If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Digital single-operator peroral cholangioscopy-guided biopsy sampling versus ERCP-guided brushing for indeterminate biliary strictures: a prospective, randomized, multicenter trial (with video)
Department of Gastroenterology and Hepatology, Radboud University Medical Center, Nijmegen, The NetherlandsAsian Institute of Gastroenterology Hospitals, Hyderabad, India
Accurately diagnosing indeterminate biliary strictures is challenging but important for patient prognostication and further management. Biopsy sampling under direct cholangioscopic vision might be superior to standard ERCP techniques such as brushing or biopsy sampling. Our aim was to investigate whether digital single-operator cholangioscopy (DSOC) compared with standard ERCP workup improves the diagnostic yield in patients with indeterminate biliary strictures.
Methods
Patients with an indeterminate biliary stricture on the basis of MRCP were randomized to standard ERCP visualization with tissue brushing (control arm [CA]) or DSOC visualization and DSOC-guided biopsy sampling (study arm [SA]). This was a prospective, international, multicenter trial with a procedure-blinded pathologist.
Results
The first sample sensitivity of DSOC-guided biopsy samples was significantly higher than ERCP-guided brushing (SA 68.2% vs CA 21.4%, P < .01). The sensitivity of visualization (SA 95.5% vs CA 66.7%, P = .02) and overall accuracy (SA 87.1% vs CA 65.5%, P = .05) were significantly higher in the SA compared with the CA, whereas specificity, positive predictive value, and negative predictive value showed no significant difference. Adverse events were equally low in both arms.
Conclusions
DSOC-guided biopsy sampling was shown to be safe and effective with a higher sensitivity compared with standard ERCP techniques in the visual and histopathologic diagnosis of indeterminate biliary strictures. (Clinical trial registration number: NCT 03140007.)
Biliary strictures arise from a range of entities from benign (eg, postsurgical or stone disease) to inflammatory (eg, sclerosing or IgG4-related cholangitis) to malignant conditions (eg, cholangiocarcinoma, hepatocellular carcinoma, or local extension of other tumors).
Because of the major differences in treatment comprising endoscopic intervention, medical treatment, extensive surgery, and palliation, a correct diagnosis, preferably at first ERCP, is crucial for patient prognosis and management. For example, although a benign postsurgical biliary stricture can often be treated successfully with an endoscopic approach, a cholangiocarcinoma needs aggressive surgical and oncologic management associated with appreciable morbidity and not insignificant mortality.
The standard evaluation of indeterminate biliary strictures therefore always involves cross-sectional imaging followed by ERCP plus biopsy sampling and/or brushing.
Transpapillary biopsy sampling and/or brushing (TPB) is easy to perform and fairly inexpensive but is limited by well-documented poor diagnostic accuracy.
Comparative effectiveness of biliary brush cytology and intraductal biopsy for detection of malignant biliary strictures: a systematic review and meta-analysis.
EUS-guided sampling is considered an additional diagnostic approach if TPB fails but is limited to distal or unresectable intrinsic biliary lesions, and seeding remains a potential concern.
Digital single-operator cholangioscopy (DSOC) is an established tool that provides high-resolution direct visualization of the bile duct, allows for histopathologic acquisition, and also enables interventional therapy.
DSOC appears to be a logical next step to increase diagnostic accuracy in indeterminate biliary strictures. Indeed, several cohort studies and a meta-analysis reported significant superior diagnostic yield and potential cost-effectiveness of DSOC compared with TPB.
The economic impact of using single-operator cholangioscopy for the treatment of difficult bile duct stones and diagnosis of indeterminate bile duct strictures.
Diagnostic accuracy of conventional and cholangioscopy-guided sampling of indeterminate biliary lesions at the time of ERCP: a prospective, long-term follow-up study.
Comparative effectiveness of biliary brush cytology and intraductal biopsy for detection of malignant biliary strictures: a systematic review and meta-analysis.
To date no prospective, randomized studies have compared standard ERCP workup (comprising TPB) with DSOC. Our aim was to investigate whether DSOC compared with standard ERCP workup improves the diagnostic yield in patients with indeterminate biliary strictures.
Methods
The objective of this study was to assess the diagnostic accuracy of DSOC using single-use cholangioscope technology (SpyGlassDS; Boston Scientific Corp, Marlboro, Mass, USA) compared with ERCP-based assessment using traditional sampling technology in patients with indeterminate biliary strictures. Indeterminate biliary stricture was defined as a suspected intrinsic biliary stricture of unknown entity (benign or malignant) based on prior MRCP. The study design was a prospective, international, multicenter, randomized controlled trial. This was an investigator-initiated study. The study was approved by the Institutional Review Board at each site and was registered (clinicaltrials.gov NCT03140007). All investigators agreed on the study protocol and could only access data from their own study center. All proceduralists were highly experienced operators performing over 200 ERCPs and over 30 SpyGlassDS per year. All authors reviewed the final manuscript.
Participants
Inclusion criteria were age ≥ 18 years, willing and able to comply with the study procedures and provide written informed consent to participate in the study, biliary obstructive symptoms, and indeterminate biliary stricture suspected to be intrinsic and proximal to the distal common bile duct based on prior MRCP. Exclusion criteria were contraindications to endoscopic intervention (coagulopathy, American Society of Anesthesiology classification 4 or greater, pregnancy), prior ERCP with TPB for assessment of indeterminate biliary stricture, extrabiliary compression identified on prior noninvasive imaging and believed to be the cause of the biliary obstructive symptoms, and age less than 18 years.
A local multidisciplinary team made up of experienced radiologists, surgeons, and endoscopists assessed the indication for endoscopic evaluation of the indeterminate biliary stricture. All imaging studies were reviewed by a dedicated hepatobiliary-certified radiologist. Treatment strategies were standardized across sites. Suspected intrinsic biliary strictures except distal common bile duct strictures were included in order to exclude patients with extrinsic biliary obstruction from a pancreatic mass. According to the MRCP, strictures were classified based on the Bismuth-Corlette classification (Fig. 1).
Patients were invited to participate in the study if they were eligible based on inclusion and exclusion criteria. Consecutive patients referred for assessment and management of their indeterminate biliary stricture were screened for inclusion. Informed consent was obtained from all patients and documented before enrollment and randomization. Before the procedure, relevant demographic information and patient medical history were recorded. Data were collected and stored securely by a local study nurse at each participating center in written form. It was subsequently sent to a central and secure, anonymous online database as a case report file.
Procedure
Patients in the control arm (CA) underwent an ERCP with cholangiography. Cholangiography-based impression of malignancy (yes, no, or indeterminate) was recorded. ERCP-guided brushing was performed with a minimum of 9 passes. The samples were fixed on a slide and sent off for cytologic analysis. The brushes used for this study were RX Cytology Brush 2.1 mm × 8F (Boston Scientific Corp) or 3-mm/6F-8F Cytomax II Double Lumen Cytology Brush (Cook Medical, Bloomington, Ind, USA).
Patients in the study arm (SA) underwent a DSOC (SpyGlassDS; Boston Scientific Corp) cholangioscopic evaluation without cholangiography. DSOC-based impression of malignancy (yes, no, or indeterminate) was recorded. A minimum of 3 DSOC-guided biopsy samples were required to be counted as adequate. Biopsy specimens were taken with a purpose-designed biopsy forceps that can be passed through the DSOC miniscope (SpyBite; Boston Scientific Corp) and sent off for histopathologic analysis in a jar. Representative images can be found in Figure 2.
Figure 2Top, MRCP of indeterminate biliary stricture. Middle, Digital single-operator cholangioscopy showing a stricture with irregular papillogranular surface, luminal contents (mass), and irregular and tortious vessels. Bottom, Histopathology with tumor cells.
Histopathologic or cytologic assessment was performed by a histopathologist with expertise in biliary histopathology at each center. The pathologist was blinded to the randomization status of the study participants. Procedural adverse events were documented, including severity and seriousness of the events, time of onset and resolution, and outcome of the events. All patients underwent a follow-up 30 days after the index procedure with a visit and/or a telephone call. Follow-up ended 6 months after the index procedure or confirmation of malignancy, whichever came first. All equipment was used on label.
Outcomes
For the primary endpoint, diagnostic accuracy of cholangioscopy or cholangiography was assessed at 6 months after the initial ERCP procedure. Malignancy was determined by cytology or histology on tissue sampling during the index or subsequent ERCP procedure or by surgical specimen histopathology up to 6 months after the index procedure. The assessed stricture was considered benign if malignancy was not confirmed by 6 months after the index procedure. Diagnostic accuracy of only the first procedure was compared between the groups, because subsequent information was not easily comparable.
There were 4 secondary endpoints. First, for the correlation between intraprocedural visual impression of malignancy and cytopathology in the CA and SA, visual impression was consistently based on previous described criteria (mass, dilated tortuous vessels, papillary or villous projections, intraductal nodules) and documented as suspected malignant or benign lesion.
Second, technical success of the procedure was defined as the ability to collect tissue deemed adequate for cytologic or histologic analysis as per the pathologist. Third, the yield of tissue acquisition was defined as benign or malignant cytology or histology results from ERCP-guided brushing and cholangioscopy-guided SpyBite biopsy samples. Fourth, the occurrence and severity of procedure-related serious adverse events from index procedure through 30 days after procedure (hospitalization and intensive care unit admissions)were recorded.
If patients had an initial benign histopathology but the visual impression of the proceduralist was suspicious for malignancy, patients would be rescheduled for a prompt re-evaluation with an additional tissue sampling. If the first histopathology was indeterminate, regardless of the visual impression, a prompt re-evaluation with an additional tissue sampling was scheduled. In these cases any type of procedure was applicable, because it would be ethically incorrect for patients to stay only in 1 arm and not use all possible procedures to get histopathology. Procedural parameters such as procedure duration (duodenoscope in to duodenoscope out) and number of procedures were recorded in both arms. Change in preprocedural patient management as a result of ERCP cholangiography or DSOC was also recorded (impact of visual impression).
Randomization
Patients were randomized in a 1:1 ratio into 2 groups: the CA/ERCP (ERCP impression and ERCP-guided brushing) and the SA/cholangioscopy (DSOC impression and DSOC-guided biopsy sampling). Randomization was conducted using study envelopes provided to each center. The randomization scheme was followed by site block randomization and was prepared by a statistician. Participants and endoscopist were blinded to randomization until the study nurse opened the envelope, revealing the allocated group before commencing the procedure.
Statistics
We tested the hypothesis that cholangioscopy-guided biopsy sampling would have a higher sensitivity than ERCP-guided brushing. We assumed that ERCP-guided brushing would have a 48% sensitivity and that cholangioscopic-guided biopsy sampling would have an 82% sensitivity. We powered the hypothesis at 80%, with a 2-sided alpha of .05, using an asymptotic test in which the number of patients required was 54. We added 10% attrition for possible loss to follow-up and dropouts to get a total sample size of 60.
Continuous data were analyzed using median with interquartile range, and differences between the groups were analyzed using a Mann-Whitney U test. However, because the number of procedures is a count variable, it was reported with mean and standard deviation and was tested using a negative binomial model. Binary variables were analyzed using frequencies, and differences were tested using asymptotic methods, except for the correlation between visualization and biopsy sampling, which was calculated using the Fisher exact test. P < .05 was considered significant, and all analyses and sample size calculations were performed using SAS version 9.4 (SAS Institute, Inc, Cary, NC, USA). Diagnostic accuracy was derived from the following components: sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy.
Results
Sixty-one patients in 3 tertiary referral centers were consecutively screened and enrolled from May 2017 to December 2018. The study was ended after reaching the primary endpoint and ended at 6 months follow-up of the last enrolled patient. Thirty-two patients were randomized to the SA and 29 to the CA. One patient in the SA was excluded from all analyses because he or she withdrew consent before the procedure was performed. The flow of participants is demonstrated in Figure 3. Demographic and baseline characteristics were equally distributed in both arms (Table 1).
In the SA 1 patient was excluded because of confirmation of extrinsic biliary compression after undergoing DSOC. This left 30 patients for the primary endpoint analysis, of which 22 patients had a malignant final diagnosis and 8 patients had a benign final diagnosis. In the CA 1 patient was excluded because of an inability to reach the duodenum due to an obstructing esophageal cancer as well as another patient in whom brushings were not performed after removal of an obstructing common bile duct stone. This left 27 patients for primary endpoint analysis, of which 14 patients had a malignant final diagnosis and 13 patients had a benign final diagnosis. In the CA, 44.8% (13/29) completed the study without malignancy and 48.2% (14/29) with malignancy. In the SA, 25% (8/32) completed the study without malignancy and 68.8% (22/32) with malignancy (Fig. 3).
Most patients received preinterventional antibiotics (CA 86.2% vs SA 96.8%; P = .14). Most patients underwent a sphincterotomy (CA 69% vs SA 74.2%; P = .64) and stent placement (CA 81.5% vs SA 77.4%; P = .7) without significant difference in the 2 arms. The mean number of procedures to get a final diagnosis was 1.16 for the CA and 1.18 for the SA (1.17 overall, P = .95). Median duration of the procedure showed no significant difference in both arms (CA 30 minutes [interquartile range, 20-87] vs SA 50 minutes [interquartile range, 26-73]; P = .2) (Table 2).
Table 2General procedure characteristics
TPB group (n = 29)
DSOC group (n = 31)
P value
Preinterventional antibiotics
86.2 (25/29)
96.8 (30/31)
.14
No. of days
2 (1-3)
1 (1-2)
.61
Location of dominant target lesion
.57
IHBD
.0 (0/29)
6.5 (2/31)
Right main
10.3 (3/29)
12.9 (4/31)
Left main
3.4 (1/29)
6.5 (2/31)
Hilar
27.6 (8/29)
29.0 (9/31)
CHD
.0 (0/29)
3.2 (1/31)
Common bile duct
55.2 (16/29)
41.9 (13/31)
Ampullary/sphinteric
.0 (0/29)
.0 (0/31)
Diffuse/multiple locations
3.4 (1/29)
.0 (0/31)
Sphincterotomy performed
69.0 (20/29)
74.2 (23/31)
.65
Stent placed
81.5 (22/27)
77.4 (24/31)
.70
Plastic stent
95.5 (21/22)
91.7 (22/24)
UC self-expandable metal stent
.0 (0/22)
4.2 (1/24)
PC self-expandable metal stent
.0 (0/22)
.0 (0/24)
FC self-expandable metal stent
4.5 (1/22)
4.2 (1/24)
Duration of procedure, min
30 (20-87)
50 (26-73)
.20
Values are %, except Age, which is median.
TPB, Transpapillary brushing; DSOC, digital single-operator cholangioscopy; IHBD, intra-hepatic bile duct; CH, common hepatic duct; UC, uncovered; PC, partially covered; FC, fully covered.
Tissue samples could be obtained in all cases at first evaluation. In the SA 1 of 31 samples was inadequate for histologic analysis, resulting in a yield for DSOC of 30 of 31 (96.8%). In this case, biopsy samples were inadequate because only 2 biopsy samplings were performed because the lesion seemed to be pulsating on DSOC imaging. On average 6 biopsy specimens were taken (range, 2.0-8.0). In the CA, brushings were adequate for cytologic analysis in all patients. In 1 patient in the CA, the proceduralist suspected a stone as a reason for the stricture with no wire passage possible. In this case DSOC was used and visualized a stone that was not seen on prior imaging and was successfully treated with DSOC-guided electrohydraulic lithotripsy. No brushings were performed in this case. Sensitivity of DSOC-guided biopsy sampling was significantly higher than ERCP-guided brushing (SA 68.2% vs CA 21.4%; P < .01). All other parameters such as specificity, positive predictive value, negative predictive value, and overall accuracy showed no significant difference (Table 3). Table 4 demonstrates the diagnostic parameters when benign and indeterminate tissue samples were combined. The main difference is that in this scenario of malignant versus nonmalignant distinction, the specificity of both TPB and DSOC increased to 100%.
Table 3Diagnostic accuracy of ERCP brushing (TPB) versus DSOC biopsy sampling on first sample
Lesion diagnosis based on ERCP brush
Final diagnosis
Lesion diagnosis based on DSOC biopsy sampling
Final diagnosis
Malignant
Benign
Malignant
Benign
Malignant
3
0
Malignant
15
0
Benign
6
11
Benign
6
5
Indeterminate
5
2
Indeterminate
1
3
Measure
TPB group
DSOC group
P value
Sensitivity
21.4 (3/14)
68.2 (15/22)
<.01
Specificity
84.6 (11/13)
62.5 (5/8)
.25
Positive predictive value
100.0 (3/3)
100.0 (15/15)
.99
Negative predictive value
64.7 (11/17)
45.5 (5/11)
.31
Overall accuracy
51.9 (14/27)
66.7 (20/30)
.25
Values are %, except Age, which is median.
TPB, Transpapillary brushing; DSOC, digital single-operator cholangioscopy.
All lesions in the SA could be visualized. In approximately half of the cases in both arms the visual impression had an impact on patient management (CA 50% vs SA 56.7%; P = .62). The sensitivity (SA 95.5% vs CA 66.7%; P = .02) and overall accuracy (SA 87.1% vs CA 65.5%; P = .05) were significantly higher in the SA compared with the CA, whereas specificity, positive predictive value, and negative predictive value showed no significant difference (Table 5). In the SA a significant correlation between visualization and biopsy sampling could be shown (P = .02) (Table 6).
Table 5Diagnostic accuracy of ERCP visual impression versus DSOC visual impression
Measure
TPB group
DSOC group
P value
Sensitivity
66.7 (10/15)
95.5 (21/22)
.02
Specificity
64.3 (9/14)
66.7 (6/9)
.91
Positive predictive value
90.9 (10/11)
100.0 (21/21)
.16
Negative predictive value
75.0 (9/12)
85.7 (6/7)
.58
Overall accuracy
65.5 (19/29)
87.1 (27/31)
.05
Values are %, except Age, which is median.
TPB, Transpapillary biopsy sampling and/or brushing; DSOC, digital single-operator cholangioscopy.
No severe adverse events were documented. Mild adverse events occurred in 3 patients in the CA (1 cholangitis, 1 cholecystitis, and 1 bleeding) and in 2 patients in the SA (2 pancreatitis). All adverse events were treated conservatively with full recovery of the patients (Table 7).
Table 7Adverse events
TPB group
DSOC group
P value
Overall
10.3 (3/29)
6.5 (2/31)
.59
Cholangitis
3.4 (1/29)
.0 (0/31)
.30
Cholecystitis
3.4 (1/29)
.0 (0/31)
.30
Pancreatitis
.0 (0/29)
6.5 (2/31)
.16
Bleeding
3.4 (1/29)
.0 (0/31)
.30
Values are %, except Age, which is median.
TPB, Transpapillary biopsy sampling and/or brushing; DSOC, digital single-operator cholangioscopy.
Our study found that compared with standard ERCP-based tissue acquisition, DSOC-guided evaluation had a higher sensitivity in the histologic diagnosis of indeterminate biliary strictures. Furthermore, DSOC-based visual impression of the indeterminate biliary stricture provided a higher diagnostic accuracy, sensitivity, and positive predictive value compared with conventional ERCP cholangiography without a difference in adverse events.
Accurate diagnosis of indeterminate biliary strictures is of paramount importance because of a potentially vastly different prognosis based on etiology. Because the diagnostic accuracy of conventional ERCP-based tissue acquisition (brushing and/or biopsy sampling) has been suboptimal, DSOC has emerged as a logical innovation that could have a role in the diagnosis of indeterminate biliary strictures. Our results confirm that DSOC is technically successful and safe for biliary visualization and collection of histopathology.
Compared with the current standard of care in most endoscopy units using brushing cytology, DSOC-directed biopsy sampling could improve diagnostic sensitivity significantly. In our study the sensitivity with DSOC was 68% compared with 21% with ERCP-guided brushing. A reason for the very low sensitivity of brushing in the CA can be explained through the intrinsic entity of the lesions. Cholangiocarcinomas have a prominent desmoplastic and hypervascularized stroma. This pathophysiology is likely the reason why nontargeted brushing of strictures is ineffective in those particular cases and highlights the potentially low ERCP-based tissue acquisition sensitivity in a prospective real-life study environment. Although DSOC-guided sensitivity is higher, it is still somewhat limited and may account for a similar number of procedures needed to confirm malignancy compared with conventional ERCP. The slight imbalance in the final diagnosis with a greater proportion of malignancy in the study arm may also have contributed to this. The impact of the size of the DSOC-guided biopsy forceps is currently unknown. It is conceivable that larger forceps could further improve histologic diagnostic accuracy because of a more suitable tissue sample with the benefit of direct visualization. The mean number of procedures to get to a final diagnosis showed no significant difference in both arms. The reason for this lies in the nature of the disease. In some cases patients either died from their cholangiocarcinoma or went into palliative care after the first procedure. This may have reduced the number of procedures to reach the final diagnosis in the ERCP arm significantly.
A significant advantage of DSOC is the ability to directly visualize the biliary tree and stricture. This is important because it gives the experienced endoscopist further information about the stricture, which has an impact on patient management such as planning further investigation despite negative or indeterminate histopathology results. In fact, 1 case was only clarified after visualizing the stricture using DSOC and finding an impacted stone as the reason for the obstruction. Despite DSOC being potentially a technically more challenging procedure, biliary visualization was successful in all DSOC cases with no statistical difference compared with standard ERCP. As described in other publications, we could reproduce a very high sensitivity and overall accuracy for the diagnosis of malignant strictures using DSOC. Direct visual impression may be convincing enough to result in surgical treatment even without definitive histopathology in select cases, such as surgically fit patients with indeterminate biliary strictures from potentially resectable tumors.
Based on the results of our study, it appears that combining the direct cholangioscopic visual impression with DSOC-guided biopsy sampling provides the highest chance of confirming malignancy in indeterminate biliary strictures. In both groups, visual impression on its own accord made a difference to patient management in at least half of the cohort. It demonstrates that this is an essential component of the diagnostic process and supplements histologic acquisition.
To our knowledge this is the first randomized controlled trial that evaluates DSOC in comparison with conventional ERCP for indeterminate biliary strictures. Strengths include the randomized, prospective nature of the study, 3 international study sites, a defined cohort with a structured follow-up, and the detailed analysis of visual and histopathologic aspects of the procedure.
Our study had several limitations. Even though comparing 60 patients meeting strict inclusion and exclusion criteria for a diagnosis of indeterminate biliary stricture seems to be an adequate number to test the a priori hypothesis set forth in this study, the absolute number of patients was small. The number of participants in each cohort was slightly imbalanced, particularly in regard to containing a diagnosis of malignancy. We believe this would be rebalanced in larger cohorts but in this study may have affected the results. In addition, histopathologic/cytologic analysis was not centrally assessed or compared between sites. Findings of the study may not be generalizable because procedures were performed by experienced expert endoscopists in tertiary referral centers. Although European Society of Gastrointestinal Endoscopy guidelines recommend that indeterminate strictures are assessed and managed in tertiary referral centers, especially when DSOC is indicated, it is unclear whether similar results would be found if less-experienced endoscopists performed these procedures.
We did not perform a cost-effectiveness analysis in this particular study because it was out of the scope of the primary objective. Purchasing a DSOC device is an additional cost to standard ERCP. Evaluating the cost-effectiveness analysis of DSOC and conventional ERCP for indeterminate biliary strictures would be a relevant separate study but is challenging because of different healthcare settings and true costs of interventions like surgery.
In conclusion, DSOC was shown to be safe and effective with a higher diagnostic sensitivity compared with standard ERCP and brushing for the diagnosis of indeterminate biliary strictures. The visual impression was shown to have a pertinent impact on patient management and should be an essential part of determining the correct diagnosis. DSOC should be considered as a viable diagnostic alternative to conventional ERCP for indeterminate biliary strictures with the potential to change patient management (Video 1, available online at www.giejournal.org). Further innovations in cholangioscopic characterization of strictures, image quality, and histologic acquisition devices are desirable and are likely to advance our understanding in this field even further.
Comparative effectiveness of biliary brush cytology and intraductal biopsy for detection of malignant biliary strictures: a systematic review and meta-analysis.
The economic impact of using single-operator cholangioscopy for the treatment of difficult bile duct stones and diagnosis of indeterminate bile duct strictures.
Diagnostic accuracy of conventional and cholangioscopy-guided sampling of indeterminate biliary lesions at the time of ERCP: a prospective, long-term follow-up study.
DISCLOSURE: The following authors disclosed financial relationships relevant to this publication: C. Gerges: Consultant for Microtech Europe GmbH. T. Beyna: Investigator for Boston Scientific– and Olympus-sponsoredtrials; consultant and lecture fees from Boston Scientific. J. Y. W. Lau, M. Ramchandani: Investigator for Boston Scientific–sponsoredtrials. R. Tang: APAC medical advisory board for Boston Scientific. All other authors disclosed no financial relationships. Statistical and databasesupportfor this study was provided by Boston Scientific.
32475-7&id=gr1.jpg){kind=link}
32475-7&id=gr2.jpg){kind=link}
32475-7&id=gr3.jpg){kind=link}