluations of Peripheral Lung Nodules (I-PROBE). Chest Pulmonary 2025;100229.
23. Pritchett MA, Bhadra K, Calcutt M, et al. Virtual or reality: divergence between preprocedural computed tomography scans and lung anatomy during guided bronchoscopy. J Thorac Dis 2020;12:1595-611. Erratum in: J Thorac Dis 2020;12:4593-5. [Crossref] [PubMed]
24. Odeh T, Pitts L, Reid M, et al. Robotic-Assisted Bronchoscopy-Guided Transbronchial Cryobiopsy Versus Transbronchial Needle Aspiration in Peripheral Pulmonary Lesions. J Bronchology Interv Pulmonol 2025;32:e1014. [Crossref] [PubMed]
25. Abia-Trujillo D, Funes-Ferrada R, Yu Lee-Mateus A, et al. Cryobiopsy versus fine-needle aspiration for shape-sensing robotic-assisted sampling of small lung nodules. Lung Cancer 2024;196:107967. [Crossref] [PubMed]
26. Oberg CL, Lau RP, Folch EE, et al. Novel Robotic-Assisted Cryobiopsy for Peripheral Pulmonary Lesions. Lung 2022;200:737-45. [Crossref] [PubMed]