A recent study showed that a low-dose computed tomography (CT) screening of the thorax reduced lung cancer mortality.1 However, it was associated with 96.4% of false-positive results for lung cancer. Tissue biopsy was needed to confirm malignancy. Bronchoscopic lung biopsy had a variable diagnostic yield depending on size and location of the peripheral lung lesion (PLL). The diagnostic sensitivities of bronchoscopy for PLLs smaller and larger than 2 cm in diameter were 0.34 and 0.63, respectively and could be <20%.2–5 Endobronchial ultrasound (EBUS) enhances the diagnostic yield of lung nodules.6 A miniature radial EBUS probe allows direct circular contact with the small airways, bringing the surrounding structures into vision. It is conveniently inserted into the peripheral airway through the working channel of a bronchoscope to locate a PLL. In a prospective trial with 150 patients, Kurimoto et al6 demonstrated that EBUS-guided bronchoscopic tissue sampling established their diagnoses in 77% of cases without any complications.
Multidetector CT-generated virtual bronchoscopy (VB) provides 3-dimensional (3D) views from the inside of the airways that simulates findings in bronchoscopy. A pathway to the PLL can be selected for individual patient in the virtual world using CT software instead of inside a patient’s body using EBUS probe to search among numerous bronchial branches.7 Existing automated bronchoscopic navigation systems (LungPoint, Broncus Technologies Inc., CA and Bf-NAV1, KGT; Olympus Medical Systems, Tokyo) are not widely available and are relatively expensive. Nonetheless, the study of bronchial path to a PLL is still possible with the CT workstations or open source software which is readily available to most Macintosh platform.8,9 Our objective was to study if VB generated by this software could shorten procedure time of EBUS-guided bronchoscopy as compared with no VB assistance.
MATERIALS AND METHODS
Consecutive patients referred to our endoscopy unit for EBUS-guided bronchoscopy would be evaluated for eligibility. The inclusion criteria were (1) age >18; (2) PLL on CT thorax defined as a lesion surrounded by lung parenchyma or parietal pleura; and (3) no evidence of endobronchial abnormalities in CT thorax. Patients without CT thorax, unfit for consent or with evidence of endobronchial abnormalities in CT thorax would be excluded from the study.
CT thorax was performed using 64 detector multislice CT (Lightspeed VCT, General Electric Medical Systems) with the following parameters: helical type, 120 kV, Smart mA, pitch of 1, large field of vies, slice thickness 0.625 mm, and rotation time 0.7 second, scanning from level of lung apex to adrenal during a single breath-hold. The set of digital imaging and communications in medicine (DICOM) CT data (0.625 mm, plain, soft tissue kernel) were transferred to a computer equipped with an advance open source processing software (OsiriX, Pixmeo, Switzerland) for generation of VB images using volume-rendering technique.
The software works on Macintosh platform and requires a Mac OS X compatible computer with an Intel processor and a minimum of 2 GB of RAM. The “3D Endoscopy” function can be found under the “3D viewer,” highlighted in blue in Figure 1. The 3D endoscopy viewer displays 4 views in a single window: 3 orthogonal multiplanar reconstruction (MPR) views (Figs. 1A–C) and a volume-rendering view (Fig. 1D). The position of the volume-rendering camera is displayed as the intersection of 2 green lines on each MPR view. Each focal point direction can be changed by clicking the pink ball (Figs. 1A–C). The camera is controlled by the 3D mouse buttons to rotate, zoom, pan, scroll, and to change window length or window width.9
To start a VB, the volume-rendering camera was put at the luminal center of lower trachea in the 3 MRP views. The 3D volume-rendering view of the lower trachea would be automatically generated as shown in Figure 1D. The camera was scrolled along the bronchial tree to the target lesion by manipulating the mouse. The most feasible route to the lesion was selected by studying the orientation and branches of each bronchial division with the volume-rendering camera in relation to the target in the MPR views. The “Fly Thru” function allows creation of an animated movie using serial images captured. Figure 2 demonstrated a series of captured images when the VB traveled from the trachea to the target in one of the patient. An online video, Supplemental Digital Content 1, http://links.lww.com/LBR/A100 of the VB video thus created is attached as supplementary material to this article.
It took 5 to 10 minutes to upload a series of CT DICOM data to the computer equipped with OsiriX. Another 5 to 15 minutes were needed to work out the VB path and compose the video. The video could then be sent to other operators by emails and be reviewed in computers or smart phones.
Bronchoscopists would perform the procedure according to usual practice for subjects who did not have a complete set of CT DICOM data. The bronchoscopists interpreted given stacks of 2-dimensional (2D) CT scan sections, either on films or by scrolling through the axial plane sections on a computer screen before bronchoscopy (non-VB group).
Bronchoscopy and EBUS Examination
Each procedure was performed by 2 bronchoscopists with 8 to 16 years of experience in bronchoscopy. Patients were sedated with 25 to 50 mg pethidine intravenously and 1 to 3 mg midazolam intravenously. Lignocaine spray 4% was used to anaesthetize the upper airway and 2% lignocaine for the tracheobronchial tree. A thin bronchoscope (BF-P260F Evis Lucera Bronchosfibervideoscope, Olympus) measuring 4 mm in the outer diameter was used for all patients. After preliminary screening of the tracheobronchial tree, the bronchoscope was advanced to the peripheral bronchus leading to the PLL. Subsequently, EBUS was employed to confirm the exact location of the PLL. A miniature radial EBUS probe (UM-S20-17S, Olympus) inside a guide sheath (GS) (K-201) was protruded to the peripheral bronchus beyond the vision of the bronchoscope. EBUS examination at the lung periphery was performed. If the lesion could not be located, the EBUS miniprobe would be withdrawn and inserted to the nearby bronchus to continue the search. When necessary, VB images could be reviewed in a laptop computer or a smart phone. Once the lesion was located, the EBUS miniprobe would be removed, leaving the GS in-situ for tissue samplings using cytology brush and biopsy forceps.
The time at insertion of bronchoscope through vocal cord, the time at the first insertion of EBUS probe, the time at the first insertion of biopsy forceps, and time at completion of bronchoscopy were recorded. The total procedure time was defined as the time between insertion of bronchoscope through the vocal cord to removal of bronchoscope at the end of procedure. The EBUS examination time was defined as the time between the first insertion of EBUS probe and the time to the first insertion of biopsy forceps. The final diagnoses and any procedure-related complications were recorded and analyzed. For patients with negative bronchoscopic results, the final diagnoses were confirmed by surgery or by interval CT thorax for stability over 2 years if the patient could not tolerate operation.
The statistical analysis was carried out with the SPSS software version 20.0. Data were expressed mean±SD unless otherwise stated. Comparison of continuous variables (including mean examination/procedure time) between the 2 groups was made by the Mann-Whitney U test, whereas comparison of categorical variable (diagnostic yield) was made by the χ2 test or the Fisher exact test when appropriate. All hypothesis tests were 2-sided and P<0.05 was considered significant.
Thirty-three consecutive subjects were studied including 16 in the VB group and 17 in the non-VB group with a mean age of 69.6±6.6 and 64.8±12.3 years, respectively. The average diameter of the PLLs were 2.88±0.93 and 2.98±1.3 cm for the VB and non-VB group, respectively (P=0.98) (Table 1).
The mean EBUS examination time and the mean total procedure time were reduced in the VB group compared with the non-VB group: 5.3±3.9 versus 10.5±7.6 minutes (P=0.04) and 22.4±3.5 versus 29.9±10.6 minutes (P=0.044), respectively (Table 2). The sensitivity for diagnosing lung cancer and the overall diagnostic accuracy was higher in the VB group than the non-VB group: 78.6% versus 71.4% and 81.3% versus 76.4%, respectively (Tables 2 and 3). However, this study was not powered to find a difference in diagnostic accuracy. Combining patients in the VB and non-VB groups, the overall diagnostic yield was 78.7% (Table 3).
VB images were produced to a median of fifth order bronchi (range, fourth to eighth order bronchi). All 16 virtual bronchoscopies demonstrated correctly the anatomy of the bronchial trees and were consistent with the flexible bronchoscopy findings. However, in 5 patients the bronchoscope would not be advanced to the most peripheral bronchi arrived by the VB.
The diagnostic yield was higher (93.3%) when the EBUS probe was seen within the lesion, than when the EBUS probe was adjacent to the lesion (73.3%) (Fig. 3). In the non-VB group, 1 lesion remained undiagnosed even though the EBUS probe was seen within the lesion. It was confirmed to be a benign hamartoma by lobectomy. All lesions not located by the EBUS probe remained undiagnosed by bronchoscopy in both VB and non-VB group.
There was 1 small apical pneumothorax in the non-VB group and drainage was not needed. No complication was observed in the VB group.
Previous studies, both in phantom models and in humans, showed improved accuracy of ultrathin bronchoscopy in identifying the path to PLL using VB navigation.10–15 In phantom model using synthetic lesions, VB-based bronchoscopy improved the diagnostic accuracy to 96% in targets located in the third to fifth order bronchi of the right upper lobe or the right middle lobe.10 When only standard 2D CT images were used for analysis, a physician’s accuracy in defining proper 3D routes was only on the order of 40% for lesions near airways at generation 4 or less, with errors beginning as early as generation 2.10,16
We have confirmed previous findings that VB shortened procedure times.17,18 In our study, the overall diagnostic yield was 78.7% and was higher than what had been achieved using flexible bronchoscopic techniques alone.2,3 Because of the small numbers of patients included in the study, we were not able to determine if there was a difference in diagnostic yield when VB was used in addition to EBUS to assist bronchoscopy (VB vs. non-VB group, 81.3% vs. 76.4%, P=0.74). For PLLs with a diameter of <3 cm, a diagnostic yield from 63.3% to 84.4% for combining VB and EBUS GS guidance had been reported.19,20 The virtual navigation system (Bf-NAV1; Cybernet Systems, Tokyo, Japan) in a Japanese trial demonstrated significant improvement in diagnostic yields when EBUS GS was combined with VB navigation (80.4% with vs. 67% without VB).17 X-ray fluoroscopy was employed to confirm correct device placement resulting in median durations of x-ray fluoroscopy exposure of 9.7 minutes (range, 1.5 to 22.7 min) in VB navigation group and 11 minutes (range, 1.3 to 31 min) for the group without VB navigation. In our study, there was no radiation exposure as fluoroscopy was not used.
EBUS miniature probe cannot navigate itself without the bronchoscope, resulting in some lesions (8% to 20.8%) not accessible by the EBUS probe.21,22 This intrinsic limitation is evidenced in our study when overall 3 lesions (9%) were not visualized by EBUS. Electomagnetic navigation (EMN) uses an electromagnetic board to generate a magnetic field around the patients. In contrast to the EBUS miniature probe, the EMN steerable catheter can be protruded out of the bronchoscope and then flexed, rotated, and navigated to the lung periphery until the target is reached. During procedure, VB images in the EMN system automatically synchronize with the live bronchoscopic images. Several studies have demonstrated the EMN diagnostic sensitivity between 69% and 74%.23–25 A recent meta-analysis reviewed the performance of EMN, VB, EBUS, GS, and ultrathin bronchoscope in the diagnostic yield of bronchoscopic biopsies of PLLs among 39 studies comprising of over 3000 subjects.26 The weighted diagnostic yield of GS was highest (73.2%), followed by VB (72%), and EBUS (71.1%). Although EMN has a VB component, the diagnostic yield of VB alone was higher than EMN in this meta-analysis. This might be related to the software that generated the virtual pictures and the operator.26 The adverse event rate that required intervention was only 0.8%. The pooled pneumothorax rate of 1.6% and chest tube of 0.7% was much lower than that of CT-guided fine needle aspiration, 25% and 5%, respectively.26,27 The complication rate of EBUS-guided bronchoscopy was 3% in our study due to 1 small pneumothorax (3%) among 33 procedures. No chest tube was needed.
Automated computer software, such as EMN and Bf-NAV1, is not widely available in South East Asia. Compared with automated software, an obvious disadvantage of our method using OsiriX has been the time needed to construct the endoluminal view and to select the path to the target. The window width and length need adjustments, especially when the virtual bronchoscope approaches the subsegmental bronchi, as it is necessary to use different threshold values when visualizing the central and the distal airway.28 MPR images must be studied carefully to ensure the presence or absence of an airway branch in 3D views. Although it is time consuming to define endobronchial routes to PLLs in the computer equipped with OsiriX, we still prefer this method to spend time inside the patient’s airway to search for the target using the EBUS miniprobe when no VB is available.
Automated VB navigation system will show the VB view in synchrony to the movement of real-time bronchoscopic image on the monitor, the VB in our system is unable to navigate simultaneously with the bronchoscopic images.13,25 To use OsiriX-generated VB efficiently, bronchoscopists must be familiar with the anatomy of the peripheral airway, at least to the subsegmental level (fourth generation). The OsiriX-generated VB provides a preview of the path to the target before procedure. When choosing the fifth generation bronchus and beyond to insert EBUS probe for target confirmation, immediate reference to the VB images may be necessary. VB images were produced to a median of fifth generation where the bronchoscope could hardly be advanced further in real life. Frequent references back to VB images were thus not necessary. For inexperienced bronchoscopists, anatomy of the airway branches can be repeatedly studied to quicken the learning process on a computer without causing discomfort to patients using this open source software.
This is the first report on using OsiriX-generated VB to assist EBUS bronchoscopy. The software works on most Macintosh platform. OsiriX can be downloaded free of charge and the cost of a FDA-cleared version had been around USD 600 to 700. Unlike expensive commercial VB systems, it is affordable and is readily available to many countries outside North America and Europe. It allows VB to be conveniently studied in laptop computers and be reviewed. This pilot study is not a comparison between different VB systems. Our study provided preliminary data showing that compared with no VB assistance, VB constructed by this software could shorten procedure time without compromising diagnostic yield or safety in EBUS-guided bronchoscopy. Further studies are warranted to confirm our findings.
The authors would like to acknowledge the Medical Board, Tung Wah Group of Hospitals, for funding the equipment in our bronchoscopy suite.
1. Aberle DR, Adams AM, Berg CD, et al..Reduced lung-cancer mortality with low-dose computed tomographic screening.N Engl J Med.2011;365:395–409.
2. Rivera MP, Mehta AC.Initial diagnosis of lung cancer
: ACCP evidence-based clinical practice guidelines (2nd edition).Chest.2007;132suppl 3131S–148S.
3. Baaklini WA, Reinoso MA, Gorin AB, et al..Diagnostic yield of fiberoptic bronchoscopy
in evaluating solitary pulmonary nodules.Chest.2000;117:1049–1054.
4. Wallace JM, Deutsch AL.Flexible fiberoptic bronchoscopy
and percutaneous needle lung aspiration for evaluating the solitary pulmonary nodule.Chest.1982;81:665–671.
5. Roth K, Hardie JA, Andreassen AH, et al..Predictors of diagnostic yield in bronchoscopy
: a retrospective cohort study comparing different combinations of sampling techniques.BMC Pulm Med.2008;8:2.
6. 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;126:959–965.
7. Yu KC, Gibbs JD, Graham MW, et al..Image-based reporting for bronchoscopy
.J Digit Imaging.2010;23:39–50.
8. Rosset A, Spadola L, Ratib O.OsiriX: an open-source software for navigating in multidimensional DICOM images.J Digit Imaging.2004;17:205–216.
9. Heuberger J, Rosset APixmeo Sarl.The 3D endoscopy.OsiriX User Manual: The Complete Reference.Switzerland:Pixmeo;197–202.
10. Dolina MY, Cornish DC, Merritt SA, et al..Interbronchoscopist variability in endobronchial path selection: a simulation study.Chest.2008;133:897–905.
11. Merritt SA, Gibbs JD, Yu KC, et al..Image-guided bronchoscopy
for peripheral lung lesions: a phantom study.Chest.2008;134:1017–1026.
12. Asano F, Matsuno Y, Matsushita T, et al..Transbronchial diagnosis of a pulmonary peripheral small lesion using an ultrathin bronchoscope with virtual bronchoscopic navigation.J Bronchol.2002;9:108–111.
13. Asano F, Matsuno Y, Shinagawa N, et al..A virtual bronchoscopic navigation system for pulmonary peripheral lesions.Chest.2006;130:559–566.
14. Shinagawa N, Yamazaki K, Onodera Y, et al..Virtual bronchoscopic navigation system shortens the examination time—feasibility study of virtual bronchoscopic navigation system.Lung Cancer
15. Shinagawa N, Yamazaki K, Onodera Y, et al..CT-guided transbronchial biopsy using an ultrathin bronchoscope with virtual bronchoscopic navigation.Chest.2004;125:1138–1143.
16. Osborne D, Vock P, Godwin JD, et al..CT identification of bronchopulmonary segments: 50 normal subjects.Am J Roentgenol.1984;142:47–52.
17. Ishida T, Asano F, Yamazaki K, et al..Virtual bronchoscopic navigation combined with endobronchial ultrasound to diagnose small peripheral pulmonary lesions: a randomised trial.Thorax.2011;66:1072–1077.
18. Oshige M, Shirakawa T, Nakamura M, et al..Clinical application of virtual bronchoscopic navigation system for peripheral lung lesions.J Bronchol Interv Pulmonol.2011;18:196–202.
19. Asahina H, Yamazaki K, Onodera Y, et al..Transbronchial biopsy using endobronchial ultrasonography
with a guide sheath and virtual bronchoscopic navigation.Chest.2005;128:1761–1765.
20. Asano F, Matsuno Y, Tsuzuku A, et al..Diagnosis of peripheral pulmonary lesions using a bronchoscope insertion guidance system combined with endobronchial ultrasonography
with a guide sheath.Lung Cancer
21. Herth FJ, Ernst A, Becker HD.Endobronchial ultrasound-guided transbronchial lung biopsy in solitary pulmonary nodules and peripheral lesions.Eur Respir J.2002;20:972–974.
22. Kikuchi E, Yamazaki K, Sukoh N, et al..Endobronchial ultrasonography
with guide-sheath for peripheral pulmonary lesions.Eur Respir J.2004;24:533–537.
23. Gildea TR, Mazzone PJ, Karnak D, et al..Electromagnetic navigation diagnostic bronchoscopy
: a prospective study.Am J Respir Crit Care Med.2006;174:982–989.
24. Schwarz Y, Greif J, Becker HD, et al..Real-time electromagnetic navigation bronchoscopy
to peripheral lung lesions using overlaid CT images: the first human study.Chest.2006;129:988–994.
25. Eberhardt R, Anantham D, Herth F, et al..Electromagnetic navigation diagnostic bronchoscopy
in peripheral lung lesions.Chest.2007;131:1800–1805.
26. Wang Memoli JS, Nietert PJ, Silvestri GA.Meta-analysis of guided bronchoscopy
for the evaluation of the pulmonary nodule.Chest.2012;142:385–393.
27. Gould MK, Fletcher J, Iannettoni MD, et al..Evaluation of patients with pulmonary nodules: when is it lung cancer
?: ACCP evidence-based clinical practice guidelines (2nd edition).Chest.2007;132suppl 3108S–130S.
28. Horton KM, Horton MR, Fishman EK.Advanced visualization of airways with 64-MDCT: 3D mapping and virtual bronchoscopy
.Am J Roentgenol.2007;189:1387–1396.
Keywords:© 2014 by Lippincott Williams & Wilkins.
virtual bronchoscopy; peripheral pulmonary lesion; endobronchial ultrasonography; bronchoscopy; lung cancer