- Email: [email protected]
Efficacy and safety of endoscopic submucosal tunnel dissection for superficial esophageal squamous cell carcinoma: a propensity score matching analysis
Accepted Manuscript Efficacy and safety of endoscopic submucosal tunnel dissection for superficial esophageal squamous cell carcinoma: a propensity score matching analysis Rui Huang, Hongwei Cai, Xin Zhao, Xiaoqiang Lu, Min Liu, Wenhao Lv, Zhiguo Liu, Kaichun Wu, Ying Han PII:
S0016-5107(17)30178-5
DOI:
10.1016/j.gie.2017.03.001
Reference:
YMGE 10447
To appear in:
Gastrointestinal Endoscopy
Received Date: 26 November 2016 Accepted Date: 1 March 2017
Please cite this article as: Huang R, Cai H, Zhao X, Lu X, Liu M, Lv W, Liu Z, Wu K, Han Y, Efficacy and safety of endoscopic submucosal tunnel dissection for superficial esophageal squamous cell carcinoma: a propensity score matching analysis, Gastrointestinal Endoscopy (2017), doi: 10.1016/ j.gie.2017.03.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Efficacy and safety of endoscopic submucosal tunnel dissection for superficial esophageal squamous cell carcinoma: a propensity score matching analysis Rui Huang1*, Hongwei Cai2*, Xin Zhao1*, Xiaoqiang Lu1, Min Liu1, Wenhao Lv1, Zhiguo Liu1, Kaichun Wu1, Ying Han1. Xijing Hospital of Digestive Diseases, Fourth Military Medical University, Xian, China
2
RI PT
1
Information Technology Department, the First Affiliated Hospital of Xi’an Jiaotong University, Xi’an, China *These three authors contributed equally to this work.
M AN U
SC
Corresponding author: Zhiguo Liu MD and Ying Han MD, Xijing Hospital of Digestive Diseases, Fourth Military Medical University, 127 Changle West Road, Xi’an, Shanxi 710032, China. Telephone: +86-29-84771535; Fax: +86-29-82539041; Email: [email protected] and [email protected] Running title: Tunnel ESD technique in superficial esophageal cancer Meeting presentation: Oral presentation in Digestive Disease Week 2016.
EP
TE D
Abbreviations : ASD, absolute standard differences; cESD, conventional ESD; CI, confidence interval; ESCC, esophageal squamous cell carcinoma; ESD, endoscopic submucosal dissection; ESTD, endoscopic submucosal tunnel dissection; HGIN, High-grade intraepithelial neoplasia; MP, muscularis propria; OR, odd ratio; PSM, propensity score matching; pT1a, Intramucosal carcinoma; pT1b, Submucosal invasive carcinoma; STER, submucosal tunnel endoscopic resection, tESD, tunnel ESD.
AC C
Declaration of interest All authors declare no conflicts of interest. The authors alone are responsible for the content and writing of the paper. Author contributions Study concept and design: Zhiguo Liu; Acquisition of data: Rui Huang, Xiaoqiang Lu; Analysis and interpretation of data: Rui Huang, Zhiguo Liu; Drafting of the manuscript: Rui Huang, Zhiguo Liu, Xin Zhao; Critical revision of the manuscript for important intellectual content: Min Liu, Wenhao Lv; Statistical analysis: Hongwei Cai; Administrative support: Kaichun Wu, Ying Han. Acknowledgments This work was supported in part by the National Natural Science Foundation of China (81272449, 1
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
81372387 and 81572820).
2
ACCEPTED MANUSCRIPT Abstract Background and Aims: Esophageal ESD is technically challenging due to the facts of thinner wall and narrow lumen. Tunnel technique was previously proposed. This current retrospective study
RI PT
aimed to evaluate the efficacy of the tunnel technique in ESD of superficial esophageal squamous cell carcinoma (ESCC).
Methods: Patients underwent ESD for superficial ESCC between October 2013 and September 2015
SC
were included in the study. Propensity score matching was used to compensate for the differences in
M AN U
age, sex, resected specimen size, and pathology. Treatment outcomes were compared with conventional statistic methods between tunnel ESD group and conventional ESD group after matching. To further explore the potential variables that relevant to procedure time, univariate and multivariate logistic regression analyses were applied.
TE D
Results: A total of 115 lesions were included in the analysis. Propensity score matching analysis created 38 matched pairs. There was no difference on en bloc resection rate, complete resection rate and the curative rate between the 2 groups. The ESD procedure time was 38.0 (29.5-46.0) minutes in
EP
tunnel ESD group and 48.0 (35.4-83.3) minutes in conventional ESD group (P = 0.006). There was
AC C
no difference on adverse events including post-procedure bleeding rate, perforation and chest pain, but a lower rate of muscular injury (28.9% vs 52.6 %; P = 0.036) and a less-frequent usage of coagulation forceps (36.8% vs 65.8%; P = 0.012) were shown in tunnel ESD group. In multivariate regression analysis for procedure time, the tunnel ESD technique (OR, 3.42; 95 % CI, 1.32 – 8.85; P = 0.011) and specimen size <40 mm (OR, 8.74; 95% CI, 1.30 – 58.5; P = 0.026) were associated with a shorter procedure time. Conclusion: The endoscopic submucosal tunnel dissection improved the efficacy and safety of ESD
3
ACCEPTED MANUSCRIPT procedure by shortening the procedure time and reducing injury to muscular layer.
RI PT
INTRODUCTION Endoscopic submucosal dissection (ESD) is an excellent therapeutic procedure in resecting localized superficial gastrointestinal neoplasms. Although ESD procedure is technically challenging and
SC
requires a longer treatment time than other conventional endoscopic mucosal resection (EMR), ESD
M AN U
enables en bloc resection and lower recurrence rate[1]. However, esophageal ESD is especially more difficult due to the thinner wall and narrow lumen, which could result in adverse events such as perforation and pneumomediastinum [2-4]. A meta-analysis on esophageal showed that the pooled estimate of perforation rate was 5.0% [5]. Meanwhile, for large lesions, conventional ESD is
TE D
time-consuming and suffering an increased risk of adverse events due to the rapid diffusion of submucosal liquid cushion and blocked lumen by the resected mucosa[6, 7]. Previously multiple strategies including dental floss traction have been applied to make the procedure safer and easier [8].
EP
However, the likelihood of clip malfunction and additional procedure steps would tend to increase
AC C
the complexity of the operation.
Recently, the tunnel technique was proposed in esophageal ESD for superficial cancer [6]. In conventional ESD (cESD) procedure, a complete circumferential cutting around the lesion was made before the submucosal dissection. But in tunnel ESD (tESD) technique, after submucosal injection in the usual way, incisions were made at the anal and oral side of the lesion, submucosal tissue was dissected then between anal and oral incisions to form a tunnel underneath the lesion and a clear vision of submucosal layer could be maintained, then two lateral mucosal incisions were made to
4
ACCEPTED MANUSCRIPT complete the ESD procedure. The theological advantage of tunnel technique was to maintain an improved view on working field, efficient submucosal dissection becomes possible, and contributing to shorten procedure time [9]. However, there were few reports to verify the efficacy of this
RI PT
technique [6, 10-12]. The aim of this retrospective study was to evaluate the efficacy and safety of the tunnel technique during ESD of superficial esophageal squamous cell carcinoma (ESCC) by a propensity score
M AN U
SC
matching analysis.
METHODS Patients
During October 2013 to September 2015, patients who were superficial ESCC treated by ESD at
TE D
Xijing Hospital of Digestive Disease (Xi’an, China) were included in our study. The study protocol was approved by the institutional review board of Xijing hospital and all patients provided written
EP
informed consent before the start of study. The data were collected and retrospectively reviewed from medical records, procedure videos and endoscopic reports. All lesions were confirmed by
AC C
histologic evaluation of biopsy specimens obtained from forceps sampling before the ESD procedure or suspected localized mucosal lesions detected by endoscopy according to Japanese classification on esophageal squamous cell carcinoma by Japan Esophageal Society [13]. Recurrent lesion or lesion located in the remnant esophagus after esophagogastrectomy were excluded. Only the largest one was included in the analysis in patients with multiple lesions. ESD procedure All patients were sedated and intubated. ESD procedures were performed by 3 endoscopists (Z.G.L., 5
ACCEPTED MANUSCRIPT R.H., and X.Q.L.) with experience of more than 100 cases on ESD operation, using a single-channel endoscope (GIF-Q260J; Olympus Optical Co, Ltd, Tokyo, Japan). The conventional ESD procedure was as previously described [14]. The procedure of tunnel ESD
RI PT
was showed in Figure 1. After submucosal injection, incisions were made at both the anal and oral side of the lesion, and the submucosa under the lesion is then dissected to create a tunnel between the distal and proximal incisions. When the endoscope has reached the lower incision, the distal end of
SC
the esophageal lumen could be visualized, then 2 lateral mucosal incisions were made, thus
M AN U
completing the ESD procedure. There was no any traction method used in all ESD cases including dental floss traction.
A saline solution containing minimal indigo carmine was used in all patients. Circumferential cutting and submucosal dissection was carried out with O-type HybridKnife (ERBE, Tübingen, Germany),
TE D
the Dual Knife (KD-650L; Olympus) was used as assistance when submucosal fibrosis was severe. Endoscopic hemostasis was performed with the dissecting knife or the hemostatic forceps (FD-410LR; Olympus). After dissection, preventive endoscopic hemostasis was performed for any
AC C
EP
oozing or exposed vessel. The VIO generator (VIO 200D; ERBE) was used for all ESD procedures.
Outcome measures
The main outcome measure in this study was the procedure time. Submucosal dissection time, rates of en bloc resection, complete resection, curative resection and the usage of coagulation forceps during ESD and rates of adverse events including muscular injury, postprocedural bleeding, perforation and chest pain were also retrieved. Furthermore, the factors associated with the procedure time, including the usage of tunnel method, were analyzed in the multivariate analysis in all lesions
6
ACCEPTED MANUSCRIPT before propensity score matching.
Definitions
RI PT
Procedure time was defined as the time from marking of the lesion till complete removal of the specimen including hemostasis. The submucosal dissection time was defined from completion of circumferential cutting till complete removal of the specimen excluding hemostasis in conventional
SC
ESD group. In tunnel ESD group, the submucosal dissection time was defined from incisions are
M AN U
made at the anal and oral side cutting of the lesion to removal of specimen, minus the time of two lateral mucosal incisions. The specimen size was defined the longest length of specimen after the lesion was completely removed and pinned onto board. Surface area was determined using the ellipse formula: Area [cm2] = (shortest length [cm]/2) x (longest length [cm]/2) x π. The dissection
TE D
speed (cm2/min) was defined as the specimen area (cm2) divided by duration (minutes). En bloc resection, complete and curative resection was defined as previously described [13]. The coagulation forceps was used to stanch bleeding of larger branch vessels and its usage was to indicate the
EP
occurrence of major bleeding events and recorded as the percentage of used case in all cases during
AC C
procedure. Postprocedural bleeding was diagnosed on the onset of clinical symptoms, such as melena or hematemesis, or based on a decrease > 2 g/dL in hemoglobin level after ESD. Muscular injury can be observed during ESD procedure under endoscopy and defined as thermal injury with any muscular defect, minimal thermal change was not regarded as muscular injury. A diagnosis of perforation was considered if the extramural organ outside of the muscle layer was visualized by endoscopy or significant pneumomediastinum was observed on radiography or CT. Propensity score matching
7
ACCEPTED MANUSCRIPT The propensity score matching (PSM) approach allows the assembly of a well-balanced cohort in all available explanatory factors [15]. During ESD procedure, lesion size was regarded as primary factor related to longer procedure time and technical difficulty as previously reported [8, 16, 17]. Thus,
RI PT
propensity score matching ( tESD group: cESD group in a 1:1 match) was conducted to adjust a set of available explanatory factors including resected specimen size determined by median, gender, age and histology, which might affect the result of the comparison between the 2 groups. A logistic
SC
regression model, measuring how each of those explanatory factors predicted being assigned to type
M AN U
of treatment, was used to calculate the propensity score for each individual Hosmer-Lemeshow test and c-statistic test were used in the evaluation of the goodness-of-fit of the propensity score model. ASD (Absolute standard differences) was used to evaluate the balance of the confounding variables
Statistical analyses
TE D
between 2 groups after PSM.
Quantitative variables were summarized by either the mean (standard deviation, SD) for normally
EP
distributed data or the median and interquartile range (IQR) for skewed distribution determined by
AC C
Shapiro-Wilk test. The χ2 test was used to compare categorical variables, the Student T test was used for continuous and normally distributed variables and the Mann-Whitney U-test was used to compare medians if data were not normally distributed. In univariate and multivariate regression analysis, continuous variables, such as age and resected specimen size were categorized into 2 levels according to the mean or median of the included data. Finally, age (≥60 years vs <60 years), sex (male vs female), resected specimen size (≥40 mm vs <40 mm), location of lesions (upper, middle or lower third), perforation (yes or no), use of tunnel method (yes or no) were included as
8
ACCEPTED MANUSCRIPT independent variables. A p value less than 0.05 in each single testing was considered statistically significant. Uncorrected p values were shown with results along with indications whenever correction for multiple testing may remove significance. All statistical analyses were performed by
RI PT
using SPSS software, version 22.0 (SPSS Inc., Chicago, Ill). Results Baseline characteristics of patients
SC
A total of 121 patients with 134 superficial ESCC treated by ESD at Xijing Hospital of Digestive Disease between October 2013 and September 2015 were retrieved. Among them, one patient with
M AN U
lesions located in the remnant esophagus after esophagogastrectomy, two patients with discrete lesions which were not suitable for ESD and 16 multiple lesions were excluded. Finally, 115 lesions in 115 patients treated by ESD were included in our study. The study selection process was shown in Figure 2. Clinicopathological characteristics between tESD group (n=38) and cESD group (n=77) were shown in Table 1. Among them, the resected specimen size, the usage of coagulation forceps,
TE D
the procedure time and submucosal dissection time showed a skewed distribution (P≤0.001) and thus median and IQR were showed. Tunnel technique was predominantly used after Jan 2015, while most cases before were performed with conventional method. There is a tendency of difference on comparing gender and pathology finding (P = 0.05). A propensity score matching was conducted to
EP
minimize the selection bias. The four parameters were set as matching covariant variables for PSM, including age (60 years vs <60 years), sex, resected specimen size determined by median (≥40 mm
AC C
vs <40 mm), and pathology of lesions (HIGH vs. T1a vs. T1b vs others). Finally, 38 patient pairs were matched in this study. The propensity score model was well calibrated (Hosmer-Lemeshow test, P = 0.74), and discriminated well between two groups (c statistic = 0.82). All treated patients were matched to the closest control within 0.15 ASD of the logit of the estimated propensity score. There were no significant differences in age, pathology finding, gender, resected specimen size or location after PSM (P > 0.05) as shown in Table 1. Treatment outcomes There was no significant difference between the tESD group and cESD group on the rates of en bloc 9
ACCEPTED MANUSCRIPT resection, complete resection and curative resection before and after PSM (P > 0.05). The procedure time and submucosal dissection time showed similar improvement in tESD group compared with those of cESD group before (38.0 [29.5 - 46.0] min vs 45.0 [34.0 - 74.0] min, P = 0.016) and after PSM (38.0 (29.5 - 46.0) min vs 48.0 [35.4 - 83.3], P = 0.006). Meaningfully, the dissection speed
[0.19-0.32) cm2/min vs 0.17 [0.10-0.19] cm2, P = 0.000).
RI PT
also showed improvement in the tunnel ESD group than conventional ESD group after PSM (0.23
With regard to adverse events, no perforation and post-procedural bleeding occurred in tESD group. In cESD group, 4 perforations and one post-procedural bleeding occurred before matching. There
SC
was no difference on perforation rate and post-procedural bleeding rate between the two groups both before and after PSM (P > 0.05, Table 2). However, the usage of coagulation forceps during ESD
M AN U
procedure in tunnel method group was fewer than conventional group (P < 0.05) which could be related to less major bleeding events occurred in tESD group (36.8% vs 65.8%; P = 0.012). The postoperative chest pain rate did also not show significant differences between tESD and cESD groups (10.5% vs 10.5 %, P = 1.000, Table 2). All perforation or bleeding events were successfully treated by endoscopy or conservative measures. Interestingly, the frequency of muscular injury
TE D
during ESD procedure in tunnel method group was lower than conventional group (28.9% vs 52.6%, P = 0.036) after PSM. P < 0.05 in a single test of significance was considered statistically significant, however, correction for multiple testing may remove that nominal significance.
EP
Univariate and multivariate analyses of procedure time To analyze the factors associated with procedure time, the univariate and multivariate regression
AC C
analysis were conducted (Table 3). The univariate analysis showed that the tunnel method and resected specimen size <40 mm were related to shorter procedure time. P < 0.05 in a single test of significance was considered statistically significant, however, correction for multiple testing may remove that nominal significance. A further multivariate analysis also confirmed that the resected specimen size <40 mm (OR, 8.74; 95% CI, 1.30 – 58.5; P = 0.026) was an independent predicting factor for shorter procedure time as previously reported [6, 10]. Meanwhile, tunnel method was confirmed to be an additional independent predicting factor for shorter procedure time (OR, 3.42; 95% CI, 1.32–8.85; P = 0.011). The tunnel method was considered to be more useful in large lesions [6]; therefore, it would be
10
ACCEPTED MANUSCRIPT interesting to know if it could improve the procedure time in smaller lesions. To confirm this, a subgroup analysis was conducted (Table 4). It was indicated that the tunnel method improved the efficacy both in lesion <40 mm and ≥40 mm, though the less significant improvement was found on procedure time of smaller lesions.
RI PT
DISCUSSION
Although considered as an excellent therapeutic procedure with a high en bloc resection rate, the
SC
ESD procedure itself was technically complicated and adverse events occur quite often [18]. Esophageal ESD was especially difficult due to the facts of narrow esophageal lumen, thin
M AN U
esophageal wall, lack of serosa, continuous movements with respiration and heartbeat, so the risk of perforation was greater for esophageal ESD compared with gastric ESD [5, 19]. Especially for large lesion, rapid diffusion of submucosal liquid cushion and the contraction of the resected mucosa make the conventional ESD procedure time-consuming and have an increased risk of adverse events [6, 7].
TE D
Previously, multiple modifications have been attempted including usage of hyaluronic acid in submucosal injection and clip assisted traction [20-22]. However, most modifications required extra
EP
steps or devices, which increased procedure time and cost, modification on procedure itself would be of advantages since no specific devices or steps required.
AC C
Per-oral endoscopic myotomy (POEM) introduced since 2010 by Inoue et al [23] was a novel endosurgical therapy for achalasia opening a new discipline of submucosal endoscopic surgery [24]. Inspired by POEM, submucosal tunnel endoscopic resection (STER) was developed for the treatment of submucosal tumors arising from the muscularis propria (MP) [12]. Interestingly, this submucosal tunnel technique maintain a favorable view on submucosal layer during tunnel making; therefore, ESD using the tunnel method was proposed [6]. Although this method is considered to be more effective and feasible in patients or living animal model, there was few study verifying the efficacy 11
ACCEPTED MANUSCRIPT and safety of endoscopic submucosal tunnel dissection compared with conventional ESD [6, 9, 10, 12, 25]. Our retrospective study was the first comparative trial to analyze the efficacy and safety of tunnel
RI PT
technique. Our study showed that ESD procedure time could be improved to over 20% in matched cases. In a descriptive study of 5 patients with esophageal lesions underwent tunnel ESD, Linghu et al [6] reported that the median procedure duration was 78.6 minutes (34-120 min) with a median
SC
diameter of 57 mm (range 40 - 80 mm) of resected specimens. Pioche et al [9] retrospectively
M AN U
reviewed 11 consecutive patients underwent tunnel ESD, the median procedure duration was 76.7 minutes for a median length of 45 mm (range 27 - 80 mm). With regard to the literature reports on conventional procedures, Lee et al [16] reported the mean duration of 92.7 (69) minutes for 24 patients conducted conventional ESD with a median specimen size of 43.1 mm (8-80 mm).
TE D
Compared with the previous studies, our data showed a similar shorter procedure time in tESD group. A similar result was revealed when submucosal dissection speed was compared. Several reasons might contribute the improved efficacy: better view for submucosal layer, more
EP
efficient vessel coagulation during submucosal dissection, decreased blood supply for
AC C
circumferential cutting. In tunnel technique, the endoscopic view was well maintained since endoscope was stably kept in the tunnel, and submucosal liquid cushion lasted longer with the preservation of lateral mucosa, therefore better dissection plane could be maintained, which would shorten the operation time and reduce the risk of perforation [9, 11] Though our small scaled study showed no difference on perforation rate, but it did provide data that less muscular injury (28.9% vs 52.6 %; P = 0.036) could be achieved in tunnel ESD group, which would be an important parameter showing the procedure safety could be improved by tESD method. The better view of submucosal
12
ACCEPTED MANUSCRIPT space made pre-coagulation possible, which was an essential strategy to avoid intraprocedure bleeding [12]. Coagulation forceps were mainly used to stanch bleeding of larger branch vessels; therefore, the usage of coagulation forceps might correlate with the occurrence of major bleeding
RI PT
event. The less usage of coagulation forceps in our study might indicate that less major bleeding events occurred in tESD group.
As previously reported, lesion size was regarded as primary factor related to procedure time during
SC
ESD procedure [8, 16, 17]. It would be interesting to know if the tunnel technique was useful in
M AN U
small lesions. A subgroup analysis according to specimen size was conducted in our study. It was shown that procedure time in both small and large lesions could be improved by tunnel method, indicating either small or large lesions should be considered to be treated by tESD. In tunnel technique, precaution should be taken on the extent of submucosal dissection since the
TE D
lateral mucosa were preserved and there was no significant marker to show the lateral boundary for dissection. Therefore, careful examination on the lateral extent of dissection would be recommended during the procedure [10, 25]. It might not be necessary to dissect the submucosal tissue exactly to
EP
the boundary underneath the lateral margin since the remaining submucosal tissue at lateral margin
AC C
could be easily removed after lateral cutting. There were limitations in our study. First, the study was retrospective in nature. Potential selection bias cannot be ignored, therefore we applied the propensity score matching method to minimized influence of patient characteristics on the clinical outcome and the PSM method was conducted to balance the confounding variables to make the results to be similar to randomized controlled studies [15]. In addition, it was a single center analysis. It is difficult to represent the characteristics of patients in general. Furthermore, only 3 expert endoscopists performed all ESDs, median procedure
13
ACCEPTED MANUSCRIPT time was much shorter than those in other studies [6, 9, 16], which may explain the procedure time had no dramatic reduction in the tunnel ESD group. Finally, the number of subjects enrolled was relatively small. Thus, a multicenter, randomized controlled trial in future would be needed to
RI PT
confirm its role in esophageal ESD. In conclusion, the endoscopic submucosal tunnel dissection is an advantageous option for the management of superficial esophageal squamous cell carcinoma. Compared with the conventional
SC
ESD procedure, it is more efficient and probably safer, presumably by providing good visualization
References
M AN U
during submucosal dissection.
[1] Onozato Y, Ishihara H, Iizuka H, et al. Endoscopic submucosal dissection for early gastric cancers and large flat adenomas. Endoscopy. 2006;38:980-6.
[2] Oyama T, Tomori A, Hotta K, et al. Endoscopic submucosal dissection of early esophageal cancer. Clin Gastroenterol
TE D
Hepatol. 2005;3:S67-70.
[3] Maeda Y, Hirasawa D, Fujita N, et al. A pilot study to assess mediastinal emphysema after esophageal endoscopic submucosal dissection with carbon dioxide insufflation. Endoscopy. 2012;44:565-71. [4] Maeda Y, Hirasawa D, Fujita N, et al. Mediastinal emphysema after esophageal endoscopic submucosal dissection: its prevalence and clinical significance. Dig Endosc. 2011;23:221-6.
EP
[5] Kim JS, Kim BW, Shin IS. Efficacy and safety of endoscopic submucosal dissection for superficial squamous esophageal neoplasia: a meta-analysis. Dig Dis Sci. 2014;59:1862-9. [6] Linghu E, Feng X, Wang X, et al. Endoscopic submucosal tunnel dissection for large esophageal neoplastic lesions.
AC C
Endoscopy. 2013;45:60-2.
[7] Tsao SK, Toyonaga T, Morita Y, et al. Modified fishing-line traction system in endoscopic submucosal dissection of large esophageal tumors. Endoscopy. 2011;43 Suppl 2 UCTN:E119. [8] Koike Y, Hirasawa D, Fujita N, et al. Usefulness of the thread-traction method in esophageal endoscopic submucosal dissection: randomized controlled trial. Dig Endosc. 2015;27:303-9. [9] Pioche M, Mais L, Guillaud O, et al. Endoscopic submucosal tunnel dissection for large esophageal neoplastic lesions. Endoscopy. 2013;45:1032-4. [10] Zhai YQ, Li HK, Linghu EQ. Endoscopic submucosal tunnel dissection for large superficial esophageal squamous cell neoplasms. World J Gastroenterol. 2016;22:435-45. [11] Gomercic C, Vanbiervliet G, Gonzalez JM, et al. Prospective randomized comparison of endoscopic submucosal tunnel dissection and conventional submucosal dissection in the resection of superficial esophageal/gastric lesions in a living porcine model. Endosc Int Open. 2015;3:E577-83. [12] Werner YB, Rosch T. POEM and Submucosal Tunneling. Curr Treat Options Gastroenterol. 2016;14:163-77.
14
ACCEPTED MANUSCRIPT [13] Society JE. Japanese Classification of Esophageal Cancer, 11th Edition: part I. Esophagus. 2017;14:1-36. [14] Honda K, Akiho H. Endoscopic submucosal dissection for superficial esophageal squamous cell neoplasms. World J Gastrointest Pathophysiol. 2012;3:44-50. [15] Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res. 2011;46:399-424. [16] Lee CT, Chang CY, Tai CM, et al. Endoscopic submucosal dissection for early esophageal neoplasia: a single center experience in South Taiwan. J Formos Med Assoc. 2012;111:132-9.
RI PT
[17] Repici A, Hassan C, Carlino A, et al. Endoscopic submucosal dissection in patients with early esophageal squamous cell carcinoma: results from a prospective Western series. Gastrointest Endosc. 2010;71:715-21.
[18] Teoh AY, Chiu PW, Wong SK, et al. Difficulties and outcomes in starting endoscopic submucosal dissection. Surg Endosc. 2010;24:1049-54.
[19] Oyama T. Esophageal ESD: technique and prevention of complications. Gastrointest Endosc Clin N Am. 2014;24:201-12.
SC
[20] Ota M, Nakamura T, Hayashi K, et al. Usefulness of clip traction in the early phase of esophageal endoscopic submucosal dissection. Dig Endosc. 2012;24:315-8.
[21] Oyama T. Counter traction makes endoscopic submucosal dissection easier. Clin Endosc. 2012;45:375-8.
M AN U
[22] Fujishiro M, Yahagi N, Nakamura M, et al. Successful outcomes of a novel endoscopic treatment for GI tumors: endoscopic submucosal dissection with a mixture of high-molecular-weight hyaluronic acid, glycerin, and sugar. Gastrointest Endosc. 2006;63:243-9.
[23] Inoue H, Minami H, Kobayashi Y, et al. Peroral endoscopic myotomy (POEM) for esophageal achalasia. Endoscopy. 2010;42:265-71.
[24] Chen WF, Li QL, Zhou PH, et al. Long-term outcomes of peroral endoscopic myotomy for achalasia in pediatric patients: a prospective, single-center study. Gastrointest Endosc. 2015;81:91-100.
TE D
[25] Lv XH, Wang CH, Xie Y. Efficacy and safety of submucosal tunneling endoscopic resection for upper gastrointestinal submucosal tumors: a systematic review and meta-analysis. Surg Endosc. 2017;31:49-63. [26] Oyama T, Inoue H, Arima M, et al. Prediction of the invasion depth of superficial squamous cell carcinoma based on
AC C
EP
microvessel morphology: magnifying endoscopic classification of the Japan Esophageal Society. Esophagus. 2016:1-8.
15
ACCEPTED MANUSCRIPT
Figure legends
Figure 1. A representative tunneling ESD case. A, A reddish flat lesion was found in
RI PT
the middle esophagus. B, Narrow-band imaging revealed B1 vessels according to Japanese Esophagus Society Classification[26], which indicating an early cancer with an invasion depth of m1/m2. C, Iodine staining showed an iodine void lesion. After
SC
marking the border of the lesion (D) normal saline solution was injected into the
M AN U
submucosal space, and mucosa incisions were made at anal (E) and oral side of the lesion. F, Submucosal tissue was dissected between anal and oral incisions to form a tunnel underneath the lesion and a clear vision of submucosal layer was maintained. Then the circumferential cutting of mucosa was complete (G and H). I, The remaining
TE D
submucosal tissue was further dissected to finish the ESD procedure.
AC C
EP
Figure 2. Flowchart of lesion selection in this study
16
ACCEPTED MANUSCRIPT
Before PSM Variables
cESD group (n = 77) 59.1 (9.2)
P value 0.839
Male/Female
33/5
54/23
0.050
cESD group (n = 38) 58.9 (10.6)
P value 0.912
33/5
34/4
0.723
M AN U
35.0 (25.5-40.0)
0.089
39.0 (30.0-48.5)
36.0 (29.5-46.3)
0.669
6.87 (4.61-11.00)
0.026
9.42 (6.21-13.25)
6.97 (4.66-11.33)
0.210
0.350
TE D
Specimen size, mm (Median, 39.0 (30.0-48.0) IQR) Surface area, cm2 (Median, 9.42 (6.21-13.24) IQR) Location of lesions
After PSM
tESD group (n = 38) 58.7 (7.8)
SC
Age , y (Mean, SD)
tESD group (n = 38) 58.7 (7.8)
RI PT
Table 1. Comparison of baseline characteristics between tunnel ESD with tunnel technique and conventional ESD groups before and after propensity score matching
0.240
3 (7.9%)
12 (15.6%)
3 (7.9%)
6 (15.8%)
Middle
22 (57.9%)
35 (45.5%)
22 (57.9%)
15 (39.5%)
Lower
13 (34.2%)
30 (38.9%)
13 (34.2%)
17 (44.7%)
HGIN
18 (47.4%)
pT1a
15 (39.5 %)
pT1b Others
AC C
Pathology
EP
Upper
0.051
0.343
24 (31.2%)
18 (47.3 %)
17 (44.5 %)
30 (38.9%)
15 (39.5 %)
11 (28.9 %)
5 (13.1 %)
11 (14.3 %)
5 (13.2 %)
8 (21.1 %)
0 (0%)
12 (15.6%)*
0 (0 %)
2 (5.3 %)# 17
ACCEPTED MANUSCRIPT
cESD, conventional ESD; HGIN, High-grade intraepithelial neoplasia; PSM, propensity score matching; pT1a, Intramucosal
RI PT
carcinoma; pT1b, Submucosal invasive carcinoma; tESD, tunnel ESD; *Including nine cases of low-grade intraepithelial neoplasia and three cases of inflammation in cESD group before PSM; #
AC C
EP
TE D
M AN U
SC
Including two cases of low-grade intraepithelial neoplasia in cESD group after PSM.
18
ACCEPTED MANUSCRIPT
Before PSM cESD group (n =77)
tESD group (n =38)
cESD group (n =38)
P value
38.0(29.5-46.0)
45.0(34.0-74.0)
0.016*
38.0(29.5-46.0)
48.0(35.4-83.3)
0.006*
Submucosal dissection time, min (Median, 30.0(24.0-38.3) IQR)
38.0(26.0-61.5)
0.013*
30.0(24.0-38.3)
40.0(29.0-71.8)
0.005*
Dissection speed, cm2/min (Median, IQR)
0.23(0.19-0.32)
0.15(0.10-0.23)
0.000*
0.23(0.19-0.32)
0.17(0.10-0.19)
0.000*
Coagulation forceps usage
14(36.8%)
54(70.1%)
0.001*
14(36.8%)
25 (65.8%)
0.012*
En bloc resection
38(100%)
76(98.7 %)
0.480
38(100 %)
38(100 %)
1.000
Complete resection
38(100%)
72(93.5%)
0.108
38(100 %)
36(94.7 %)
0.152
Curative resection
36(94.7 %)
74(96.1%)
0.735
36(94.7 %)
35(92.1 %)
0.644
Muscular injury during ESD
11(28.9%)
35(45.4%)
0.089
11(28.9 %)
20(52.6 %)
0.036*
0(0%)
1(1.3%)
0.480
0(0 %)
1(2.6 %)
0.314
0(0%)
4(5.2 %)
0.153
0(0 %)
3(7.9 %)
0.077
4(10.5%)
11(14.3 %)
0.573
4(10.5 %)
4(10.5 %)
1.000
Post-procedure bleeding Perforation Chest pain
M AN U
TE D
EP
Procedure time, min (Median, IQR)
P value
SC
tESD group (n =38)
After PSM
AC C
Variables
RI PT
Table 2. Treatment outcomes before and after propensity score matching
cESD, conventional ESD; PSM, propensity score matching; tESD, tunnel ESD. *P < 0.05 in a single test of significance; however, correction for multiple testing may remove that nominal significance. 19
ACCEPTED MANUSCRIPT
Table 3. Univariate and multivariate regression analysis for procedure time 95%CI
RI PT
OR Univariate regression analysis
19.25
Location of lesions Perforation (yes or no)
0.84 3.28
Use of tunnel method (yes or no)
2.43
Multivariate regression analysis
2.18
3.42 8.74
3.10 1.25
0.294 0.147 0.008*
169.78
0.27 0.33
2.64 32.54
1.08
5.45
1.32 1.30
8.85 58.5
0.772 0.310 0.031*
0.011 0.026
EP
Use of tunnel method (yes or no) Specimen size, mm (<40 or≥ 40)
SC
Specimen size, mm (<40 or ≥40)
0.71 0.22
M AN U
1.48 0.53
TE D
Age (<60 y or ≥60 y) Sex (male or female)
P value
AC C
*P < 0.05 in a single test of significance; however, correction for multiple testing may remove that nominal significance.
20
ACCEPTED MANUSCRIPT
Table 4. The subgroup analysis according to specimen size ＜40 mm
cESD group (n =18) 67.5 (43.3-96.5)
0.012
34.0 (21.0-44.0) 0.044
36.0 (27.0-44.0)
59.5 (32.0-86.0)
0.008
RI PT
tESD group (n =19) 44.0 (36.0-52.0)
P value
AC C
EP
TE D
M AN U
Procedure time, min (Median, IQR) Submucosal dissection time, min 25.0 (19.0-33.0) (Median, IQR)
cESD group P value (n =20) 43.0 (28.0-53.5) 0.036
SC
tESD group (n =19) 32.5 (25.0-39.0)
≥40 mm
21
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0016-5107(17)30178-5
DOI:
10.1016/j.gie.2017.03.001
Reference:
YMGE 10447
To appear in:
Gastrointestinal Endoscopy
Received Date: 26 November 2016 Accepted Date: 1 March 2017
Please cite this article as: Huang R, Cai H, Zhao X, Lu X, Liu M, Lv W, Liu Z, Wu K, Han Y, Efficacy and safety of endoscopic submucosal tunnel dissection for superficial esophageal squamous cell carcinoma: a propensity score matching analysis, Gastrointestinal Endoscopy (2017), doi: 10.1016/ j.gie.2017.03.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Efficacy and safety of endoscopic submucosal tunnel dissection for superficial esophageal squamous cell carcinoma: a propensity score matching analysis Rui Huang1*, Hongwei Cai2*, Xin Zhao1*, Xiaoqiang Lu1, Min Liu1, Wenhao Lv1, Zhiguo Liu1, Kaichun Wu1, Ying Han1. Xijing Hospital of Digestive Diseases, Fourth Military Medical University, Xian, China
2
RI PT
1
Information Technology Department, the First Affiliated Hospital of Xi’an Jiaotong University, Xi’an, China *These three authors contributed equally to this work.
M AN U
SC
Corresponding author: Zhiguo Liu MD and Ying Han MD, Xijing Hospital of Digestive Diseases, Fourth Military Medical University, 127 Changle West Road, Xi’an, Shanxi 710032, China. Telephone: +86-29-84771535; Fax: +86-29-82539041; Email: [email protected] and [email protected] Running title: Tunnel ESD technique in superficial esophageal cancer Meeting presentation: Oral presentation in Digestive Disease Week 2016.
EP
TE D
Abbreviations : ASD, absolute standard differences; cESD, conventional ESD; CI, confidence interval; ESCC, esophageal squamous cell carcinoma; ESD, endoscopic submucosal dissection; ESTD, endoscopic submucosal tunnel dissection; HGIN, High-grade intraepithelial neoplasia; MP, muscularis propria; OR, odd ratio; PSM, propensity score matching; pT1a, Intramucosal carcinoma; pT1b, Submucosal invasive carcinoma; STER, submucosal tunnel endoscopic resection, tESD, tunnel ESD.
AC C
Declaration of interest All authors declare no conflicts of interest. The authors alone are responsible for the content and writing of the paper. Author contributions Study concept and design: Zhiguo Liu; Acquisition of data: Rui Huang, Xiaoqiang Lu; Analysis and interpretation of data: Rui Huang, Zhiguo Liu; Drafting of the manuscript: Rui Huang, Zhiguo Liu, Xin Zhao; Critical revision of the manuscript for important intellectual content: Min Liu, Wenhao Lv; Statistical analysis: Hongwei Cai; Administrative support: Kaichun Wu, Ying Han. Acknowledgments This work was supported in part by the National Natural Science Foundation of China (81272449, 1
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
81372387 and 81572820).
2
ACCEPTED MANUSCRIPT Abstract Background and Aims: Esophageal ESD is technically challenging due to the facts of thinner wall and narrow lumen. Tunnel technique was previously proposed. This current retrospective study
RI PT
aimed to evaluate the efficacy of the tunnel technique in ESD of superficial esophageal squamous cell carcinoma (ESCC).
Methods: Patients underwent ESD for superficial ESCC between October 2013 and September 2015
SC
were included in the study. Propensity score matching was used to compensate for the differences in
M AN U
age, sex, resected specimen size, and pathology. Treatment outcomes were compared with conventional statistic methods between tunnel ESD group and conventional ESD group after matching. To further explore the potential variables that relevant to procedure time, univariate and multivariate logistic regression analyses were applied.
TE D
Results: A total of 115 lesions were included in the analysis. Propensity score matching analysis created 38 matched pairs. There was no difference on en bloc resection rate, complete resection rate and the curative rate between the 2 groups. The ESD procedure time was 38.0 (29.5-46.0) minutes in
EP
tunnel ESD group and 48.0 (35.4-83.3) minutes in conventional ESD group (P = 0.006). There was
AC C
no difference on adverse events including post-procedure bleeding rate, perforation and chest pain, but a lower rate of muscular injury (28.9% vs 52.6 %; P = 0.036) and a less-frequent usage of coagulation forceps (36.8% vs 65.8%; P = 0.012) were shown in tunnel ESD group. In multivariate regression analysis for procedure time, the tunnel ESD technique (OR, 3.42; 95 % CI, 1.32 – 8.85; P = 0.011) and specimen size <40 mm (OR, 8.74; 95% CI, 1.30 – 58.5; P = 0.026) were associated with a shorter procedure time. Conclusion: The endoscopic submucosal tunnel dissection improved the efficacy and safety of ESD
3
ACCEPTED MANUSCRIPT procedure by shortening the procedure time and reducing injury to muscular layer.
RI PT
INTRODUCTION Endoscopic submucosal dissection (ESD) is an excellent therapeutic procedure in resecting localized superficial gastrointestinal neoplasms. Although ESD procedure is technically challenging and
SC
requires a longer treatment time than other conventional endoscopic mucosal resection (EMR), ESD
M AN U
enables en bloc resection and lower recurrence rate[1]. However, esophageal ESD is especially more difficult due to the thinner wall and narrow lumen, which could result in adverse events such as perforation and pneumomediastinum [2-4]. A meta-analysis on esophageal showed that the pooled estimate of perforation rate was 5.0% [5]. Meanwhile, for large lesions, conventional ESD is
TE D
time-consuming and suffering an increased risk of adverse events due to the rapid diffusion of submucosal liquid cushion and blocked lumen by the resected mucosa[6, 7]. Previously multiple strategies including dental floss traction have been applied to make the procedure safer and easier [8].
EP
However, the likelihood of clip malfunction and additional procedure steps would tend to increase
AC C
the complexity of the operation.
Recently, the tunnel technique was proposed in esophageal ESD for superficial cancer [6]. In conventional ESD (cESD) procedure, a complete circumferential cutting around the lesion was made before the submucosal dissection. But in tunnel ESD (tESD) technique, after submucosal injection in the usual way, incisions were made at the anal and oral side of the lesion, submucosal tissue was dissected then between anal and oral incisions to form a tunnel underneath the lesion and a clear vision of submucosal layer could be maintained, then two lateral mucosal incisions were made to
4
ACCEPTED MANUSCRIPT complete the ESD procedure. The theological advantage of tunnel technique was to maintain an improved view on working field, efficient submucosal dissection becomes possible, and contributing to shorten procedure time [9]. However, there were few reports to verify the efficacy of this
RI PT
technique [6, 10-12]. The aim of this retrospective study was to evaluate the efficacy and safety of the tunnel technique during ESD of superficial esophageal squamous cell carcinoma (ESCC) by a propensity score
M AN U
SC
matching analysis.
METHODS Patients
During October 2013 to September 2015, patients who were superficial ESCC treated by ESD at
TE D
Xijing Hospital of Digestive Disease (Xi’an, China) were included in our study. The study protocol was approved by the institutional review board of Xijing hospital and all patients provided written
EP
informed consent before the start of study. The data were collected and retrospectively reviewed from medical records, procedure videos and endoscopic reports. All lesions were confirmed by
AC C
histologic evaluation of biopsy specimens obtained from forceps sampling before the ESD procedure or suspected localized mucosal lesions detected by endoscopy according to Japanese classification on esophageal squamous cell carcinoma by Japan Esophageal Society [13]. Recurrent lesion or lesion located in the remnant esophagus after esophagogastrectomy were excluded. Only the largest one was included in the analysis in patients with multiple lesions. ESD procedure All patients were sedated and intubated. ESD procedures were performed by 3 endoscopists (Z.G.L., 5
ACCEPTED MANUSCRIPT R.H., and X.Q.L.) with experience of more than 100 cases on ESD operation, using a single-channel endoscope (GIF-Q260J; Olympus Optical Co, Ltd, Tokyo, Japan). The conventional ESD procedure was as previously described [14]. The procedure of tunnel ESD
RI PT
was showed in Figure 1. After submucosal injection, incisions were made at both the anal and oral side of the lesion, and the submucosa under the lesion is then dissected to create a tunnel between the distal and proximal incisions. When the endoscope has reached the lower incision, the distal end of
SC
the esophageal lumen could be visualized, then 2 lateral mucosal incisions were made, thus
M AN U
completing the ESD procedure. There was no any traction method used in all ESD cases including dental floss traction.
A saline solution containing minimal indigo carmine was used in all patients. Circumferential cutting and submucosal dissection was carried out with O-type HybridKnife (ERBE, Tübingen, Germany),
TE D
the Dual Knife (KD-650L; Olympus) was used as assistance when submucosal fibrosis was severe. Endoscopic hemostasis was performed with the dissecting knife or the hemostatic forceps (FD-410LR; Olympus). After dissection, preventive endoscopic hemostasis was performed for any
AC C
EP
oozing or exposed vessel. The VIO generator (VIO 200D; ERBE) was used for all ESD procedures.
Outcome measures
The main outcome measure in this study was the procedure time. Submucosal dissection time, rates of en bloc resection, complete resection, curative resection and the usage of coagulation forceps during ESD and rates of adverse events including muscular injury, postprocedural bleeding, perforation and chest pain were also retrieved. Furthermore, the factors associated with the procedure time, including the usage of tunnel method, were analyzed in the multivariate analysis in all lesions
6
ACCEPTED MANUSCRIPT before propensity score matching.
Definitions
RI PT
Procedure time was defined as the time from marking of the lesion till complete removal of the specimen including hemostasis. The submucosal dissection time was defined from completion of circumferential cutting till complete removal of the specimen excluding hemostasis in conventional
SC
ESD group. In tunnel ESD group, the submucosal dissection time was defined from incisions are
M AN U
made at the anal and oral side cutting of the lesion to removal of specimen, minus the time of two lateral mucosal incisions. The specimen size was defined the longest length of specimen after the lesion was completely removed and pinned onto board. Surface area was determined using the ellipse formula: Area [cm2] = (shortest length [cm]/2) x (longest length [cm]/2) x π. The dissection
TE D
speed (cm2/min) was defined as the specimen area (cm2) divided by duration (minutes). En bloc resection, complete and curative resection was defined as previously described [13]. The coagulation forceps was used to stanch bleeding of larger branch vessels and its usage was to indicate the
EP
occurrence of major bleeding events and recorded as the percentage of used case in all cases during
AC C
procedure. Postprocedural bleeding was diagnosed on the onset of clinical symptoms, such as melena or hematemesis, or based on a decrease > 2 g/dL in hemoglobin level after ESD. Muscular injury can be observed during ESD procedure under endoscopy and defined as thermal injury with any muscular defect, minimal thermal change was not regarded as muscular injury. A diagnosis of perforation was considered if the extramural organ outside of the muscle layer was visualized by endoscopy or significant pneumomediastinum was observed on radiography or CT. Propensity score matching
7
ACCEPTED MANUSCRIPT The propensity score matching (PSM) approach allows the assembly of a well-balanced cohort in all available explanatory factors [15]. During ESD procedure, lesion size was regarded as primary factor related to longer procedure time and technical difficulty as previously reported [8, 16, 17]. Thus,
RI PT
propensity score matching ( tESD group: cESD group in a 1:1 match) was conducted to adjust a set of available explanatory factors including resected specimen size determined by median, gender, age and histology, which might affect the result of the comparison between the 2 groups. A logistic
SC
regression model, measuring how each of those explanatory factors predicted being assigned to type
M AN U
of treatment, was used to calculate the propensity score for each individual Hosmer-Lemeshow test and c-statistic test were used in the evaluation of the goodness-of-fit of the propensity score model. ASD (Absolute standard differences) was used to evaluate the balance of the confounding variables
Statistical analyses
TE D
between 2 groups after PSM.
Quantitative variables were summarized by either the mean (standard deviation, SD) for normally
EP
distributed data or the median and interquartile range (IQR) for skewed distribution determined by
AC C
Shapiro-Wilk test. The χ2 test was used to compare categorical variables, the Student T test was used for continuous and normally distributed variables and the Mann-Whitney U-test was used to compare medians if data were not normally distributed. In univariate and multivariate regression analysis, continuous variables, such as age and resected specimen size were categorized into 2 levels according to the mean or median of the included data. Finally, age (≥60 years vs <60 years), sex (male vs female), resected specimen size (≥40 mm vs <40 mm), location of lesions (upper, middle or lower third), perforation (yes or no), use of tunnel method (yes or no) were included as
8
ACCEPTED MANUSCRIPT independent variables. A p value less than 0.05 in each single testing was considered statistically significant. Uncorrected p values were shown with results along with indications whenever correction for multiple testing may remove significance. All statistical analyses were performed by
RI PT
using SPSS software, version 22.0 (SPSS Inc., Chicago, Ill). Results Baseline characteristics of patients
SC
A total of 121 patients with 134 superficial ESCC treated by ESD at Xijing Hospital of Digestive Disease between October 2013 and September 2015 were retrieved. Among them, one patient with
M AN U
lesions located in the remnant esophagus after esophagogastrectomy, two patients with discrete lesions which were not suitable for ESD and 16 multiple lesions were excluded. Finally, 115 lesions in 115 patients treated by ESD were included in our study. The study selection process was shown in Figure 2. Clinicopathological characteristics between tESD group (n=38) and cESD group (n=77) were shown in Table 1. Among them, the resected specimen size, the usage of coagulation forceps,
TE D
the procedure time and submucosal dissection time showed a skewed distribution (P≤0.001) and thus median and IQR were showed. Tunnel technique was predominantly used after Jan 2015, while most cases before were performed with conventional method. There is a tendency of difference on comparing gender and pathology finding (P = 0.05). A propensity score matching was conducted to
EP
minimize the selection bias. The four parameters were set as matching covariant variables for PSM, including age (60 years vs <60 years), sex, resected specimen size determined by median (≥40 mm
AC C
vs <40 mm), and pathology of lesions (HIGH vs. T1a vs. T1b vs others). Finally, 38 patient pairs were matched in this study. The propensity score model was well calibrated (Hosmer-Lemeshow test, P = 0.74), and discriminated well between two groups (c statistic = 0.82). All treated patients were matched to the closest control within 0.15 ASD of the logit of the estimated propensity score. There were no significant differences in age, pathology finding, gender, resected specimen size or location after PSM (P > 0.05) as shown in Table 1. Treatment outcomes There was no significant difference between the tESD group and cESD group on the rates of en bloc 9
ACCEPTED MANUSCRIPT resection, complete resection and curative resection before and after PSM (P > 0.05). The procedure time and submucosal dissection time showed similar improvement in tESD group compared with those of cESD group before (38.0 [29.5 - 46.0] min vs 45.0 [34.0 - 74.0] min, P = 0.016) and after PSM (38.0 (29.5 - 46.0) min vs 48.0 [35.4 - 83.3], P = 0.006). Meaningfully, the dissection speed
[0.19-0.32) cm2/min vs 0.17 [0.10-0.19] cm2, P = 0.000).
RI PT
also showed improvement in the tunnel ESD group than conventional ESD group after PSM (0.23
With regard to adverse events, no perforation and post-procedural bleeding occurred in tESD group. In cESD group, 4 perforations and one post-procedural bleeding occurred before matching. There
SC
was no difference on perforation rate and post-procedural bleeding rate between the two groups both before and after PSM (P > 0.05, Table 2). However, the usage of coagulation forceps during ESD
M AN U
procedure in tunnel method group was fewer than conventional group (P < 0.05) which could be related to less major bleeding events occurred in tESD group (36.8% vs 65.8%; P = 0.012). The postoperative chest pain rate did also not show significant differences between tESD and cESD groups (10.5% vs 10.5 %, P = 1.000, Table 2). All perforation or bleeding events were successfully treated by endoscopy or conservative measures. Interestingly, the frequency of muscular injury
TE D
during ESD procedure in tunnel method group was lower than conventional group (28.9% vs 52.6%, P = 0.036) after PSM. P < 0.05 in a single test of significance was considered statistically significant, however, correction for multiple testing may remove that nominal significance.
EP
Univariate and multivariate analyses of procedure time To analyze the factors associated with procedure time, the univariate and multivariate regression
AC C
analysis were conducted (Table 3). The univariate analysis showed that the tunnel method and resected specimen size <40 mm were related to shorter procedure time. P < 0.05 in a single test of significance was considered statistically significant, however, correction for multiple testing may remove that nominal significance. A further multivariate analysis also confirmed that the resected specimen size <40 mm (OR, 8.74; 95% CI, 1.30 – 58.5; P = 0.026) was an independent predicting factor for shorter procedure time as previously reported [6, 10]. Meanwhile, tunnel method was confirmed to be an additional independent predicting factor for shorter procedure time (OR, 3.42; 95% CI, 1.32–8.85; P = 0.011). The tunnel method was considered to be more useful in large lesions [6]; therefore, it would be
10
ACCEPTED MANUSCRIPT interesting to know if it could improve the procedure time in smaller lesions. To confirm this, a subgroup analysis was conducted (Table 4). It was indicated that the tunnel method improved the efficacy both in lesion <40 mm and ≥40 mm, though the less significant improvement was found on procedure time of smaller lesions.
RI PT
DISCUSSION
Although considered as an excellent therapeutic procedure with a high en bloc resection rate, the
SC
ESD procedure itself was technically complicated and adverse events occur quite often [18]. Esophageal ESD was especially difficult due to the facts of narrow esophageal lumen, thin
M AN U
esophageal wall, lack of serosa, continuous movements with respiration and heartbeat, so the risk of perforation was greater for esophageal ESD compared with gastric ESD [5, 19]. Especially for large lesion, rapid diffusion of submucosal liquid cushion and the contraction of the resected mucosa make the conventional ESD procedure time-consuming and have an increased risk of adverse events [6, 7].
TE D
Previously, multiple modifications have been attempted including usage of hyaluronic acid in submucosal injection and clip assisted traction [20-22]. However, most modifications required extra
EP
steps or devices, which increased procedure time and cost, modification on procedure itself would be of advantages since no specific devices or steps required.
AC C
Per-oral endoscopic myotomy (POEM) introduced since 2010 by Inoue et al [23] was a novel endosurgical therapy for achalasia opening a new discipline of submucosal endoscopic surgery [24]. Inspired by POEM, submucosal tunnel endoscopic resection (STER) was developed for the treatment of submucosal tumors arising from the muscularis propria (MP) [12]. Interestingly, this submucosal tunnel technique maintain a favorable view on submucosal layer during tunnel making; therefore, ESD using the tunnel method was proposed [6]. Although this method is considered to be more effective and feasible in patients or living animal model, there was few study verifying the efficacy 11
ACCEPTED MANUSCRIPT and safety of endoscopic submucosal tunnel dissection compared with conventional ESD [6, 9, 10, 12, 25]. Our retrospective study was the first comparative trial to analyze the efficacy and safety of tunnel
RI PT
technique. Our study showed that ESD procedure time could be improved to over 20% in matched cases. In a descriptive study of 5 patients with esophageal lesions underwent tunnel ESD, Linghu et al [6] reported that the median procedure duration was 78.6 minutes (34-120 min) with a median
SC
diameter of 57 mm (range 40 - 80 mm) of resected specimens. Pioche et al [9] retrospectively
M AN U
reviewed 11 consecutive patients underwent tunnel ESD, the median procedure duration was 76.7 minutes for a median length of 45 mm (range 27 - 80 mm). With regard to the literature reports on conventional procedures, Lee et al [16] reported the mean duration of 92.7 (69) minutes for 24 patients conducted conventional ESD with a median specimen size of 43.1 mm (8-80 mm).
TE D
Compared with the previous studies, our data showed a similar shorter procedure time in tESD group. A similar result was revealed when submucosal dissection speed was compared. Several reasons might contribute the improved efficacy: better view for submucosal layer, more
EP
efficient vessel coagulation during submucosal dissection, decreased blood supply for
AC C
circumferential cutting. In tunnel technique, the endoscopic view was well maintained since endoscope was stably kept in the tunnel, and submucosal liquid cushion lasted longer with the preservation of lateral mucosa, therefore better dissection plane could be maintained, which would shorten the operation time and reduce the risk of perforation [9, 11] Though our small scaled study showed no difference on perforation rate, but it did provide data that less muscular injury (28.9% vs 52.6 %; P = 0.036) could be achieved in tunnel ESD group, which would be an important parameter showing the procedure safety could be improved by tESD method. The better view of submucosal
12
ACCEPTED MANUSCRIPT space made pre-coagulation possible, which was an essential strategy to avoid intraprocedure bleeding [12]. Coagulation forceps were mainly used to stanch bleeding of larger branch vessels; therefore, the usage of coagulation forceps might correlate with the occurrence of major bleeding
RI PT
event. The less usage of coagulation forceps in our study might indicate that less major bleeding events occurred in tESD group.
As previously reported, lesion size was regarded as primary factor related to procedure time during
SC
ESD procedure [8, 16, 17]. It would be interesting to know if the tunnel technique was useful in
M AN U
small lesions. A subgroup analysis according to specimen size was conducted in our study. It was shown that procedure time in both small and large lesions could be improved by tunnel method, indicating either small or large lesions should be considered to be treated by tESD. In tunnel technique, precaution should be taken on the extent of submucosal dissection since the
TE D
lateral mucosa were preserved and there was no significant marker to show the lateral boundary for dissection. Therefore, careful examination on the lateral extent of dissection would be recommended during the procedure [10, 25]. It might not be necessary to dissect the submucosal tissue exactly to
EP
the boundary underneath the lateral margin since the remaining submucosal tissue at lateral margin
AC C
could be easily removed after lateral cutting. There were limitations in our study. First, the study was retrospective in nature. Potential selection bias cannot be ignored, therefore we applied the propensity score matching method to minimized influence of patient characteristics on the clinical outcome and the PSM method was conducted to balance the confounding variables to make the results to be similar to randomized controlled studies [15]. In addition, it was a single center analysis. It is difficult to represent the characteristics of patients in general. Furthermore, only 3 expert endoscopists performed all ESDs, median procedure
13
ACCEPTED MANUSCRIPT time was much shorter than those in other studies [6, 9, 16], which may explain the procedure time had no dramatic reduction in the tunnel ESD group. Finally, the number of subjects enrolled was relatively small. Thus, a multicenter, randomized controlled trial in future would be needed to
RI PT
confirm its role in esophageal ESD. In conclusion, the endoscopic submucosal tunnel dissection is an advantageous option for the management of superficial esophageal squamous cell carcinoma. Compared with the conventional
SC
ESD procedure, it is more efficient and probably safer, presumably by providing good visualization
References
M AN U
during submucosal dissection.
[1] Onozato Y, Ishihara H, Iizuka H, et al. Endoscopic submucosal dissection for early gastric cancers and large flat adenomas. Endoscopy. 2006;38:980-6.
[2] Oyama T, Tomori A, Hotta K, et al. Endoscopic submucosal dissection of early esophageal cancer. Clin Gastroenterol
TE D
Hepatol. 2005;3:S67-70.
[3] Maeda Y, Hirasawa D, Fujita N, et al. A pilot study to assess mediastinal emphysema after esophageal endoscopic submucosal dissection with carbon dioxide insufflation. Endoscopy. 2012;44:565-71. [4] Maeda Y, Hirasawa D, Fujita N, et al. Mediastinal emphysema after esophageal endoscopic submucosal dissection: its prevalence and clinical significance. Dig Endosc. 2011;23:221-6.
EP
[5] Kim JS, Kim BW, Shin IS. Efficacy and safety of endoscopic submucosal dissection for superficial squamous esophageal neoplasia: a meta-analysis. Dig Dis Sci. 2014;59:1862-9. [6] Linghu E, Feng X, Wang X, et al. Endoscopic submucosal tunnel dissection for large esophageal neoplastic lesions.
AC C
Endoscopy. 2013;45:60-2.
[7] Tsao SK, Toyonaga T, Morita Y, et al. Modified fishing-line traction system in endoscopic submucosal dissection of large esophageal tumors. Endoscopy. 2011;43 Suppl 2 UCTN:E119. [8] Koike Y, Hirasawa D, Fujita N, et al. Usefulness of the thread-traction method in esophageal endoscopic submucosal dissection: randomized controlled trial. Dig Endosc. 2015;27:303-9. [9] Pioche M, Mais L, Guillaud O, et al. Endoscopic submucosal tunnel dissection for large esophageal neoplastic lesions. Endoscopy. 2013;45:1032-4. [10] Zhai YQ, Li HK, Linghu EQ. Endoscopic submucosal tunnel dissection for large superficial esophageal squamous cell neoplasms. World J Gastroenterol. 2016;22:435-45. [11] Gomercic C, Vanbiervliet G, Gonzalez JM, et al. Prospective randomized comparison of endoscopic submucosal tunnel dissection and conventional submucosal dissection in the resection of superficial esophageal/gastric lesions in a living porcine model. Endosc Int Open. 2015;3:E577-83. [12] Werner YB, Rosch T. POEM and Submucosal Tunneling. Curr Treat Options Gastroenterol. 2016;14:163-77.
14
ACCEPTED MANUSCRIPT [13] Society JE. Japanese Classification of Esophageal Cancer, 11th Edition: part I. Esophagus. 2017;14:1-36. [14] Honda K, Akiho H. Endoscopic submucosal dissection for superficial esophageal squamous cell neoplasms. World J Gastrointest Pathophysiol. 2012;3:44-50. [15] Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res. 2011;46:399-424. [16] Lee CT, Chang CY, Tai CM, et al. Endoscopic submucosal dissection for early esophageal neoplasia: a single center experience in South Taiwan. J Formos Med Assoc. 2012;111:132-9.
RI PT
[17] Repici A, Hassan C, Carlino A, et al. Endoscopic submucosal dissection in patients with early esophageal squamous cell carcinoma: results from a prospective Western series. Gastrointest Endosc. 2010;71:715-21.
[18] Teoh AY, Chiu PW, Wong SK, et al. Difficulties and outcomes in starting endoscopic submucosal dissection. Surg Endosc. 2010;24:1049-54.
[19] Oyama T. Esophageal ESD: technique and prevention of complications. Gastrointest Endosc Clin N Am. 2014;24:201-12.
SC
[20] Ota M, Nakamura T, Hayashi K, et al. Usefulness of clip traction in the early phase of esophageal endoscopic submucosal dissection. Dig Endosc. 2012;24:315-8.
[21] Oyama T. Counter traction makes endoscopic submucosal dissection easier. Clin Endosc. 2012;45:375-8.
M AN U
[22] Fujishiro M, Yahagi N, Nakamura M, et al. Successful outcomes of a novel endoscopic treatment for GI tumors: endoscopic submucosal dissection with a mixture of high-molecular-weight hyaluronic acid, glycerin, and sugar. Gastrointest Endosc. 2006;63:243-9.
[23] Inoue H, Minami H, Kobayashi Y, et al. Peroral endoscopic myotomy (POEM) for esophageal achalasia. Endoscopy. 2010;42:265-71.
[24] Chen WF, Li QL, Zhou PH, et al. Long-term outcomes of peroral endoscopic myotomy for achalasia in pediatric patients: a prospective, single-center study. Gastrointest Endosc. 2015;81:91-100.
TE D
[25] Lv XH, Wang CH, Xie Y. Efficacy and safety of submucosal tunneling endoscopic resection for upper gastrointestinal submucosal tumors: a systematic review and meta-analysis. Surg Endosc. 2017;31:49-63. [26] Oyama T, Inoue H, Arima M, et al. Prediction of the invasion depth of superficial squamous cell carcinoma based on
AC C
EP
microvessel morphology: magnifying endoscopic classification of the Japan Esophageal Society. Esophagus. 2016:1-8.
15
ACCEPTED MANUSCRIPT
Figure legends
Figure 1. A representative tunneling ESD case. A, A reddish flat lesion was found in
RI PT
the middle esophagus. B, Narrow-band imaging revealed B1 vessels according to Japanese Esophagus Society Classification[26], which indicating an early cancer with an invasion depth of m1/m2. C, Iodine staining showed an iodine void lesion. After
SC
marking the border of the lesion (D) normal saline solution was injected into the
M AN U
submucosal space, and mucosa incisions were made at anal (E) and oral side of the lesion. F, Submucosal tissue was dissected between anal and oral incisions to form a tunnel underneath the lesion and a clear vision of submucosal layer was maintained. Then the circumferential cutting of mucosa was complete (G and H). I, The remaining
TE D
submucosal tissue was further dissected to finish the ESD procedure.
AC C
EP
Figure 2. Flowchart of lesion selection in this study
16
ACCEPTED MANUSCRIPT
Before PSM Variables
cESD group (n = 77) 59.1 (9.2)
P value 0.839
Male/Female
33/5
54/23
0.050
cESD group (n = 38) 58.9 (10.6)
P value 0.912
33/5
34/4
0.723
M AN U
35.0 (25.5-40.0)
0.089
39.0 (30.0-48.5)
36.0 (29.5-46.3)
0.669
6.87 (4.61-11.00)
0.026
9.42 (6.21-13.25)
6.97 (4.66-11.33)
0.210
0.350
TE D
Specimen size, mm (Median, 39.0 (30.0-48.0) IQR) Surface area, cm2 (Median, 9.42 (6.21-13.24) IQR) Location of lesions
After PSM
tESD group (n = 38) 58.7 (7.8)
SC
Age , y (Mean, SD)
tESD group (n = 38) 58.7 (7.8)
RI PT
Table 1. Comparison of baseline characteristics between tunnel ESD with tunnel technique and conventional ESD groups before and after propensity score matching
0.240
3 (7.9%)
12 (15.6%)
3 (7.9%)
6 (15.8%)
Middle
22 (57.9%)
35 (45.5%)
22 (57.9%)
15 (39.5%)
Lower
13 (34.2%)
30 (38.9%)
13 (34.2%)
17 (44.7%)
HGIN
18 (47.4%)
pT1a
15 (39.5 %)
pT1b Others
AC C
Pathology
EP
Upper
0.051
0.343
24 (31.2%)
18 (47.3 %)
17 (44.5 %)
30 (38.9%)
15 (39.5 %)
11 (28.9 %)
5 (13.1 %)
11 (14.3 %)
5 (13.2 %)
8 (21.1 %)
0 (0%)
12 (15.6%)*
0 (0 %)
2 (5.3 %)# 17
ACCEPTED MANUSCRIPT
cESD, conventional ESD; HGIN, High-grade intraepithelial neoplasia; PSM, propensity score matching; pT1a, Intramucosal
RI PT
carcinoma; pT1b, Submucosal invasive carcinoma; tESD, tunnel ESD; *Including nine cases of low-grade intraepithelial neoplasia and three cases of inflammation in cESD group before PSM; #
AC C
EP
TE D
M AN U
SC
Including two cases of low-grade intraepithelial neoplasia in cESD group after PSM.
18
ACCEPTED MANUSCRIPT
Before PSM cESD group (n =77)
tESD group (n =38)
cESD group (n =38)
P value
38.0(29.5-46.0)
45.0(34.0-74.0)
0.016*
38.0(29.5-46.0)
48.0(35.4-83.3)
0.006*
Submucosal dissection time, min (Median, 30.0(24.0-38.3) IQR)
38.0(26.0-61.5)
0.013*
30.0(24.0-38.3)
40.0(29.0-71.8)
0.005*
Dissection speed, cm2/min (Median, IQR)
0.23(0.19-0.32)
0.15(0.10-0.23)
0.000*
0.23(0.19-0.32)
0.17(0.10-0.19)
0.000*
Coagulation forceps usage
14(36.8%)
54(70.1%)
0.001*
14(36.8%)
25 (65.8%)
0.012*
En bloc resection
38(100%)
76(98.7 %)
0.480
38(100 %)
38(100 %)
1.000
Complete resection
38(100%)
72(93.5%)
0.108
38(100 %)
36(94.7 %)
0.152
Curative resection
36(94.7 %)
74(96.1%)
0.735
36(94.7 %)
35(92.1 %)
0.644
Muscular injury during ESD
11(28.9%)
35(45.4%)
0.089
11(28.9 %)
20(52.6 %)
0.036*
0(0%)
1(1.3%)
0.480
0(0 %)
1(2.6 %)
0.314
0(0%)
4(5.2 %)
0.153
0(0 %)
3(7.9 %)
0.077
4(10.5%)
11(14.3 %)
0.573
4(10.5 %)
4(10.5 %)
1.000
Post-procedure bleeding Perforation Chest pain
M AN U
TE D
EP
Procedure time, min (Median, IQR)
P value
SC
tESD group (n =38)
After PSM
AC C
Variables
RI PT
Table 2. Treatment outcomes before and after propensity score matching
cESD, conventional ESD; PSM, propensity score matching; tESD, tunnel ESD. *P < 0.05 in a single test of significance; however, correction for multiple testing may remove that nominal significance. 19
ACCEPTED MANUSCRIPT
Table 3. Univariate and multivariate regression analysis for procedure time 95%CI
RI PT
OR Univariate regression analysis
19.25
Location of lesions Perforation (yes or no)
0.84 3.28
Use of tunnel method (yes or no)
2.43
Multivariate regression analysis
2.18
3.42 8.74
3.10 1.25
0.294 0.147 0.008*
169.78
0.27 0.33
2.64 32.54
1.08
5.45
1.32 1.30
8.85 58.5
0.772 0.310 0.031*
0.011 0.026
EP
Use of tunnel method (yes or no) Specimen size, mm (<40 or≥ 40)
SC
Specimen size, mm (<40 or ≥40)
0.71 0.22
M AN U
1.48 0.53
TE D
Age (<60 y or ≥60 y) Sex (male or female)
P value
AC C
*P < 0.05 in a single test of significance; however, correction for multiple testing may remove that nominal significance.
20
ACCEPTED MANUSCRIPT
Table 4. The subgroup analysis according to specimen size ＜40 mm
cESD group (n =18) 67.5 (43.3-96.5)
0.012
34.0 (21.0-44.0) 0.044
36.0 (27.0-44.0)
59.5 (32.0-86.0)
0.008
RI PT
tESD group (n =19) 44.0 (36.0-52.0)
P value
AC C
EP
TE D
M AN U
Procedure time, min (Median, IQR) Submucosal dissection time, min 25.0 (19.0-33.0) (Median, IQR)
cESD group P value (n =20) 43.0 (28.0-53.5) 0.036
SC
tESD group (n =19) 32.5 (25.0-39.0)
≥40 mm
21
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT