3D computed tomography airway geometry for predicting bronchoscopic accessibility in peripheral pulmonary nodules: a prospective study

Introduction

Diagnosis of peripheral pulmonary nodules (PPNs) through bronchoscopy has been extensively investigated in recent literature (1). Bronchoscopic interventions, such as radial probe endobronchial ultrasound (rEBUS)-guided transbronchial lung biopsy (TBLB) based on computed tomography (CT)-guided approaches, have been extensively explored (2). Preprocedural chest CT imaging enables precise procedural planning by delineating key anatomical structures, such as lesion location and bronchial pathways, whereas rEBUS provides intraprocedural real-time localization of the lesion relative to the probe, enhancing the accuracy of imaging-guided bronchoscopic interventions when other forms of intraprocedural imaging are unavailable (3,4).

However, PPN identification with rEBUS depends heavily on operator knowledge of bronchial anatomy and procedural difficulty may vary according to the anatomical course of the target bronchus (5,6). Certain bronchial pathways may present greater mechanical challenges during scope advancement, underscoring the importance of additional preprocedural information for procedural planning. When a bronchoscope is advanced into an upper lobar bronchus, the direction of pushing force applied by the operator is opposite to the direction in which the bronchoscope needs to proceed (6). Consequently, accessibility to PPNs in the upper lobes can be lower than in other lobes. These anatomical constraints underscore the need for additional preprocedural information to support bronchoscopic navigation.

Lung airway morphometry using chest CT has emerged as a powerful tool for quantifying airway geometry and assessing structural abnormalities (7). Prior research has proposed CT-derived quantitative airway measures as predictors of lung function decline and disease progression (8-10). However, those findings have largely been limited to observational studies, and their utility in guiding real-world clinical procedures remains underexplored. In our previous retrospective study, we demonstrated the feasibility of CT-based airway analysis in preoperative planning (11). Although our results suggested potential clinical value, their application in prospective clinical settings had not been validated.

Therefore, in this study, we extend our previous work by implementing a prospective design to develop and clinically validate a predictive model for estimating the accessibility of PPNs during rEBUS-TBLB. We further assessed the utility of our predictive model as a preprocedural guide in routine bronchoscopic workflows. We present this article in accordance with the TRIPOD reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0133/rc).

Methods

Study participants

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Samsung Medical Center (No. 2018-03-021). Informed consent was taken from all the patients (or a statement that it was not required and why). Patient information was de-identified and anonymized prior to the analysis.

We collected data from 212 patients whose rEBUS-TBLB procedures were planned during a prospective trial at Samsung Medical Center, a single tertiary referral center, between December 2021 and February 2023. Since the procedures were performed using rEBUS alone—without additional guidance from fluoroscopy, cone-beam CT, guide sheaths, or navigation systems—patients were primarily referred for the procedure when a bronchus leading to the lesion was identifiable on pre-procedural CT to ensure procedural feasibility. During the initial clinical screening, and prior to any quantitative airway analysis, 10 patients were excluded for the following reasons: the operator determined that the lesion was clinically unsuitable for biopsy (n=6), the lesions were small pure ground-glass opacities (n=3), or the lesion improved on follow-up imaging before the procedure (n=1). Quantitative airway analysis was then attempted in the remaining 202 patients. However, three patients were further excluded because of processing failures due to technical or software issues. Consequently, a total of 219 PPNs from 199 patients were included in the final analysis. Because several patients presented with multiple PPNs, each lesion was prospectively assessed and included in the analysis to reflect the full clinical spectrum encountered in real-world practice. The overall study flow is illustrated in Figure 1.

Figure 1 Flow chart of patient inclusion. EBUS-TBLB, endobronchial ultrasound-guided transbronchial lung biopsy; GGN, ground-glass nodule.

CT acquisition and preoperative evaluation

All patients underwent chest CT scanning prior to rEBUS-TBLB as part of the prospective study protocol. CT scans were acquired with standard thin slices (≤1 mm) and high-resolution reconstruction; detailed scanning parameters are available in a supplementary appendix online (Appendix 1).

Several radiologic characteristics were visually assessed on chest CT: lesion size, lesion depth from the pleural surface, number of bronchial branches from the trachea to the lesion, and lesion solidity (solid, sub-solid, or cavitary). The lesion location was categorized into five anatomical regions: right upper, right middle, right lower, left upper, and left lower lobes. Furthermore, to account for anatomical and physiological factors affecting procedural difficulty, we categorized the segments into three bronchial clusters: (I) segments requiring acute angulation for access, specifically the apical and superior segments (RB1, LB1, RB6, LB6); (II) segments with high respiratory-induced mobility, including the basal segments (RB8–10, LB8–10); and (III) all other segments. This classification reflects established challenges in bronchoscopy, such as the acute angles of the upper lobes (6) and the significant respiratory displacement common in the lower lobes (12). All CT image assessments were performed by a board-certified thoracic radiologist (H.Y.L.) with 18 years of experience in thoracic imaging.

To assess preoperative airway information, we performed three-dimensional airway segmentation and lesion localization using chest CT data and commercial deep-learning software (AVIEW; Coreline Soft, Seoul, Korea). The segmented airway and lesion volumes were reviewed by an expert radiologist (H.Y.L.). Quantitative airway geometrical features were extracted using in-house software (AirGeo) that was developed and examined in a previous study (11). The extracted features included cross-sectional dimensions (e.g., minimum diameter, perimeter, circularity), spatial orientation (e.g., curvature, torsion), and branching characteristics (e.g., bifurcation angle). The detailed quantitative analysis process and a comprehensive list of airway geometry variables are provided in Appendix 1 and Table S1.

The extracted CT-based airway information was provided to the bronchoscopist prior to the procedure, enabling preoperative planning. It is important to note that the AirGeo-based analysis provided only raw quantitative data (e.g., subsegmental routes and bronchial angles) for preoperative planning. It did not offer any qualitative assessment of predicted difficulty. Thus, it served as an informative tool to guide the anatomical approach, rather than as a behavioral intervention. The proceduralist’s effort was standardized across all cases, and the study’s objective was to determine which of these objective preoperative parameters were significantly associated with the actual technical difficulty encountered during the procedure. Figure 2 shows the overall workflow of our study.

Figure 2 Overall workflow of the prospective study design. CT, computed tomography; EBUS-TBLB, endobronchial ultrasound-guided transbronchial lung biopsy; LA, luminal area; PPNs, peripheral pulmonary nodules.

Procedure for rEBUS-TBLB

All rEBUS-TBLB procedures were performed on patients under conscious sedation using midazolam and fentanyl, as previously reported (13). Briefly, a 4.1-mm bronchoscope (BF P260F; Olympus, Tokyo, Japan) was advanced to the sub-subsegmental bronchus nearest the lesion after reviewing the CT. Then, the rEBUS probe (1.4-mm, 20-MHz, UM S20-17S; Olympus) was inserted through the bronchoscope’s working channel. Upon confirming lesion visualization via ultrasound, a 1.8-mm biopsy forceps was inserted through the working channel for TBLB after the probe was removed. All procedures were performed by one operator (B.H.J.), a bronchoscopist with 11 years of clinical experience, without a guide sheath, fluoroscopy, or navigation system.

When the target was identified on ultrasound, the findings were classified as “within” (the radial probe was within the target lesion, surrounded by lesion material) or “adjacent” (the probe did not pierce the lesion but was nearby). When the nodule was not able to be visualized by the probe, it was defined as “invisible”, indicating failure of endobronchial navigation using thin bronchoscopy.

Independent of the relationship between the probe and the lesion, the difficulty in reaching the lesion was defined by the operator as “easily accessible”, “difficult but accessible”, or “inaccessible”. To provide a more objective assessment, “easily accessible” was defined as reaching the target lesion within one or two attempts after identifying the bronchial path on preprocedural CT. “Difficult” access (including “difficult but accessible” and “inaccessible”) was defined as cases requiring more than three attempts (>3 attempts) to successfully navigate the probe to the target.

We analyzed 219 procedures from 199 patients (some with multiple lesions) to validate the prediction model. Additionally, the procedure time was recorded from radial probe insertion to biopsy completion to assess procedural efficiency.

Statistical analysis

In this study, we focused on airway geometric characteristics as three levels of feature sets. First, six global level features were defined using five first-order statistics about the bifurcation angles of the total branches and the total length of each sectional branch. Next, 64 features were considered from the last two branches leading to the target lesion because they are likely most relevant for accessing the lesion. All cases were grouped by final accessibility (easily accessible vs. difficult or inaccessible).

For the prediction model, multicollinearity among the extracted features was assessed using the variance inflation factor (VIF), excluding features with VIF >4 (14). Univariate logistic regression identified variables associated with accessibility, and redundant variables (P<0.05) were removed. Feature selection and predictive scoring were performed using logistic-least absolute shrinkage and selection operator (LASSO) with 10-fold cross-validation to determine the optimal penalty term. The predictive score for each patient was calculated as a weighted sum of the selected features.

We developed three prediction models: clinical, airway geometrical, and composite (combining both feature sets). Model performance was evaluated using receiver operating characteristic (ROC) curves and corresponding area under the curve (AUC) values. DeLong’s test was applied to compare model classification performance (15). Calibration was assessed by plotting observed versus predicted probabilities, where an ideal nomogram would closely follow the 45-degree reference line (16).

Normality of continuous variables was assessed using the Shapiro-Wilk test. Because most continuous variables were not normally distributed, they are presented as median [interquartile range (IQR)] and were compared using the Mann-Whitney U test. Categorical variables were compared using the Chi-squared test. All statistical analyses were performed using both SPSS version 20.0.0.0 (IBM Corp., Armonk, NY, USA) and R 3.2.5 software (R Foundation for Statistical Computing, Vienna, Austria). A P value less than 0.05 was considered statistically significant.

Results

Characteristics of cohorts

Demographics for the enrolled patients are shown in Table 1. The patients [median age, 70.0 years (IQR, 62.0–77.0 years); 118 male and 81 female] had lesions with three types of solidity: solid (n=150, 68.5%), sub-solid (n=47, 21.5%), and cavitary (n=22, 10.0%). Compared with easily accessible cases, difficult or inaccessible cases had significantly smaller target lesions [19.2 mm (IQR, 15.6–22.8 mm) vs. 22.8 mm (IQR, 17.7–29.5 mm), P=0.008]. There was no significant difference between groups in the number of branch levels (P=0.56). Although lesion depth from the pleural surface did not reach statistical significance, it was numerically smaller in difficult or inaccessible cases [4.3 mm (IQR, 0–10.4 mm) vs. 6.9 mm (IQR, 0–16.9 mm), P=0.08]. The distribution of the PPNs in the five lobes did not differ statistically between the groups. However, in terms of bronchus clustering, more difficult or inaccessible cases than easily accessible cases tended to be found in the RB1/LB1/RB6/LB6 cluster (45.9% vs. 25.8%, P=0.059). The ‘invisible’ bronchus sign was significantly more frequent in the difficult or inaccessible group than the easily accessible group (16.2% vs. 3.3%; P=0.007). While the “within” was more frequently observed in the easily accessible group (155/182, 85.2%) than in the difficult or inaccessible group (27/37, 73.0%), its presence in a substantial proportion of the difficult or inaccessible group cases was notable.

Procedure results are shown in Table S2. The procedure time was longer in the difficult or inaccessible cases than in the easily accessible cases (15.8±8.7 vs. 9.4±5.0 min, P<0.001). Radial probe positioning within the target lesion during the procedure was more common in the easily accessible cases than in the difficult or inaccessible cases (89.0% vs. 45.9%, P<0.001). Diagnostic results from pathology were obtained for 136/182 (74.7%) easily accessible cases and only 10/37 (27.0%) difficult or inaccessible cases (P<0.001).

Risk factors associated with difficult or inaccessible PPNs

We performed univariable logistic regression analyses to identify clinical variables (Table 2) associated with difficult or inaccessible PPNs. Larger lesion size [per 1 mm, odds ratio (OR) =0.931, P=0.007] and centrality (1 mm farther from pleura, OR =0.963, P=0.046) are likely to increase target lesion accessibility. For bronchus clustering, lower accessibility was observed if the targeted lesion was located at RB1, RB6, LB1, or LB6 (OR =2.288, P=0.044). When the bronchus sign on preprocedural CT was ‘invisible’ (OR =5.741, P=0.004) or the procedure time was longer (per 1 min, OR =1.156, P<0.001), bronchoscopic access to the target lesion was more likely to be difficult.

Several airway geometrical features were significantly associated with difficult or inaccessible PPNs (Table S3). Among bifurcation angle metrics at the total branch level, higher sum (OR =1.505, P=0.02), maximum (OR =1.595, P=0.01), and average (OR =1.636, P=0.006) values were linked to reduced accessibility. At the branch preceding the final segment, a higher minimum of the min-max diameter ratio indicated greater accessibility (OR =0.641, P=0.02). At the final branch level, increased values of the average max inscribed sphere R (OR =0.623, P=0.01), maximum hydraulic luminal diameter (OR =0.349, P=0.040), and average hydraulic luminal diameter (OR =0.496, P=0.008) were associated with better accessibility. Conversely, greater maximum and average curvature were linked to inaccessibility (OR =1.512, P=0.01; OR =1.474, P=0.03, respectively). Representative cases illustrating these associations are shown in Figure 3.

Figure 3 Representative cases of peripheral pulmonary nodules demonstrating airway pathway visualization. Left: cavitary lesion in LB3 of a 66-year-old male patient, located close to the pleural surface. Right: solid lesion in LB4 of a 75-year-old male patient, positioned farther from the pleural surface within the lung parenchyma. Both cases showed a “within” bronchus sign on preprocedural computed tomography. Airway geometry maps display the reconstructed pathway from the trachea to each lesion, with segmental coloring indicating sequential airway branches and an enlarged view highlighting the distal approach route. LB3, anterior segmental bronchus of the left upper lobe; LB4, superior lingular bronchus of the left upper lobe.

Performance of the prediction models

To identify patients with PPNs that were difficult to access with rEBUS, we applied the logistic-LASSO method to develop a prediction model and assessed the performance metrics using 10-fold cross-validation (Table 3). In the final composite model, smaller lesion size; closeness to the pleural surface; location in RB1, RB6, LB1, or LB6; invisible bronchus sign; larger bifurcation angle-in; smaller min-max diameter ratio; larger curvature; and smaller hydraulic luminal diameter all predicted difficult PPN access for rEBUS.

The composite model, which combines clinical variables with airway geometrical variables, performed better than either model alone [AUC (95% CI); composite model =0.84 (0.76–0.92), clinical model =0.68 (0.60–0.81), airway geometrical model =0.79 (0.74–0.84)] (Table 4, Figure 4). The Delong test showed significant model differences (P=0.05). Calibration curves showed good agreement between predicted and actual accessibility, indicating well-developed models (Figure S1). The composite model showed greater consistency between observed and predicted probabilities.

Figure 4 ROC curves of the three predictive models. AUC values are presented as AUC (95% confidence interval, 95% CI). AUC, area under the curve; ROC, receiver operating characteristic.

Discussion

In this study, we prospectively evaluated the accessibility of 219 PPNs (median size, 22.4 mm; median distance from the pleura, 6.8 mm) to rEBUS. To create a model for predicting accessibility, we analyzed both clinical and airway geometrical factors. The composite model showed superior predictive performance (AUC =0.84) compared with the clinical model (0.68). These results suggest that objective quantitative assessment using CT-based airway geometrical data may be more useful than clinical information alone.

Several meta-analyses have reported pooled diagnostic yields of approximately 70% for PPNs measuring 15–40 mm when using rEBUS (17-19). In our study, the median lesion size was 22.4 mm and the diagnostic yield was 66.7%, suggesting comparable lesion characteristics. Unlike previous studies focusing on diagnostic yield, we analyzed the accessibility of PPNs to rEBUS. Because diagnostic yield varies substantially depending on malignancy status and biopsy technique (20), we focused on anatomical accessibility using CT-based airway information. Clinical factor analysis showed that larger lesion size, farther distance from the pleura, and presence of the bronchus sign were associated with easier access—findings consistent with previous meta-analyses (18,21,22). Moreover, we assessed accessibility by bronchus clustering rather than lobe-wise locations and found that apical segments of the upper lobes (RB1, LB1) and posterior segments of the lower lobes (RB6, LB6) were significantly less accessible, likely due to their greater curvature (5).

In the composite model, six airway geometrical features were selected as significant risk factors for difficult or inaccessible cases (Table S4). These included larger bifurcation angles, lower min-max diameter ratio, greater curvature, and smaller hydraulic luminal diameter. Clinically, these features indicate that more acutely angled, irregularly shaped, and narrowed distal airways present barriers to successful rEBUS navigation—findings that are consistent with our previous pilot study (11). The present results not only validate these risk factors in a prospective cohort but also extend prior work by integrating both clinical and quantitative airway data within a unified predictive model.

Unlike earlier studies, our analysis directly compared the predictive value of clinical variables, airway geometrical features, and their combination. The composite model achieved a higher AUC (0.84) than either the clinical (0.68) or airway geometrical (0.79) models alone, with notably improved sensitivity. These findings support the clinical utility of combining quantitative CT-based airway measurements with conventional factors to better predict bronchoscopic accessibility.

Notably, 73.0% of the lesions classified as difficult or inaccessible exhibited a ‘within’ bronchus sign on preprocedural CT, highlighting the limitations of conventional visual assessment. The bronchus sign can often fail to capture underlying anatomical complexities such as sharp bifurcation angles, severe curvature, or airway narrowing. In contrast, our quantitative CT-based airway analysis could provide a more objective and comprehensive assessment, enabling early identification of difficult cases—even when CT appearances seemed favorable. This approach has the potential to enhance preprocedural planning and procedural success, supporting its integration into real-world bronchoscopic workflows.

The proposed CT-based airway geometric analysis is not intended to replace existing bronchoscopic navigation systems, but rather to complement them by providing objective preprocedural information. Unlike many navigation technologies that are applied primarily during bronchoscopy, our approach uses quantitative CT-derived airway measurements to estimate procedural difficulty before the intervention, thereby supporting earlier and more structured decision-making. Because this framework relies on reproducible geometric parameters rather than visual bronchial assessment alone, it may reduce subjectivity in procedural planning. Furthermore, by identifying lesions likely to be difficult to access in advance, this method may help clinicians selectively allocate adjunctive navigation or robotic systems to cases in which they are most likely to provide incremental benefit, thereby improving procedural efficiency and potential cost-effectiveness.

The procedures in this study were performed under conscious sedation without the use of a guide sheath, fluoroscopy, or navigation systems. While these adjuncts have been shown to enhance diagnostic yield in several studies (5,23-25), their routine use across all cases may be limited by medical costs and resource availability (26). Recent meta-analyses of high-cost robotic-assisted bronchoscopy (RAB) reported diagnostic yields approaching 80% (27,28). However, as noted in previous literature, RAB studies often involve a higher proportion of challenging targets, including smaller nodules, bronchus sign-negative lesions, and subsolid opacities, which are inherently more difficult to sample than the lesions typically included in standalone rEBUS studies. Therefore, rather than a direct comparison of diagnostic yields, we propose that our CT-driven airway geometrical analysis (AirGeo) could serve as a tool for procedural triage. By preoperatively identifying lesions predicted to be “inaccessible” via standalone rEBUS, clinicians could strategically refer these high-difficulty cases for advanced platforms like RAB or navigation-guided bronchoscopy under general anesthesia. Such a risk-stratified approach could optimize both clinical outcomes and cost-effectiveness by matching the complexity of the lesion with the appropriate level of technological intervention.

Our study has several limitations. First, although the study was conducted prospectively, it was performed at a single center, and all bronchoscopic procedures were carried out by one experienced bronchoscopist. This design may limit the generalizability of the findings to other institutions, procedural settings, and operators with different levels of experience. However, the use of a single operator also reduced procedural heterogeneity and enabled a more consistent assessment of the relationship between CT-derived airway geometrical features and bronchoscopic accessibility. Future multicenter studies including bronchoscopists with varying levels of expertise are warranted to confirm the external validity and reproducibility of the proposed approach. Another limitation is that the operational definition of bronchoscopic accessibility—although based on the number of procedural attempts, and therefore more structured than a purely subjective assessment—may still be affected by operator-dependent factors such as procedural technique, persistence, and intraprocedural judgment. Accordingly, this definition should be interpreted as a pragmatic procedural surrogate rather than a fully anatomy-specific measure of lesion accessibility. In this study, the use of a single experienced bronchoscopist helped reduce inter-operator variability; however, residual operator dependence cannot be excluded. Future studies are needed to validate this concept using more standardized and less operator-sensitive criteria. Second, our proposed model was not externally validated due to the limited dataset. However, to minimize overfitting and improve model robustness, we conducted internal validation using cross-validation during logistic-LASSO regression modeling. Third, we evaluated only accessibility and not the diagnostic yield of rEBUS for PPNs. As noted in recent literature, diagnostic yield is heavily influenced by confounding variables such as the type of biopsy instruments used (e.g., forceps vs. cryobiopsy), the prevalence of malignancy, and the pathological nature of the lesion (19,20). By focusing strictly on accessibility—defined as the ability to navigate the probe to the target—we aimed to isolate the impact of airway geometric factors from these variables and provide a purer assessment of navigational predictability. Despite those limitations, our study identified the potential of CT-based quantitative assessment for enhancing bronchoscopic procedures and improving PPN accessibility to rEBUS.

Conclusions

In conclusion, this study prospectively evaluated the bronchoscopic accessibility of PPNs by constructing predictive models that incorporate CT-derived airway geometrical features. By integrating these imaging-derived variables with clinical information, we demonstrated their clinical significance in preprocedural planning. Our findings validated the utility of a quantitative airway analysis as an objective tool for assessing procedural difficulty and support its potential role in real-world bronchoscopic workflows. The prospective study design further strengthens the evidence that CT-based airway geometry holds practical relevance and promising potential for improving bronchoscopic strategy and decision-making in clinical practice.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0133/rc

Data Sharing Statement: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0133/dss

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0133/prf

Funding: This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (No. NRF-2022R1A2C1003999) and by the Future Medicine 20*30 Project of Samsung Medical Center (No. SMO1250061) and partly supported by an Institute of Information & Communications Technology Planning & Evaluation grant funded by the Korean government (No. RS-2021-II212068, Artificial Intelligence Innovation Hub).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0133/coif). The authors have 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. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Samsung Medical Center (No. 2018-03-021). Informed consent was taken from all the patients (or a statement that it was not required and why).

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. Shen YC, Chen CH, Tu CY. Advances in Diagnostic Bronchoscopy. Diagnostics (Basel) 2021;11:1984. [Crossref] [PubMed]
2. Steinfort DP, Khor YH, Manser RL, et al. Radial probe endobronchial ultrasound for the diagnosis of peripheral lung cancer: systematic review and meta-analysis. Eur Respir J 2011;37:902-10. [Crossref] [PubMed]
3. Chen A, Chenna P, Loiselle A, et al. Radial probe endobronchial ultrasound for peripheral pulmonary lesions. A 5-year institutional experience. Ann Am Thorac Soc 2014;11:578-82.
4. 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.
5. Kurimoto N, Miyazawa T, Okimasa S, et al. Endobronchial ultrasonography using a guide sheath increases the ability to diagnose peripheral pulmonary lesions endoscopically. Chest 2004;126:959-65. [Crossref] [PubMed]
6. Oki M, Saka H, Imabayashi T, et al. Guide sheath versus non-guide sheath method for endobronchial ultrasound-guided biopsy of peripheral pulmonary lesions: a multicentre randomised trial. Eur Respir J 2022;59:2101678. [Crossref] [PubMed]
7. Kirby M, Smith BM, Quantitative CT. Scan Imaging of the Airways for Diagnosis and Management of Lung Disease. Chest 2023;164:1150-8. [Crossref] [PubMed]
8. Diaz AA, Rahaghi FN, Ross JC, et al. Understanding the contribution of native tracheobronchial structure to lung function: CT assessment of airway morphology in never smokers. Respir Res 2015;16:23. [Crossref] [PubMed]
9. Nakano Y, Muro S, Sakai H, et al. Computed tomographic measurements of airway dimensions and emphysema in smokers. Correlation with lung function. Am J Respir Crit Care Med 2000;162:1102-8.
10. Washko GR, Diaz AA, Kim V, et al. Computed tomographic measures of airway morphology in smokers and never-smoking normals. J Appl Physiol 1985;2014:668-73. [Crossref] [PubMed]
11. Cho HH, Choe J, Kim J, et al. 3D airway geometry analysis of factors in airway navigation failure for lung nodules. Cancer Imaging 2024;24:84. [Crossref] [PubMed]
12. 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]
13. Moon SM, Choe J, Jeong BH, et al. Diagnostic Performance of Radial Probe Endobronchial Ultrasound without a Guide-Sheath and the Feasibility of Molecular Analysis. Tuberc Respir Dis (Seoul) 2019;82:319-27. [Crossref] [PubMed]
14. Miles J. Tolerance and variance inflation factor. In: Everitt BS, Howell DC. editors. Encyclopedia of Statistics in Behavioral Science. John Wiley & Sons, Ltd.; 2005. Available online: https://onlinelibrary.wiley.com/doi/epdf/10.1002/9781118445112.stat06593
15. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45.
16. Lagendijk J, Steyerberg EW, Daalderop LA, et al. Validation of a prognostic model for adverse perinatal health outcomes. Sci Rep 2020;10:11243. [Crossref] [PubMed]
17. McGuire AL, Myers R, Grant K, et al. The Diagnostic Accuracy and Sensitivity for Malignancy of Radial-Endobronchial Ultrasound and Electromagnetic Navigation Bronchoscopy for Sampling of Peripheral Pulmonary Lesions: Systematic Review and Meta-analysis. J Bronchology Interv Pulmonol 2020;27:106-21. [Crossref] [PubMed]
18. Lee J, Song JU. Diagnostic yield of radial probe endobronchial ultrasonography-guided transbronchial biopsy without fluoroscopy in peripheral pulmonary lesions: A systematic review and meta-analysis. Thorac Cancer 2023;14:195-205. [Crossref] [PubMed]
19. Sryma PB, Mittal S, Madan NK, et al. Efficacy of Radial Endobronchial Ultrasound (R-EBUS) guided transbronchial cryobiopsy for peripheral pulmonary lesions (PPL...s): A systematic review and meta-analysis. Pulmonology 2023;29:50-64. [Crossref] [PubMed]
20. Ali MS, Trick W, Mba BI, et al. Radial endobronchial ultrasound for the diagnosis of peripheral pulmonary lesions: A systematic review and meta-analysis. Respirology 2017;22:443-53. [Crossref] [PubMed]
21. Tian S, Li X, Liu J, et al. Radial endobronchial ultrasound - guided bronchoscopy for the diagnosis of peripheral pulmonary lesions: A systematic review and meta-analysis of prospective trials. Heliyon 2024;10:e29446. [Crossref] [PubMed]
22. Kim SH, Chung HS, Kim J, et al. Development of the Korean Association for Lung Cancer Clinical Practice Guidelines: Recommendations on Radial Probe Endobronchial Ultrasound for Diagnosing Lung Cancer - An Updated Meta-Analysis. Cancer Res Treat 2024;56:464-83. [Crossref] [PubMed]
23. Jiang S, Xie F, Mao X, et al. The value of navigation bronchoscopy in the diagnosis of peripheral pulmonary lesions: A meta-analysis. Thorac Cancer 2020;11:1191-201. [Crossref] [PubMed]
24. Tamiya M, Okamoto N, Sasada S, et al. Diagnostic yield of combined bronchoscopy and endobronchial ultrasonography, under LungPoint guidance for small peripheral pulmonary lesions. Respirology 2013;18:834-9. [Crossref] [PubMed]
25. Eberhardt R, Anantham D, Ernst A, et al. Multimodality bronchoscopic diagnosis of peripheral lung lesions: a randomized controlled trial. Am J Respir Crit Care Med 2007;176:36-41. [Crossref] [PubMed]
26. Vakil E, Jackson N, Sainz-Zuñega PV, et al. Optimizing Diagnostic and Staging Pathways for Suspected Lung Cancer: A Decision Analysis. Chest 2021;160:2304-23. [Crossref] [PubMed]
27. Zhang C, Xie F, Li R, et al. Robotic-assisted bronchoscopy for the diagnosis of peripheral pulmonary lesions: A systematic review and meta-analysis. Thorac Cancer 2024;15:505-12. [Crossref] [PubMed]
28. Ali MS, Ghori UK, Wayne MT, et al. Diagnostic Performance and Safety Profile of Robotic-assisted Bronchoscopy: A Systematic Review and Meta-Analysis. Ann Am Thorac Soc 2023;20:1801-12. [Crossref] [PubMed]