Published ahead of print on March 22, 2007, doi:10.1164/rccm.200612-1866OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200612-1866OC
Multimodality Bronchoscopic Diagnosis of Peripheral Lung LesionsA Randomized Controlled Trial1 Department of Pneumology and Critical Care Medicine, Thoraxklinik, University of Heidelberg, Heidelberg, Germany; and 2 Interventional Pulmonology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts Correspondence and requests for reprints should be addressed to Armin Ernst, M.D., Chief, Interventional Pulmonology, Beth Israel Deaconess Medical Center, Harvard Medical School, 330 Brookline Avenue, Boston MA 02215. E-mail: aernst{at}bidmc.harvard.edu
Rationale: Endobronchial ultrasound (EBUS) and electromagnetic navigation bronchoscopy (ENB) have increased the diagnostic yield of bronchoscopic diagnosis of peripheral lung lesions. However, the role of combining these modalities to overcome each individual technique's limitations and, consequently, to further increase the diagnostic yield remains untested. Objectives: A prospective randomized controlled trial involving three diagnostic arms: EBUS only, ENB only, and a combined procedure. Methods: All procedures were performed via flexible bronchoscopy and transbronchial forceps biopsies were obtained without fluoroscopic guidance. In the combined group, after electromagnetic navigation, the ultrasound probe was passed through an extended working channel to visualize the lesion. Biopsies were taken if ultrasound visualization showed that the extended working channel was within the target. Primary outcome was diagnostic yield. The reference "gold standard" was a surgical biopsy if bronchoscopic biopsy did not reveal a definite histological diagnosis compatible with the clinical presentation. Secondary outcomes were yields by size, lobar distribution, and lesion pathology. Complication rates were also documented. Measurements and Main Results: Of the 120 patients recruited, 118 had a definitive histological diagnosis and were included in the final analysis. The diagnostic yield of the combined procedure (88%) was greater than EBUS (69%) or ENB alone (59%; p = 0.02). The combined procedure's yield was independent of lesion size or lobar distribution. The pneumothorax rates ranged from 5 to 8%, with no significant differences between the groups. Conclusions: Combined EBUS and ENB improves the diagnostic yield of flexible bronchoscopy in peripheral lung lesions without compromising safety.
Key Words: electromagnetic navigation bronchoscopy • ultrasound, interventional • solitary pulmonary nodule • transbronchial lung biopsy
Endobronchial ultrasound (EBUS) and electromagnetic navigation bronchoscopy (ENB) have increased the yield of flexible bronchoscopy in the diagnosis of peripheral lung lesions and solitary pulmonary nodules. The reported sensitivity of flexible bronchoscopy for the diagnosis of peripheral bronchogenic carcinoma ranges from 36 to 86% and is dependent on size (1, 2). Diagnostic yields for EBUS using a radial probe have been reported to be 58.3 to 80% (Table 1) (3–9), whereas ENB has reported yields of 69 to 74% (Table 2) (10–12). The yields of both these procedures are independent of lesion size (3–8, 12).
EBUS enables direct visualization of the target lesion before attempting biopsy. However, EBUS lacks a navigation system and requires the operator to maneuver the bronchoscope blindly to the lesion with the knowledge of prior radiological investigations like computed tomography (CT) scans. In previous studies, 11 to 24% of lesions could not be localized by EBUS (3, 5, 7, 9). ENB consists of four components: an electromagnetic location board, a locatable sensor probe with an eight-way steering mechanism that is able to navigate the bronchial tree, an extending working channel (EWC) that can carry either the sensor probe or a flexible forceps, and computer software that converts CT scans into multiplanar images with three-dimensional virtual bronchoscopy reconstruction. This system enables real-time navigation guidance within the lungs to endobronchially invisible targets and subsequent biopsy through the EWC. However, biopsies using ENB have not always resulted in a diagnosis despite accurate navigation in the vast majority of cases to within 10 mm of the target center (10–12). Respiratory variations causing larger than anticipated navigation errors (10) and dislodgement of the EWC when biopsy instruments were introduced (12) may account for this lower than expected diagnostic yield. ENB lacks a means to directly visualize lesions before biopsy. The role of combining EBUS with ENB to gain the benefits and minimize the limitations of either technique has never been reported. We performed a prospective randomized controlled trial comprising three arms with EBUS only, ENB only, and combined EBUS/ENB to test this hypothesis.
We recruited 120 patients who were referred to the interventional pulmonology service at our centers between January 2003 and August 2006. Inclusion criteria were subjects above the age of 18, who signed informed consent and who were candidates for elective bronchoscopy or surgery. All subjects had evidence of peripheral lung lesions or solitary pulmonary nodules on CT scans. Peripheral pulmonary lesions were defined as lesions that are surrounded by normal lung parenchyma without any CT evidence of endobronchial abnormalities. Pregnant patients and those with implantable pacemakers or defibrillators were excluded because of their untested nature in ENB. Randomization was achieved through a computer-generated random list. Both institutional review boards of the participating centers (Thoraxklinik and BIDMC) approved the data collection and analysis. Primary outcome was diagnostic yield. If transbronchial lung biopsy failed to yield a definitive histological diagnosis that was consistent with the clinical presentation, then patients were referred for a surgical biopsy, which was considered the reference ‘gold standard’. All patients with failed bronchoscopic diagnosis and who were unwilling or unable to have a surgical biopsy were excluded from final analysis to exclude possible confounders. Secondary outcomes included analysis of yield by lesion size, lobar location and lesion pathology (malignant versus benign). Safety of the procedures was documented by tracking all complications. Bronchoscopy was performed in all three diagnostic arms of the study via the oral route using an Olympus IT160 adult therapeutic bronchoscope (Olympus, Tokyo, Japan) with a 2.8-mm working channel. Either moderate sedation or general anesthesia was used at the discretion of the operator. Patients who had general anesthesia were intubated before bronchoscopy. All cases were performed in an outpatient setting.
EBUS Procedure
ENB Procedure The patient was then placed on the electromagnetic location board (470 x 560 mm). Endobronchial mapping was achieved when the virtual fiducial registration points were linked to the actual position in the patient's thorax by a sensor probe (outer diameter, 1.9 mm). The software then documented the registration error, which represents the radius of the expected difference in location between the tip of the sensor probe in the actual patient and where the tip is expected to be. The registration error could then be reduced by either repositioning a misplaced landmark or by eliminating the landmarks with the greatest deviation. A registration error of 6 mm or less was considered acceptable. After endobronchial inspection, navigation began by wedging the bronchoscope in the suspected bronchial segment and steering the sensor probe with the EWC to the lesion using the multiplanar CT images and the "tip-view" orientation. The EWC has a working length of 945 mm and requires a minimum bronchoscope instrument channel width of 2.0 mm. After navigation to the lesion was complete, specimens were obtained through the EWC by transbronchial forceps biopsy.
Combined EBUS/ENB Only forceps biopsies were taken in all three study arms and fluoroscopy was not used to guide transbronchial biopsies. Chest radiographs were taken after all procedures to exclude iatrogenic pneumothorax. Rapid onsite cytopathological evaluation was not used.
Data Analysis
Of the 120 recruited patients, 2 patients with a nondiagnostic bronchoscopic procedure declined surgical biopsy and were excluded from analysis. Both of these patients showed radiological stability in the size of lesion on clinical follow-up. Among the remaining 118 patients, 85 (72%) had a positive diagnostic result via bronchoscopy and the remaining 33 required a subsequent surgical biopsy to establish histological diagnosis. Although pulmonary function testing was not performed on all patients, these patients were all evaluated to be good surgical candidates by our thoracic surgeons. Females accounted 42% of the 118 patients. The mean age was 53 ± 13 years, with a range of 19 to 81 years. No endobronchial lesions were seen in any of the patients. The overall prevalence of malignancy in our study was 78% (92/118). The mean number of biopsies taken was 4.1 ± 0.8, with a range of 2 to 5. There were no statistically significant differences among the three diagnostic arms of the study in terms of patient baseline characteristics, number of biopsies taken, type of sedation, lobar distribution of lesions, or pathology (i.e., malignant vs. benign) (Table 3). The final histological diagnoses are listed in Table 4 together with the respective bronchoscopic yields.
The mean lesion size was 26 ± 6 mm, with a range of 13 to 58 mm. The mean lesion size was larger in the group who underwent ENB only (28 ± 8 mm) compared with EBUS only (25 ± 5 mm) or combined EBUS/ENB (24 ± 5 mm; p = 0.01). However, there were no significant differences in size distribution across all three study arms when lesion size was classified as less than 20, 20 to 30, and more than 30 mm (p = 0.19).
Diagnostic Yield
EBUS and combined EBUS/ENB had diagnostic yields that were independent of lesion size and lobar distribution. Although the results of ENB alone were also independent of lesion size, the yields from the lower lobes were significantly lower (29%; p = 0.01).
Safety
Multimodality investigation by combining EBUS with ENB enhances the diagnostic yield of flexible bronchoscopy in peripheral lung lesions compared with either procedure alone. The improved yield is unaffected by lesion size or lobar distribution. The 69% yield that was achieved in our EBUS-alone group was comparable to previous studies despite not using fluoroscopic guidance (3–9). Although our ENB-alone yield appears to be marginally lower (59 vs. 69–74%) compared with historical data, this may be attributable to factors other than fluoroscopy (10–12). In previous data, surgical biopsies were not always performed after nondiagnostic ENB procedures (12). Therefore, the definition of what constitutes a positive or negative yield is questionable. Furthermore, these studies used endobronchial brushes together with forceps, which could also have enhanced the yield (10–12). The improved yield of the joint procedure is attributed to combining the ability of EBUS to directly visualize the internal structure of peripheral lung lesions with the precise navigation capabilities of ENB. As an adjunct to ENB, EBUS is superior to fluoroscopy in the detection of "fluoroscopically invisible" small lesions (9) and has the added advantage of being radiation free. Conversely, ENB enhances EBUS by providing real-time and subtle navigation through the steering mechanism of the locatable guide. This navigation capability is better than that afforded by either fluoroscopy (4, 5, 7) (yield, 58.3–77.3%), curettes (4, 5, 7) (yield, 58.3–77.3%), or virtual bronchoscopy (7) (yield, 63.3%). By guiding the EBUS probe to within the lesion rather than adjacent to it, yield is improved (4). The strength of this study is that a gold-standard diagnosis was achieved in all analyzed patients when bronchoscopic results were inconclusive. This removed any ambiguity over how positive and negative yields were defined. By restricting biopsy technique to only forceps biopsy and not using other tools, such as needle, brush, and washings, possible confounders in yield and complications were eliminated. Although these biopsy techniques would be used in different combinations in usual clinical practice, varying biopsy instruments, number of passes, volume of lavage (injected and aspirated), and specimen handling might have introduced too many variables into our study. Although there was a statistically significant difference in the size of the lesions in the ENB-alone arm of this study, the 3- to 4-mm difference in mean size may be clinically irrelevant. Furthermore, there was no difference in the distribution of lesion size across the three arms. Our results also confirm data collected from previous studies that yields of both EBUS and ENB are independent of size (3–8, 12). Hence, lesion size is unlikely to confound our findings. The greatly diminished lower lobe yield of 29% in the ENB-alone arm could be attributed to navigation error. Navigation in the lower lobes may be more affected by diaphragmatic movement during breathing. This is because the planning data are based on CT images acquired in a single breath hold and cannot compensate for respiratory movements (10). There was also no increase in pneumothorax rate by combining EBUS with ENB. The pneumothorax rate in either EBUS (5%) or ENB (5%) was not greatly increased compared with that reported in previous studies (0–4.2%) (5, 7, 12). Fluoroscopy was not used in this study because earlier data have shown that it does not decrease the rate of iatrogenic pneumothorax after transbronchial lung biopsy using regular flexible bronchoscopy (13, 14). Moreover, previous studies on EBUS-guided diagnosis of peripheral lung lesions did not use fluoroscopy because the target lesions were small and would have been difficult to visualize on fluoroscopy (9). The other disadvantages of fluoroscopy are radiation exposure and space constraints in the bronchoscopy suite. Furthermore, the 8% pneumothorax rate of combined ENB/EBUS compares favorably with the 23 to 38% pneumothorax rate reported in CT-guided transthoracic biopsy (15, 16). Significant endobronchial bleeding was also not encountered. The EWC, which enables multiple biopsy samples to be taken from the same area after navigation, also facilitates tamponade of potential bleeding by allowing the scope to remain wedged at the subsegmental bronchi throughout the biopsy process (5, 7). There are possible concerns that, in combined EBUS/ENB, the procedure duration may be extended. Although procedure duration was not studied in this trial, no adverse events related to sedation or anesthesia were documented. Previous data comparing EBUS-directed biopsy with fluoroscopically directed biopsy show only a marginal increase in biopsy time (9.8 vs. 8.1 min) (8). In ENB, the average reported registration times are 2 to 3 minutes and navigation times are about 7 minutes (10, 12). There may also be some time saved in ENB if C-arm fluoroscopy is not used. Therefore, the combined EBUS/ENB procedure is likely to require additional time, but this may only be a marginal increase. Before embarking on advanced diagnostic bronchoscopy, issues of costs and training need to be addressed. A detailed cost-analysis model is beyond the scope of this study. At the time this manuscript was prepared, the EBUS EU-M30S-K processor cost $31,000 and the UM-BS20–26R-3 radial probes cost $6,250. This probe can be used for approximately 50 cases. The superDimension/Bronchus system for ENB costs $129,450 and the disposable sensor probes together with the EWC cost $995. Reimbursement issues for ENB have also not been clarified. The recommended training requirement for EBUS is at least 50 proctored procedures (17). No guidelines are currently available for ENB.
Multimodality diagnosis with the joint use of EBUS with ENB has pushed the diagnostic yield of flexible bronchoscopic procedures closer to the sensitivity obtainable through either transthoracic CT-guided (92%) (1) or surgical (
Micelle Secic, from Secic Statistical Consulting, Inc., assisted in the statistical analysis. Arthur Dea and Robert Garland from Interventional Pulmonology, Beth Israel Deaconess Medical Center, Boston, assisted with manuscript editing and data collection. The locatable sensor probes for electromagnetic navigation bronchoscopy at both Thoraxklinik and Beth Israel Deaconess Medical Center were provided free of charge by superDimension/Bronchus. SuperDimension/Bronchus has supported CME courses at the Thoraxklinik Heidelberg and Harvard University through unrestricted educational grants.
Originally Published in Press as DOI: 10.1164/rccm.200612-1866OC on March 22, 2007 Conflict of Interest Statement: R.E. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. D.A. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.E. is a member of the Scientific Advisory Board of superDimension/Bronchus and has been reimbursed for time and travel expenses related to that function; he had received stock options for superDimension, but has not been involved in the consenting and randomization of patients. D.F.-K. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. F.H. is a member of the Scientific Advisory Board of superDimension/Bronchus and has been reimbursed for time and travel expenses related to that function. Received in original form December 21, 2006; accepted in final form March 22, 2007
Related articles in AJRCCM:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
100%) (