possible. Second, once this has been accomplished, one must ensure that the C-arm can complete its rotation around the patient and table given the location of the isocenter. If the isocenter is in somewhat of an extreme location, the C-arm may run into the patient or table as it attempts to rotate around the isocenter. In such a case, the patient may need to be moved such that the isocenter is closer to the center of the table, making it more likely that the C-arm will be able to clear its rotation around both sides of the table. Raising the patient’s arms above the head may also help ensure C-arm clearance around the chest, particularly in larger patients. This may have the added beneft of reducing X-ray interference from the arms, especially with posterior lung lesions that may lay in the same plane as the arms when at the patient’s side. If the arms are raised, care must be taken to minimize both the degree of manipulation ( exion/extension of joints) and the length of time during which the arms are manipulated in order to reduce the risk of injury.

The probability of intra-operative atelectasis is another matter of which to be aware. Atelectasis has been demonstrated in up to 90% of anesthetized patients undergoing various procedures [36, 37]. This can obscure the visualization of a given target by CBCT if the lesion becomes buried within collapsed lung parenchyma. Indeed, the authors recently published a prospective, observational study of peripheral bronchoscopy to determine the incidence of atelectasis developed during bronchoscopy as detected by radial probe EBUS [25]. The posterior segments of the upper lobes as well as the superior, lateral, and posterior segments of the lower lobes were evaluated. Of 57 patients examined, 89% developed atelectasis in at least one of these segments and 1 in 3 developed atelectasis in 6 of the 8 examined segments. To attempt to minimize such effects, different ventilatory interventions have been advocated including the use of higher tidal volumes (e.g., 8–12 mL/kg of ideal body weight), higher levels of positive end-expiratory pressure (e.g., 10–15 cm H2O), the use of an endotracheal tube rather than a laryngeal mask airway (in order to better preserve intrathoracic pressures), avoiding

hyperoxia, and the use of recruitment maneuvers [26, 35, 38–40]. A clinical trial is currently underway comparing two specifc ventilatory strategies and hopefully will provide higher quality data in this regard specifc to peripheral bronchoscopy[41]. Finally, placing the patient in a lateral decubitus position with the target side up can also be considered, particularly for very posterior lesions that are at high risk of being obscured by atelectasis.

Hardware

Fixed (i.e., non-portable) CBCT systems exist in various designs: oor-mounted, ceiling-mounted, biplane (both ceilingand oor-mounted components), and robotic (Fig. 25.4). Floor-mounted and biplane systems are attached to the oor at the head of the patient table so this may interfere with procedural work ow for bronchoscopy, specifcally. They also scan near the level of the head of the patient and so may be unable to capture images in the more caudal lung felds. Ceiling-­ mounted and robotic systems are a bit more versatile in their ability to be moved away from the patient and associated tools in use (this may be particularly useful when using electromagnetic navigation as the presence of the X-ray detector may interfere with the magnetic feld generated by the EMN equipment). Nevertheless, all of these systems can be used for bronchoscopy at the discretion of the physician.

Mobile C-arm-based systems have been designed as well and are newer on the market compared to conventional CBCT. The scan area provided by these devices is somewhat smaller than CBCT with a slightly lesser image quality but both are nevertheless adequate for the purposes of bronchoscopy and guiding the biopsy of lung lesions. They do offer certain advantages over usual CBCT, however. The footprint of these devices is only slightly larger than the typical C-arms currently used for uoroscopy, obviating any need to make adjustments to the setup of the room in order to incorporate cone-beam technology (as with fxed systems, discussed below). Moreover, they are fully mobile, mounted on

wheels. As such, they can be brought in and out of any bronchoscopy suite that presently usesuoroscopy. This can be particularly useful to pulmonary procedural programs that run more than one procedure room at the same time and may need to share equipment concurrently. Mobile C-arm-based systems are also lower cost compared to conventional CBCT.

Finally, a radiolucent patient table (most often a carbon-fber table) is required for CBCT, regardless of whether a fxed or mobile system is used. If a metallic or otherwise radio-opaque table is used, this will generate signifcant degrees of artifact as the X-rays are absorbed by the table [35].

Practical Considerations

Multiple personnel are often required for diagnostic bronchoscopy using cone-beam technology including: the bronchoscopist, anesthesia, bronchoscopy technicians to aid with the procedure, a radiology technician to manipulate the scanner, and often cytology staff as well. Members of the team need to coordinate their

efforts and adapt their roles to this new technology in order to obtain a better outcome. Aside from lead aprons and other personal protective equipment, CT rooms are equipped with a control room that is both impervious to radiation and also can serve as an area to review the acquired images and to plan subsequent steps as needed. In the case of bronchoscopy suites whose prior design may not have been mindful of the future addition of CT technology, portable radiation shields can be wheeled into the room to protect staff that must remain during the scan while other staff can temporarily exit the room.

Planning room setup and work ow is important as well. The main concept requiring careful attention is that interference with the C-arm does not occur by ancillary equipment. Anesthesia equipment often is sizeable but longer tubing and/or wires may allow for creative ways to keep the ventilator and other larger machines in an accommodating location (to that end, the authors use Velcro straps to keep these together and out of the way of the rotating C-arm). The bronchoscope and its tower/cart and any other necessary technology (e.g., navigation equipment) must be close enough to the head of the patient to allow

proper use yet located or angled far enough away to not hinder the C-arm. Video monitor(s) must also be situated in a location visible to the bronchoscopist. Finally, consideration must be given to how the bronchoscope (and any tools within its working channel) will be maintained in a secure and stationary position during the scan when the bronchoscopist has moved to a radiation-safe location. Any number of creative methods can be employed with some previously described and/or commercially available [42–44]. This is obviously not needed when robotic bronchoscopy is being utilized. A typical room setup used by the authors is illustrated in Fig. 25.5.

Radiation Dose

The total dose of radiation to which a patient is exposed is directly related to the number of CT acquisitions as well as the number of images captured within a given acquisition and the radiation dose per image. The total dose increases linearly with each acquisition [35].

When comparing exposure dosages between studies in the medical literature, one must be aware that these have been reported dissimilarly

at times, making comparisons diffcult. For example, the most appropriate metrics for radiation dose with CBCT are the air Kerma (Kar) but most studies report the effective dose (E). The problem with the latter is that E carries with it a conversion factor and this is specifc not only to the system being used but also is a factor of anatomic location. This has been dealt with extensively previously [42]. Nevertheless, the effective radiation dose with CBCT acquisitions of 248, 312, and 419 projection images has been translated into levels of 0.98 mSv, 1.33 mSv, and 3.32 mSv, respectively [45–47]. These levels of radiation are the same as or slightly above those associated with low-dose CT for lung cancer screening.

Limitations andChallenges

One of the main limitations in the use of cone-­ beam technology is that of access. An individual CBCT system costs in the millions of dollars. Moreover, outftting a bronchoscopy suite for CBCT will typically require a complete redesign of the room—if not creation of a new lab alto- gether—in order to allow suffcient space and

infrastructure, particularly for ceiling-mounted devices [48]. Space and budget limitations could make this prohibitive for many pulmonary services. An alternative, then, could be the aforementioned mobile CT systems. These provide adequate imaging quality and can be used in any current procedure suite that can accommodate a traditional C-arm—and at lower cost.

Added procedure time is another consideration. Of course, using an additional technology/tool consumes time that would not have been used otherwise. However, it may be that CBCT ultimately can save time as a bronchoscopist can know with certainty if the desired target has been reached or not and only obtain biopsies when “on target,” avoiding multiple non-diagnostic samples.

Literature Review: TheState

of the Art

Over the last few years, original investigation of CBCT has been growing at an accelerated pace, not only in the form of case reports and series but predominantly in much more sophisticated studies [43, 49]. These studies lay the foundation and generate the hypotheses for future research that will more defnitively identify the role of true real time, three-dimensional image guidance in the diagnosis of pulmonary lesions. We will provide a review of the salient points herein. The studies are summarized in Table 25.1.

The earliest study of CBCT in the evaluation of pulmonary nodules was by Hohenforst-­ Schmidt et al., published in 2014 [50]. This prospective, feasibility study included 33 nodules in as many patients. Mean diameter was 25 mm. A steering catheter was allowed to be used at the bronchoscopist’s discretion. The only biopsy tool used was forceps. Navigational yield, defned as the forceps being located within the nodule, was 91% overall. However, diagnostic yield (not explicitly defned) was lower at 70%. Regarding complications, two patients developed pneumothorax. The radiation dose was not specifcally reported but was referred to as <2 mSv. The same group reported levels between 0.98 and 1.15 mSv

in an earlier phantom study of the same technology and referenced that paper as a precursor to the current study [45]. Overall, this study made the frst case for the utility of real-time intraoperative imaging.

Subsequent studies of CBCT did not appear until three years later in 2017. Park and colleagues retrospectively reviewed data on 59 patients who underwent peripheral bronchoscopy in their institution in South Korea [51]. Again, the only biopsy tool used was forceps. The bronchoscope was advanced to the segmental bronchus in question after which forceps were further advanced under uoroscopic guidance. A CT was then performed and the forceps adjusted as needed. A maximum of two CT scans were allowed in order to limit radiation exposure. Average lesion size was 33 mm. Overall, diagnostic yield was 71.2% with a sensitivity for malignancy of 85.7%. Multivariate regression analysis was performed to identify factors associated with diagnostic yield; unsurprisingly, the only factor found to have a signifcant association was when forceps were identifed by CT to be located within the target lesion compared to not having reached the lesion. This same year, Bowling et al. also reported a smaller cohort of patients (n = 14) who underwent EMN-guided peripheral bronchoscopy with CBCT but with the added use of a transparenchymal access tool designed to reach lesions without a bronchus sign [44]. A steerable, extended working channel (Edge catheter, Medtronic, Inc) was also used in all cases. Overall diagnostic yield was 71%. Of note, the investigators made a change to their procedural protocol at the midpoint of patient recruitment. Initially, they would deploy the transparenchymal access tool and/or the biopsy tool and then obtain the CBCT. However, they noted that this could lead to distortion of the image of the target lesion due to bleeding that had been incited by the intervention. This made it diffcult to know how to redirect the tools to better attempt to reach the lesion. As a result, they changed their approach by frst obtaining a CBCT in order to determine the location of the extended working channel relative to the target before deploying the transparenchymal access tool.

Table 25.1  Summary of cone beam CT studies in peripheral bronchoscopy

442

Casal .F .R and Sabath .F .B

com/.https://meduniver сайта язык русский на перевода для списке в находится книга Данная

N number of nodules sampled, DY diagnostic yield. See text for defnition, if described, NR not reported, TPAT transparenchymal access tool, EWC extended working channel, EMN electromagnetic navigation, AF augmented uoroscopy, BAL bronchoalveolar lavage, RP-EBUS radial probe endobronchial ultrasound, VBN virtual bronchoscopic navigation

aValues are only for EMN + CBCT group as the comparison group did not use CT

bValue for target size is for the primarily-CBCT group

cValues are for the AF group as the comparison group did not use CT

dValues represent the latter part of the study as the design was to evaluate how outcomes changed over time

Bronchoscopy Guided-Tomography Computed Beam Cone  25

443

Interestingly, the remaining 7 patients managed under this new protocol had a 100% diagnostic yield.

The year 2018 saw the publication of three separate studies on cone-beam CT for peripheral bronchoscopy. Pritchett et al. described their retrospective experience with 93 nodules biopsied in 75 consecutive patients [52]. All nodules were sampled in a single sitting, guided by electromagnetic navigation with augmented uoroscopy. The aforementioned curved, steerable catheter was also used. After intubation, an initial CT was obtained. The data were uploaded into software through which the nodule underwent three-dimensional segmentation. This segmentation then could be overlayed on live uoroscopic imaging that guided the advancement of biopsy tools once the bronchoscope was steered to the appropriate location using EMN guidance. The tools used included standard cytology brush, fne needle for aspiration, forceps, GenCut core biopsy tool (Medtronic), cytology brushes, and bronchoalveolar lavage (BAL). Median lesion size was 16 mm with less than half being visible by standard uoroscopy. Diagnostic yield was 83.7% with diagnostic accuracy of 93.5%. The former was defned as the bronchoscopy procedure providing a specifc malignant or benign diagnosis, excluding non-specifc fndings such as in ammation; the defnition for the latter included those lesions that were subsequently confrmed as benign by clinical and radiographic follow-up. Most cases required only a single CT scan which the authors attribute to the fact that they used augmented uoroscopy as an additional guidance tool. Of note, in this study CBCT was mostly used for navigation rather than for confrmation. Soon after this report, Sobieszczyk and colleagues also described their retrospective review of 22 patients with 22 nodules [53]. Electromagnetic navigation was used to reach the lesions with subsequent radial EBUS and CBCT to confrm the location they had attained. Only a single CT scan was performed in each case. A transparenchymal sampling tool was used as needed. In addition to the Edge catheter, a variety of tools were used at the discretion of the bronchoscopist, including either forceps, fne needle,

brush, or GenCut. Mean nodule size was 21 mm. Overall diagnostic yield was 77.2% with a statistical trend toward increased yield with increasing size of the lesion, and there were no complications. The defnition of diagnostic yield was not described. When used, the transparenchymal tool led to a diagnosis in all 7 cases in which it was used. Finally in 2018, Casal et al. published the frst study of CBCT combined with thin/ultrathin bronchoscopy and radial EBUS [42]. Entry criteria limited nodules to 10–30 mm in size and only if located within the outer 2/3 of the lung feld. Diagnostic yield was defned as the proportion of patients in whom a malignant or benign process was identifed, the latter including pathology that was confrmed benign either surgically or clinicoradiographically at six-month follow-up. Additionally—and distinct from previous stud- ies—these investigators specifed the added beneft of CBCT. Termed “post-CBCT yields,” these were defned as the proportion of cases in which CBCT allowed the operator to reach the lesion when it had not been reached prior to the scan (“post-CBCT navigational yield”) and, thereby, to obtain a diagnosis that would not have been made prior to the scan (“post-CBCT diagnostic yield”). When the CT demonstrated that initial navigation to the nodule was unsuccessful, additional maneuvers were employed such as re-­ navigation,changingtoanultrathinbronchoscope, or using additional tools. A second CBCT would then be performed to evaluate the result of the adjustments made. Median nodule size was 21 mm. Sixty percent of targets were not visible by uoroscopy. Pre-CBCT navigation and diagnostic yields were both 50%; additional post-­ CBCT maneuvers increased the navigation yield to 75% and diagnostic yield to 70%. The additional nodules that were reached and diagnosed with the use of CBCT were all smaller than the average of the other cases and 75% were not visible by uoroscopy. Interestingly, an additional beneft of CBCT described in this study was that it allowed for the incidental discovery of the development of intra-procedural atelectasis, obscuring the target in 20% of cases. This observation led to the subsequent study described above that specifcally investigated and further

unfolded how atelectasis can develop during bronchoscopy [25]. One pneumothorax occurred but otherwise no adverse events were observed.

The following year also saw the publication of a study combining CBCT with ultrathin bronchoscopy,­ but with the use of virtual navigation (VBN) technology [54]. Patients were included with solid or subsolid nodules ≤30 mm but nodules without a bronchus sign were excluded. VNB software was used to guide bronchoscopy as close as deemed possible to the target bronchus of interest and forceps were introduced. Thereafter, CBCT was performed to demonstrate the location reached; CBCT was repeated as needed after adjustments were made to approximate the biopsy tool to the lesion as closely as possible. Forceps biopsies, brushings, and/or bronchoalveolar lavage were performed depending on how closely the nodule was reached (if CT confrmed that the nodule was not reached at all, only bronchoalveolar lavage was performed). In all, 40 patients were enrolled with a median tumor size of 20 mm. Eighty percent of nodules had a bronchus leading to the center of the lesion. An average of 1.8 scans were performed after forceps insertion (range 1–5). Overall diagnostic yield was 90%, subdivided into lesions with an airway leading to their center (diagnostic yield 96.9%) and those with an airway leading to the nodule periphery (diagnostic yield 62.5%). However, diagnostic yield was not clearly defned. Nevertheless, this higher yield compared to other studies was postulated by the authors to be likely related to patient selection as those without a bronchus sign were excluded. One patient developed a pneumothorax and lung abscess.

With the background of these initial studies showing promise, investigation into CBCT continued to accelerate, with 2021 producing the most publications to date on this technology. Kheir et al. in the United States reported an interesting retrospective study design in which the diagnostic yield for EMN-guided bronchoscopy with augmented uoroscopy was determined before and after the introduction of CBCT into their practice [55]. Sixty-two patients (31 with EMN alone before CBCT, and 31 with EMN plus

CBCT) were evaluated. The aforementioned steerable Edge catheter was also used in all cases. Patients underwent usual EMN-guided bronchoscopy with the use of radial probe EBUS for confrmation of having reached the nodule in both groups. In the EMN-only group, at that point, sampling was obtained under conventional C-arm uoroscopy. In the EMN plus CBCT group, however, after radial probe confrmation, the aspirating needle was deployed and a CBCT was performed. Based on the location of the needle tip as revealed by the CT, adjustments were made under uoroscopic guidance. Repeat CBCT was performed as needed. Once satisfactory needle position was obtained, samples were taken. Tools used included cytology brush, fne needle, forceps, and BAL.Diagnostic yield was calculated as the number of true positive malignant or specifc benign diagnoses divided by the total number of procedures for that study arm; non-­ specifc in ammation, atypical cells, or normal lung parenchyma were considered non-­ diagnostic. The median size of target nodules was 16 mm in the EMN-CBCT group but 21.5 mm in the EMN-only group. Diagnostic yield was 74.2% in the EMN-CBCT group compared to 51.6% in the other. Procedures were also a median of 16 min shorter when CBCT was used. When multivariate regression analysis was performed adjusting for nodule size, presence of a bronchus sign, and distance from the pleura, EMN-CBCT was found to yield a diagnosis 3.4-­ fold more often than EMN alone (i.e., odds ratio 3.4; 95% CI 1.03–11.26, p = 0.04). Two patients in each arm developed a pneumothorax.

Benn et al. then published the frst study of CBCT combined with robotic bronchoscopy [56]. Performed at a single center, ffty-nine nodules in 52 patients (seven patients had two nodules) underwent bronchoscopy with navigation initially guided by the robotic platform followed by CBCT for secondary confrmation. Mean largest diameter in any dimension was 21.9 mm. Biopsy tools were 21-gauge needle and forceps. Overall diagnostic yield was not explicitly defned but was based on all available biopsy and imaging data, including two nodules which had resolved or regressed on follow-up CT. As such,

diagnostic yield was reported as 86% with a sensitivity for malignancy of 84%. Overall radiation exposure was reported as 1.69 mGy with a dose-­ length product of 750 mGycm. Two pneumothoraces were observed.

Verhoeven and colleagues in the Netherlands then reported their comparison of two approaches to peripheral nodules, in a crossover design: a primarily CBCT-guided approach and a primarily EMN-guided approach [57]. In the former, CBCT with augmented uoroscopy was used to reach the lesion with radial EBUS for initial confrmation. CBCT was used as needed to defnitively confrm location after which sampling was done. If this approach was unsuccessful in reaching the lesion, the case crossed over to the other study arm and EMN guidance was added. Of note, straight catheters with a steerable curette or a catheter with a preformed curvature were also used. For the primarily EMN-guided approach, usual electromagnetic navigation was used to attempt to reach the lesion. Once the lesion appeared to have been attained, radial EBUS was used for confrmation. If radial EBUS also was favorable, CBCT was performed for additional confrmation. If CBCT was needed for subsequent navigation readjustments, the case was considered to have crossed over into the CBCT arm. Biopsy tools included a cytology brush, needle, forceps, and even cryobiopsy on a case-­by-­case basis if considered safe. Navigation was deemed successful if CBCT confrmed that a tool was within a lesion or in contact with its outer boundary. Ultimately, in the primarily-CBCT-­ guided arm, 47 patients with 59 lesions were included with an average size of 16.6 mm. In the primarily-EMN-guided group, 40 patients with 48 lesions were included with an average size of 14.2 mm. Navigation was successful in 76.3% of lesions in the primarily-CBCT group but only 52.2% in the primarily-EMN group (p = 0.016). When EMN was added to CBCT group cases, navigation success increased by 13.6–89.9%; when CBCT was added to EMN group cases, navigation success increased by 35.3–87.5%. Despite high navigation success, diagnostic

accuracy was 70.2% for the primarily-CBCT group and 75% for the primarily-EMN group (p = 0.797).

The next CBCT study of 2021 was published by Yu et al. out of Taiwan [58]. This study focused on the augmented uoroscopy technology that can CBCT provide. In a propensity-matched analysis, they retrospectively evaluated patients that underwent bronchoscopy for peripheral pulmonary nodules that underwent transbronchial biopsy either with a combination of CBCT-­derived augmenteduoroscopy (AF) with radial EBUS or transbronchial biopsy with radial EBUS alone. In the AF group, a CBCT scan was performed at the beginning of the procedure. This allowed the physicians to highlight the area of interest using annotation software which then transferred and projected the annotated markers to live uoroscopy. This augmented uoroscopy, along with radial EBUS, was then used to guide bronchoscopy until a satisfactory position was attained. At that point, biopsy forceps, brushings, or bronchial washings were used at the discretion of the bronchoscopist. In the radial-EBUS-only group, radial probe EBUS was used without augmented uoroscopy, though presumably conventional two-dimensional uoroscopy was used to visualize probe positioning within the lung (however, this was not explicitly stated in the report). Diagnostic yield was based on whether a specifc malignant or benign process was identifed. Final diagnoses were based on pathologic evidence from biopsies (bronchoscopic or otherwise) as well as microbiological results or clinical follow-­up of at least one year post-bron- choscopy. In an attempt to minimize confounding in the analysis, propensity scores were generated using the following factors: age, gender, smoking, lesion size, lesion location relative to the hilum (i.e., central, middle, peripheral), and presence of bronchus sign. Ultimately, 53 pairs were matched between the two groups. Median nodule size was 28 mm in the AF group and 29 mm in the radial EBUS-­only group. Diagnostic yield was 75.5% in the AF group compared to 52.8% in the radial EBUS-­only group (p= 0.015). There was a statistically signifcant difference favoring the AF group in terms of diagnosis of malignancy but no differ-