European experience of colorectal endoscopic submucosal dissection: a systematic review of clinical efficacy and safety

Abstract

Background: Endoscopic submucosal dissection (ESD) is an advanced method allowing en bloc resection of large and complex lesions in colon and rectum. Herein, the European experience of colorectal ESD was systematically reviewed in the medical literature to determine the clinical efficacy and safety of colorectal ESD in Europe.

Material and methods: A systematic search of PubMed for full-text studies including more than 20 cases of colorectal ESD emanating from European centres was performed. Data were independently extracted by two authors using predefined data fields, including efficacy and safety.

Results: We included 15 studies containing a total of 1404 colorectal ESD cases (41% in the colon) performed between 2007 and 2018. Lesion size was 40 mm (range 24–59 mm) and procedure time was 102 min (range 48–176 min). En bloc resection rate was 83% (range 67–93%) and R0 resection rate was 70% (range 35–91%). Perforation rate was 7% (range 0–19%) and bleeding rate was 5% (range 0–12%). The percentage of ESD cases undergoing emergency surgery was 2% (range 0–6%). Additional elective surgery was performed in 3% of all cases due to histopathological findings showing deep submucosal invasion or more advanced cancer. The recurrence rate was 4% (range 0–12%) after a median follow-up time of 12 months (range 3–24 months).

Conclusions: This review shows that ESD is effective and safe for treating large and complex colorectal lesions in Europe although there is room for improvement. Thus, it is important to develop standardized and high-quality educational programs in colorectal ESD in Europe.

Introduction

Endoscopic removal of neoplasia in the colon and rectum has been proven to reduce incidence and mortality of colorectal cancer (CRC) [1,2]. Endoscopic submucosal dissection (ESD) is an advanced endoscopic technique allowing en bloc resection of large and complex lesions in the gastrointestinal tract. En bloc resection has the advantage of yielding correct histological diagnosis and reduce tumour recurrence as compared to piecemeal resection, which can help to avoid unnecessary surgery [3–6]. Initially, ESD was developed in Japan to treat superficial neoplasia in the stomach but later ESD practice was extended to also include lesions in the colon and rectum. Although being technically challenging, ESD is now established as standard care of patients with large and complex colorectal lesions with suspicion of superficially invasive cancer in many Asian countries and the efficacy has been well documented in numerous reports [5,7–9]. Due to the relatively difficult manoeuvrability of the colon and its thin wall, colorectal ESD is associated with increased procedure time and risk of perforation, which has contributed to the slow dissemination of ESD in western countries [10]. Major obstacles for implementing ESD in Europe comprise long learning curve, high risk of complications, lack of structured training programs, few suitable starting cases in the stomach, and lack of experts [11–17].

European Society of Gastrointestinal Endoscopy (ESGE) suggests that colorectal ESD should be considered for lesions with high suspicion of superficial invasion according to morphologic and endoscopic criteria, especially for tumours larger than 20 mm [18]. Moreover, according to ESGE guidelines, ESD can be considered for colorectal lesions that would not otherwise be optimally and radically removed by snare-based methods [18]. Nonetheless, endoscopic mucosal resection (EMR) is currently the most widely used minimally invasive technique for non-invasive colorectal tumours in Europe although a growing body of evidence indicates that with qualified training and adequate patient selection, ESD could be equally safe as other minimally invasive alternatives [16]. Thus, it is likely that the use of colorectal ESD will increase in Europe in the future. In order to summarize the experience of colorectal ESD in Europe, we have performed a systematic review of available publications on the clinical outcome and safety of colorectal ESD in European centres.

Material and methods

Study selection criteria

Publications containing patients with a diagnosis of non-pedunculated (type Is and/or II lesions according to Paris classification) colorectal tumours undergoing standard ESD in a European centre were included in this review. Data on hybrid ESD approaches and additional treatment methods, such as EMR, were not considered. Prospective and retrospective cohort studies, published as full-text papers, including more than 20 patients were considered. We performed a comprehensive literature search of PubMed for studies accepted or published up to July 2018. Our search query for PubMed was (‘endoscopic submucosal dissection’ OR ‘endoscopic submucosal resection’ OR ‘submucosal dissection’ OR ‘ESD’ AND (‘colon’ OR ‘colorectal’ OR ‘colo*’ OR ‘large bowel’. This search strategy identified 759 publications. All investigators screened all titles and abstracts for relevance to our study and reviewed the full-text of these articles and applied our predefined inclusion/exclusion criteria. In cases where relevant references in selected papers were found were these also considered for inclusion in the study. Studies with fewer than 20 patients and studies not published in the English language were excluded. Studies with carcinoid lesions or only rectal tumours were excluded. Animal studies, letters, meeting abstracts, commentaries and reviews were also excluded. After application of these search criteria, 19 full-text articles on colorectal ESD emanating from European centres were identified.

Extraction of data

A predefined standardized data extraction protocol was used to extract data from the selected studies, including publication information (author name, year of publication, type of publication, study characteristics (country, name of endoscopic centre, design, number of patients, patient age and gender, year of data collection, demographics), lesion characteristics (anatomical location, numbers, size, macroscopic and microscopic details), ESD procedural details (en bloc and R0 resection rate, duration of procedure), complications (perforation and bleeding rates), follow-up time and recurrence rates. R0 resection was defined as removal of lesions with vertical and horizontal margins free of neoplasia. All perforations regardless of clinical consequences were included. Only bleedings causing abortion of the procedure or requiring transfusion, endoscopic re-exploration or hospitalization were included.

Study endpoints

We analysed measures of efficacy and complications associated with colorectal ESD. The primary outcome was the R0 resection rate on a per-lesion basis. Secondary outcomes included en bloc resection, duration of ESD procedure, number of patients undergoing surgery after ESD, complications and recurrence rate. For all endpoints, the rates were evaluated as percentage of number of tumours operated. Data are given as median (range). Pearson correlation was used to evaluate correlation between the percentage of colonic localization, lesion size or procedure duration on one hand and en bloc or R0 resection rates on the other hand.

Results

Literature search

Searching the literature identified 19 full-text articles on colorectal ESD emanating from European centres. These 19 full-text articles were scrutinized in detail and another four were excluded. Two single institution studies were excluded due to obvious overlap with subsequent studies from the same institute [19,20]. The latest published or the one with highest numbers of colorectal ESD cases were included herein. In addition, one study containing only out-patients and excluding 27 patients with planned or unplanned hospitalization from analysis were also excluded in this review [21]. One study was not included because the R0 resection rate of colorectal ESD could not be determined [22]. Thus, after application of our inclusion and exclusion criteria, 15 European full-text studies of colorectal ESD published between 2007 and 2018 were included in this review (Table 1) [15,23–36]. All these included studies were case series. A total of 1404 colorectal lesions in patients (61% male) with a median age of 67 years (range 64–72 years) undergoing colorectal ESD was included (Table 1). There was a wide variation of the number of ESD cases between the investigations ranging from 23 to 301 cases with a median number of 58 cases (Table 1). Colonic localization of ESD lesions was 41%. Median lesion size was 40 mm (range 24–59 min) and median procedure time was 102 min (range 48–176 mm) (Table 1).

Table 1. European studies on colorectal ESD.

Author (origin)	Year (Ref)	ESD (N)	Size (mm)	Colon (%)	En Bloc (%)	R0 (%)	Duration (min)	Bleeding (%)	Perforation (%)	Emergency surgery (%)	Additional surgery (%)	Cancer (%)	Follow-up (months)	Recurrence (%)
Hurlstone (UK)	2007 [27]	42	24	67	79	74	48	12	2	2	NS	5	6	11
Farhat (France)	2011 [24]	85	32^a	33^a	67	62	105	11	18	6	NS	18^a	NS	NS
Probst (Germany)	2012 [27]	76	46	17^a	82	70	176	8	1	0	14	17	24	9
Thorlacius (Sweden)	2013 [28]	29	28	41	72	69	142	3	7	0	7	17	5	0
Bialek (Poland)	2013 [30]	34	NS	18	88	91	82	NS	NS	0	6	6	NS	3
Bialek (Poland)	2014 [29]	37	37	32	87	81	70	11	0	5	0	NS	NS	NS
Sauer (Germany)	2016 [31]	155	41^a	65^a	88	63	127^a	3^a	19	0	2	7^a	7^a	4^a
Barret (France)	2017 [32]	142	NS	37	85	54	NS	NS	NS	NS	NS	NS	6^a	6
Spychalski (Poland)	2017 [15]	228	40	50	79	79	104	5	8	NS	2	18	12	3
Iacopini (Italy)	2017 [33]	140	NS	79	83	71	74	1	3	1	4	11	NS	NS
Rönnow (Sweden)	2018 [26]	29	40	45	83	69	89	3	14	3	21	100	13	3
Emmanuel (UK)	2018 [20]	58	59	31	93	91	125^a	7	9	2	3	14	12	5
Milano (Italy)	2018 [35]	23	26	48	83	35	120	0	9	4	0	9	24	12
Rönnow (Sweden)	2018 [25]	301	40	43	80	69	98	2	6	2	0	0	12	3
Ebigbo (Germany)	2018 [36]	25	NS	12	92	91	100	4	4	4	0	16	3	0
Median (Range)		58 (23–301)	40 (24–59)	41 (12–79)	83 (67–93)	70 (35–91)	102 (48–176)	5 (0–12)	7 (0–19)	2 (0–6)	3 (0–21)	15 (0–100)	12 (3–24)	4 (0–12)

NS: Not stated in article. Ref: number in the reference list.

^aArticle containing more than conventional colorectal ESD, data based on all patients in the study.

Clinical efficacy

En bloc and R0 resection were defined in a similar manner and given in all 15 studies. En bloc resection rate was 83% with a range between 67–93% (Table 1). Moreover, R0 resection rate was 70% ranging from 35 to 91% (Table 1). There was no correlation between the percentage of colonic localization, lesion size or procedure duration on one hand and en bloc or R0 resection rates on the other hand (Table 1).

Complications and follow-up

Perforation and bleeding were the most frequent complications in all studies. Notably, one study [23] reported three cases of ileus and another study [33] described 10 cases of electrocoagulation syndrome. Perforation and bleeding rate were not given in two of the studies [30,32]. In the remaining 13 reports median perforation rate was 7% with a range between 0 and 19% (Table 1). The great majority of perforations were closed by use of endoclips and treated conservatively. The percentage of ESD cases undergoing emergency surgery was 2% (range 0–6%) which was predominately due to perforation. Significant bleeding leading to abortion of the procedure, intervention (transfusion or endoscopic re-exploration) or hospitalization were reported in 13 studies and amounted to a median rate of 5% with a range between 0 and 12%. (Table 1). Data on additional elective surgery was given in 12 studies with a median rate of 3%, which was mainly due to histopathological findings showing deep submucosal invasion or more advanced cancer. The overall median frequency of cancer in lesions treated with ESD was 15% (Table 1). Endoscopic follow-up was recorded in 12 studies (Table 1). Recurrence rate was provided on 12 studies and median recurrence rate was 4% ranging between 0 to 12%. The median follow-up time was 12 months (range 3–24 months).

Discussion

Our review of the European experience of colorectal ESD revealed a RO resection rate of 70% and a low rate of complications requiring emergency surgery. Moreover, the risk of recurrence was only 4% after 12 months of follow-up. These findings show that ESD is an effective and safe procedure to remove large and complex colorectal lesion but also that there is significant room for refinement of clinical efficacy in Europe.

15 published full-text articles were identified containing 1404 cases of colorectal ESD. Clinical efficacy was determined by the rate of en bloc and R0 resection rate. The median rate of en bloc resection was 83% with a relatively low variation between studies. R0 resection rate was 70% but there was a larger variation between the European centres compared to en bloc resection, suggesting that patient selection or details in the resection technique might require local improvement. In this study, we did not find any relation between tumour size, location or procedure time on one hand and clinical efficacy on the other hand. Although these data show that colorectal ESD can also be an effective procedure outside of Asia some caution should be emphasized for the possibility that the published literature is biased towards more successful results and that less successful outcomes might not have been reported in the literature. Nonetheless, the European en bloc and R0 resection rates are lower than those reported from Asian countries. One recent meta-analysis found that en bloc and R0 resection rates were 93% and 86%, respectively, in reports on colorectal ESD from Asia [13]. This is perhaps not surprising knowing that ESD was invented in Japan and spread early to other Asian countries. More importantly, these data underline the fact that enhancing clinical efficacy of colorectal ESD in Europe is an important challenge. Programs for systematic training and concentration of ESD cases to expert centres are pivotal for successful implementation of colorectal ESD in Europe. In this context, it should be mentioned that so-called hybrid-ESD (mucosal incision and snare resection) does not seem to be useful to improve clinical efficacy compared to standard ESD [13].

One significant obstacle for the dissemination of colorectal ESD in western countries has been the fear of high risk of complications [10]. We found that perforation and bleeding were the most common complications associated with ESD. Notably, one study reported three cases of ileus due to ESD but it is difficult to explain the causality of this observation [23]. One investigation reported several cases of abdominal pain due to electrocoagulation [33]. This complication is rarely mentioned in ESD publications and could thus be underestimated. We observed a perforation rate of 7% in the European studies of colorectal ESD. The great majority of perforations were treated by use of endoclips and managed conservatively. Only 2% of all patients undergoing colorectal ESD required emergency surgery. Additional elective surgery was also performed in 3% of all colorectal ESD cases, which was mainly due to non-radical resection of cancer or due to risk of concomitant lymph node metastasis. The percentage of cancers in the colorectal ESD cases was found to be 15%, which is similar to levels reported in ESD studies from Asia [3,5]. In the present study, we found that bleeding as a complication occurred in 5% of all ESD cases in the colon and rectum. These bleedings rarely required emergency surgery but rather endoscopic re-exploration, transfusion or observation were sufficient in most cases. Complication rates in Asian studies have been reported to amount to 5% perforation, 2.4% bleeding and 1% emergency surgery, which are similar to those observed in the present study, indicating that colorectal ESD in Europe is a relatively safe method at least in expert centres [13,37].

One of the most appealing features of ESD is the relatively low risk of tumour recurrence [5,37–40]. After a median follow-up time of 12 months we found a local recurrence rate of 4%, which is clearly higher than that observed in Asian countries (0.4%) [37]. The reason behind this 10-fold higher recurrence rate in Europe compared to Asia can not merely be explained by the relatively small differences in R0 resection rates and requires further studies in the future. Nonetheless, this level of recurrence is clearly lower than that observed using EMR being close to 14% [41]. Although ESD is superior to EMR in terms of R0 resection and recurrence rates for large colorectal lesion (>2 cm) the recurrence rate after resection of smaller tumours using EMR and ESD is similar, which makes EMR more suitable for colorectal lesions smaller than 2 cm [18,42].

There are some limitations to this study. First, the number of studies is relatively small on colorectal ESD in Europe which is associated with heterogeneity between studies and makes detailed analysis difficult. There is a possibility that only the best results are published as case series causing a risk of bias in reviews containing only case series. Moreover, our study excluded published abstracts and meeting reports, which could have expanded this study but we considered that peer-reviewed publications can serve as a more solid base for further developments of colorectal ESD in Europe. It is well-known that ESD in the proximal colon is more difficult and results in fewer R0 resections compared to ESD of rectal lesions [25]. However, most studies herein did not stratify clinical efficacy and complication rates according to the localization in the colorectum. Lastly, knowing that recurrence rate is time-dependent, large variations in follow-up time between and within studies complicates conclusions about actual recurrence rate after colorectal ESD in Europe.

In conclusion, our review shows that ESD is effective and safe for treating large and complex colorectal lesion in Europe. However, when comparing with the clinical efficacy of colorectal ESD in Asian countries it is clear that there is significant room for improvement in Europe. This review underlines the importance to develop standardized and high-quality educational programs in colorectal ESD in Europe.

Disclosure statement

No potential conflict of interest was reported by the authors.