- Email: [email protected]
Endoscopic full-thickness resection for early colorectal cancer
ORIGINAL ARTICLE
Endoscopic full-thickness resection for early colorectal cancer Armin Kuellmer, MD,1,* Julius Mueller,1,* Karel Caca, MD,2 Patrick Aepli, MD,3 David Albers, MD,4 Brigitte Schumacher, MD,4 Anne Glitsch, MD,5 Claus Schäfer,6 Ingo Wallstabe, MD,7 Christopher Hofmann, MD,8 Andreas Erhardt,9 Benjamin Meier, MD,2 Dominik Bettinger, MD,1 Robert Thimme, MD,1 Arthur Schmidt, MD1, the FTRD study group Freiburg, Germany
Background and Aims: Current international guidelines recommend endoscopic resection for T1 colorectal cancer (CRC) with low-risk histology features and oncologic resection for those at high risk of lymphatic metastasis. Exact risk stratification is therefore crucial to avoid under-treatment as well as over-treatment. Endoscopic full-thickness resection (EFTR) has shown to be effective for treatment of non-lifting benign lesions. In this multicenter, retrospective study we aimed to evaluate efficacy, safety, and clinical value of EFTR for early CRC. Methods: Records of 1234 patients undergoing EFTR for various indications at 96 centers were screened for eligibility. A total of 156 patients with histologic evidence of adenocarcinoma were identified. This cohort included 64 cases undergoing EFTR after incomplete resection of a malignant polyp (group 1) and 92 non-lifting lesions (group 2). Endpoints of the study were: technical success, R0-resection, adverse events, and successful discrimination of high-risk versus low-risk tumors. Results: Technical success was achieved in 144 out of 156 (92.3%). Mean procedural time was 42 minutes. R0 resection was achieved in 112 of 156 (71.8%). Subgroup analysis showed a R0 resection rate of 87.5% in Group 1 and 60.9% in Group 2 (p < .001). Severe procedure-related adverse events were recorded in 3.9% of patients. Discrimination between high-risk versus low-risk tumor was successful in 155 of 156 cases (99.3%). In Group 1, 84.1% were identified as low-risk lesions, whereas 16.3% in group 2 had low-risk features. In total, 53 patients (34%) underwent oncologic resection due to high-risk features whereas 98 patients (62%) were followed endoscopically. Conclusions: In early colorectal cancer, EFTR is technically feasible and safe. It allows exact histological risk stratification and can avoid surgery for low-risk lesions. Prospective studies are required to further define indications for EFTR in malignant colorectal lesions and to evaluate long-term outcome. (Gastrointest Endosc 2019;-:1-10.)
Abbreviations: CRC, colorectal cancer; EFTR, endoscopic full-thickness resection; ESD, endoscopic submucosal dissection; FTRD, full-thickness resection device; R0, microscopically complete resection; T1, Tumor stage according to TNM classification. DISCLOSURE: K. Caca received lecture fees and study grants from Ovesco Endoscopy. A. Schmidt received lecture fees from Ovesco Endoscopy. All other authors disclosed no financial relationships relevant to this publication. *Drs Kuellmer and Mueller contributed equally to this article. Copyright ª 2019 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 https://doi.org/10.1016/j.gie.2018.12.025 Received September 29, 2018. Accepted December 29, 2018.
www.giejournal.org
Current affiliations: Department of Medicine II, Medical CenterdUniversity of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg (1), Department of Gastroenterology and Oncology, Klinikum Ludwigsburg, Ludwigsburg, Germany (2), Gastroenterology and Hepatology Unit, Luzerner Kantonsspital, Lucerne, Switzerland (3), Department of Gastroenterology, Elisabeth-Krankenhaus Essen, Teaching Hospital of the University of Duisburg-Essen, Essen (4), Department of Surgery, University Hospital Greifswald, Greifswald, (5), Medical Clinic II, Klinikum Neumarkt, Neumarkt (6), Department for Gastroenterology, Hepatology, Diabetology und Endocrinology, Klinikum St. Georg gGmbH, Leipzig (7), Department for Gastroenterology, Katholisches Klinikum Mainz, Mainz (8), Department for Gastroenterology, Hepatology und Diabetology, Petrus-Krankenhaus, Wuppertal, Germany (9). Reprint requests: Arthur Schmidt, Department of Medicine II, Medical Center, Faculty of Medicine, University of Freiburg, Hugstetter Strasse 55, 79106 Freiburg, Germany.
Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 1
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
Colorectal cancer (CRC) is the third most commonly diagnosed cancer and the second leading cause of cancer death.1 In recent years, the incidence of CRC as well as cancer-associated deaths could be reduced significantly. This is mostly due to an improved screening strategy that allows detection and treatment in earlier stages. For early (T1) CRC with a low risk of lymphatic metastasis, current international guidelines recommend endoscopic treatment, whereas oncologic surgical resection is recommended for lesions with high-risk histology features.2-4 It is therefore crucial to correctly discriminate low-risk versus high-risk lesions to assign patients to the best treatment strategy. However, low-risk versus high-risk lesions are difficult to discriminate before resection because criteria are based on histology features. Moreover, T1 cancer is often found incidentally after resection of non-suspicious polyps, and further treatment after incomplete resection is not welldefined. Endoscopic full-thickness resection (EFTR) is an emerging technique for removal of non-lifting colorectal lesions but data on resection of malignant tumors are limited.5-12 Here we report on efficacy and safety of EFTR for staging and treatment of early CRC.
METHODS Study design and patients We conducted a retrospective observational study. Data of 1234 patients who had undergone EFTR with the fullthickness resection device (FTRD) system (Ovesco Endoscopy, Tübingen, Germany) at 96 centers were screened for eligibility. Inclusion criteria were: (1) repeat resection of incidentally found malignant polyps after incomplete endoscopic resection; (2) non-lifting lesions initially classified as benign with final histology of adenocarcinoma; and (3) non-lifting lesions with known malignant histology before EFTR. Exclusion criteria were (1) malignant histology other than adenocarcinoma (eg, neuroendocrine tumors); (2) no evidence of malignancy before or after EFTR; (3) lesions treated with combination of EFTR and other endoscopic resection techniques; and (4) incomplete data on histology results in the resected EFTR specimen. Study data were collected anonymously from the following sources (Fig. 1): the German Online FTRD Registry, which was initiated in September 2015 by Ovesco Endoscopy within a post-marketing clinical follow-up. It is accessible to all German endoscopists trained in full-thickness resection with the FTRD. Until July 2018, it was made up of data from 950 FTRD cases from 83 centers. We further screened 181 patients who had undergone EFTR within the WALL-RESECT study (NCT 02362126).10 Data were obtained from the coordinating study center in Ludwigsburg, Germany. To avoid the risk of doublecounting, all centers participating in the WALL-RESECT 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
study were thoroughly instructed not to enter data regarding these patients into the registry. During data collection, we confirmed this policy again with the responsible physicians at the participating centers. An additional 30 patients (103 screened) undergoing EFTR for CRC were identified from 7 endoscopic centers in Germany and Switzerland that were neither documented in the FTRD registry nor in the WALL-RESECT study. Centers with special expertise in full-thickness resection were contacted and asked for cases meeting the inclusion criteria. Written informed consent on the EFTR procedure was obtained in all cases. Informed consent was obtained from participants of the WALL-RESECT study. For all other patients, including those from the German Registry, no specific written consent had been obtained. Data acquisition and analysis for our study was approved by our institutional review board with the requirement of using only anonymized data.
Endpoints and definitions The primary endpoints were (1) technical success, defined as reaching the lesion with the FTRD, successful clip application, and macroscopically complete resection of the lesion with the integrated snare, and (2) R0 resection: complete histologic resection, defined as tumor-free lateral and deep resection margins. Cases with no evidence of neoplastic tissue in the EFTR specimen (after incomplete removal with standard methods) were also considered as R0. Secondary endpoints were (1) procedure-related adverse events. Mild adverse events were defined as those not requiring medical or repeated endoscopic intervention and not prolonging hospital admission. Moderate adverse events were defined as those requiring medical or repeated endoscopic intervention and/or prolonging hospital admission. Severe adverse events were defined as those requiring surgical therapy and/or potentially life threatening. Histology findings of EFTR specimens confirm the following: (1) histologically confirm full-thickness resection, (2) successful discrimination of low-risk versus high-risk tumors, (3) evidence of residual or recurrent tumors in endoscopic follow-up, (4) requirement for oncologic surgical resection, (5) evidence of residual tumor and lymph node involvement in the surgical resection specimen, T-stage and N-stage. According to international guidelines,2-4 patients were classified as low risk (in terms of risk for lymph node metastasis) if all of the following histology features were present: submucosal infiltration <1000 mm, no lymphovascular infiltration, no infiltration of blood vessels, well or moderate differentiation/grading, and microscopically complete resection (R0). Patients who did not meet all of these criteria were classified as high risk. www.giejournal.org
Kuellmer et al
Endoscopic full-thickness resection for early colorectal cancer
Patients screened n = 1234 Patients excluded, n = 1078 No adenocarcinoma, n = 770 NET/ NEC, n = 59 Hybrid technique, n = 6 Location other than colon/ rectum, n = 6 Incomplete information, n = 237
Patients enrolled n = 156
Indication
Re- resections (Group 1)
Non- lifting lesions (Group 2)
n = 64
“Low risk,” n = 53
n = 92
“High-risk,” n = 10
“High-risk,” n = 77
“Low- risk,” n = 15
-R0. but sm >1000, n = 4 -Rx and “high-risk,” n = 2 -Rx, otherwise low-risk n=1 -Rx, no further information, n=3
- R0, but sm >1000, n = 42 -Rx and “high- risk,” n = 25 -Rx, otherwise “low-risk,” n=5 - Rx, no further information, n=5
Surgery* n=9
3 month – F/U n = 37
Recurrent lesions n=2
Surgery* n = 44
3 month – F/U n=0
Recurrent/ residual lesions
3 month – F/U* * n = 19
3 month – F/U n = 13
Recurrent lesions n=1
Residual lesions n=1
n=0
End. therapy, n = 1 radiochemoth., n = 1
End. therapy n=1
Lost in F/U n=1
Figure 1. Patient enrollment, distribution in groups 1 and 2, and outcome are shown. *Surgery due to oncologic re-resection. ySeven patients refusing surgery. Three patients not further treated due to comorbidities. Four patients received endoscopic treatment without information about reasons. Five patients with no further information. Three patients were not treated further because of comorbidities. Four patients received endoscopic treatment, without information about reasons. In 5 patients, no further information was available. NET, Neuroendocrine tumor; NEC, neuroendocrine carcinoma; R0, microscopically complete resection; F/U, Endoscopic follow-up (Patients not undergoing surgery); Rx, microscopically incomplete resection; End, endoscopic; (Patients, not undergoing surgery).
www.giejournal.org
Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 3
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
Figure 2. Key steps of the endoscopic full-thickness resection procedure are shown. A, B, C, Repeat resection after an incomplete standard resection of a T1 carcinoma (group 1). A, The lesion is marked. B, The lesion is pulled into the over the scope–loaded cap. C, The lesion is resected after clip application. Histology revealed only scar tissue. The patient was classified as low risk. D, A non-lifting lesion in the cecum without prior therapy (group 2). E, The lesion was resected. F, Image of the full-thickness resection specimen. Histology showed a T1 tumor with no lymphovascular infiltration but deep submucosal infiltration and venous infiltration. Although a microscopically complete resection was achieved, the patient was classified as high risk and referred to surgery.
FTRD and the EFTR procedure
Data management and statistical analysis
For resection of lesions, the FTRD (Ovesco Endoscopy) was used in all cases. For description of the device, technique, and peri-interventional management, we referred to previous studies9-10 and Figure 2.
The study data were collected and analyzed at the University of Freiburg and were used exactly as presented by the contributing investigators who were responsible for the correctness of the data at their investigation sites. Continuous
4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
www.giejournal.org
Kuellmer et al
Endoscopic full-thickness resection for early colorectal cancer
TABLE 1. Baseline characteristics
Variable Age, mean ( SD), y
TABLE 2. Procedure data and adverse events Entire cohort (n [ 156) 72 9.6
Sex, no. (%) Female
55 (35.3%)
Male
101 (64.7%)
Indication for EFTR, no. (%) Repeat resection of malignant polyp (group 1) Non-lifting lesion (group 2)
64 (41%) 92 (59%) 19 (20.7%)
Initial classified as (benign) adenoma
73 (79.3%)
Location, no. (%) Cecum
11 (7.1%)
Ascending colon
17 (10.9%)
Transverse colon
18 (11.5%)
Descending colon
12 (7.69%)
Sigmoid colon
48 (30.8%)
Rectum
50 (32.1%) 20 (2-45)
SD, Standard deviation; EFTR, endoscopic full-thickness resection.
variables were reported as mean with standard deviation, whereas categorical variables were expressed as frequencies and percentages (in parentheses) unless stated otherwise. For categorical variables, differences were determined by using c2 tests or the Fisher exact test, as appropriate. P values < .05 were considered significant. No interim analysis was performed. Data collection was performed with Microsoft Excel 2016 for Mac Os (version 15.21; Microsoft, Redmond, Wash). Statistical analyses were performed with SPSS (version 24.0; IBM, New York, NY) and GraphPad Prism (version 6; GraphPad Software, San Diego, Calif).
RESULTS Patients A total of 1234 patients at 96 centers were screened for eligibility, and 156 patients were included in the study. Mean age was 72 ( 9.6) years. Indications for the FTRD intervention were repeat resection of malignant polyps (group 1, 41%) and non-lifting lesions (group 2, 59%). Group 2 made up 21% (n Z 19) of adenocarcinomas on which biopsies were performed; the rest (n Z 73, 79%) were initially classified as benign. Most lesions (62.9%) were located in the rectosigmoid colon. Median lesion size was 20 mm (range 2-45 mm). Patient and lesion characteristics are shown in Table 1.
Procedural data (entire cohort) Technical success was achieved in 144 of 156 (92.3%) patients (Table 2). Reasons for technical failure (n Z 12) www.giejournal.org
Procedure time, mean ( SD) min*
42.1 26.6
Technical success,y no. (%)
144 (92.3%)
Adverse events, total
21 (14.1%)
Mild adverse eventsz Moderate adverse eventsx
Known carcinoma
Lesion size, median (range), mm
Entire cohort (n [ 156)
Variable
7 (4.5%) 8 (5.1%)
Bleeding with successful endoscopic treatment
6/8 (75%)
Infection with a need for antibiotics
2/8 (25%)
Severe adverse eventsk
6 (3.9%)
Perforation (1x immediate, 5x secondary)
6/6 (100%)
Immediate with successful endoscopic closure
1/6 (16.7%)
Secondary with a need for surgery
5/6 (83.3%)*
SD, Standard deviation. *Total procedure time including colonoscopy. yLesion successfully reached with FTRD, successful clip application and macroscopically complete resection. zManaged conservatively without intervention. xNeed for endoscopic hemostasis or antibiotics. kNeed for surgery and/or potentially life-threatening condition.
were dysfunction of the snare (10/12) and incorrect application of the FTRD clip (2/12). Mean total procedure time (including colonoscopy with marking of the lesion) was 42 minutes (standard deviation [SD] 26.6). In 33 cases the procedure time was not reported.
Histology The entire cohort. In the entire cohort of patients (n Z 156), R0 resection was achieved in 112 cases (71.8%). Full-thickness resection was histologically confirmed in 135 cases (86.5%). Definitive discrimination between high-risk and low-risk lesions by histology examination of the EFTR specimen according to the definition given previously was possible in all cases (99.4%) except one. A total of 68 cases (43.9%) were histologically classified as low risk, and 87 cases (56.1%) were classified as high risk. In 8 cases of high-risk tumors, not all histology criteria were available for analysis. Most high-risk lesions (43/79, 54%) were resected completely, but the submucosal infiltration depth was >1000 mm. In 29 of 79 cases (34%), resection was incomplete, and histology showed additional high-risk criteria. In 3 of 79 cases (4%), deep submucosal and lymphovascular infiltration appeared synchronously. Six of 79 cases (8%) had low-risk features but were not R0 resected. Detailed results of the entire cohort and subgroup analysis are shown in Tables 3 and 4. Group 1 (repeat resection of malignant polyp, n [ 64). In this subgroup, the R0 resection rate was achieved in 56 of 64 cases (87.5%). Histology findings were scar tissue in 49 of 64 (77%), residual adenocarcinoma in 14 of 64 (22%), and adenoma with low-grade interaepithelial dysplasia (LGIEN) in 1 of 64 (2%). According to the definition given previously, 53 patients Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 5
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
TABLE 3. Histology features of EFTR specimens
TABLE 4. Histologic features of high-risk tumors
Entire cohort (n [ 156)
Group 1 (n [ 64)
Group 2 (n [ 92)
Histology
Full-thickness resection, no. (%)
135 (86.5%)
54 (84.4%)
81 (88%)
Microscopically complete resection, no. (%)
112 (71.8%)
56 (87.5%)
56 (60.9%)
Variable
Histology Scar, no. (%)
Entire cohort (n [ 79)
Group 1 (n [ 7)
Group 2 (n [72)
R0 but sm >1000 mm
46 (58%)
4 (57%)
42 (58%)
Rx and other high-risk features
27 (34%)
2 (29%)
25 (35%)
Rx but otherwise lowrisk features
6 (8%)
1 (15%)
5 (7%)
R0, Microscopically complete resection; Rx, microscopically incomplete resection.
49 (31.4%)
49 (76.6%)
Adenoma, no. (%)
1 (0.6%)
1 (1.6%)
Carcinoma, no. (%)
106 (68%)
14 (21.9%)
92 (100%)
Low risk, no. (%)
68/155 (43.9%)
53/63 (84.1%)
15/92 (16.3%)
High risk, no. (%)
87/155 (56.1%)
10/63 (15.9%)
77/92 (83.7%)
EFTR, Endoscopic full-thickness resection.
(84.1%) were classified as low risk, whereas 10 patients (16%) had high-risk features. In 7 of these, all histology criteria were available for further workup: 4 of 7 (57%) had been resected completely (R0) but showed deep submucosal infiltration. In 2 of 7 (29%), resection was incomplete, and other high-risk criteria were present. In 1 of 7 (15%), resection was incomplete, but other histology features were low risk (Table 4). Group 2 (non-lifting lesions, n [ 92). The R0 resection rate for group 2 was 60.9% (56/92). Adenocarcinoma was present in all specimens (100%). Fifteen patients (16.3%) were classified as low risk, and 77 (83.7%) had high-risk features. In the high-risk group, T1 carcinoma was found in 55 of 77 cases (71.4%), whereas T2 or T3 carcinoma was diagnosed in 22 patients (28.6%). Seventy-two of 77 cases were available for further work-up because all histology features were present. In this high-risk group, 58% (42/72) had been resected completely but showed deep submucosal infiltration. Thirtyfive percent (25/72) could not be resected completely and met other high-risk criteria, and 7% (5/72) showed low-risk criteria but histological resection was incomplete (Rx/R1). Risk factors for incomplete resection (entire cohort). The R0 resection rate was significantly higher in group 1 than in group 2 (87.5% vs 60.9%; P < .001). We did not find any significant differences for smaller versus larger lesions. Moreover, there was no significant correlation between depth of tumor infiltration and the R0 resection rate (P 0.471). There was a trend toward higher rates of R0 resection in lesions in the rectum versus other lesions (P Z .058). The results are shown in Figure 3.
Adverse events In total, 21 procedure-related adverse events occurred (14.1%) (Table 2). Seven of these could be managed conservatively without intervention and were classified as 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
mild adverse events (4.49%). Moderate adverse events such as bleeding with the need for endoscopic hemostasis (n Z 6) or local peritonitis treated with antibiotics (n Z 2) occurred in 8 patients (5.1%). Severe adverse events (need for surgery and/or potentially life-threatening condition) were recorded for 6 patients (3.9%). In total, 6 perforations occurred. One perforation was diagnosed immediately during EFTR and was related to incorrect application of the clip. This perforation was closed successfully endoscopically with an overthe-scope clip. Five secondary perforations were observed. Median time to diagnosis was 7 days (range 1-8 days). All of those cases required surgical intervention.
Clinical outcome (entire cohort) Data on outcome of patients undergoing surgery or endoscopic follow-up are shown in Table 5. A total of 53 of 156 patients (34%) underwent oncologic resection because of high-risk histology features. Of those, 45 were available for detailed analysis. Five patients (11%) required surgery because of histologically incomplete resection (Rx/ R1) but had otherwise low-risk features. Twenty-two patients (49%) had incomplete EFTR and at least 1 additional high-risk criteria. In 18 cases (40%), the lesion was resected completely with the FTRD system but showed other highrisk features in histology. In 44 of the 53 cases undergoing oncologic resection, information about the histology of the surgical resection specimen was available. Residual tumor was found in 27 cases (61%), whereas in 17 cases (39%) no residual carcinoma was found. Ninety-eight of 156 patients (62%) did not undergo surgical resection. Of those, 65 patients (66%) showed lowrisk features and did not require further therapy. A total of 33 patients (34%) had high-risk criteria, but surgery was not performed because of the following reasons: refusal by the patient (n Z 18), comorbidity (n Z 5), and death caused by coexisting disease (n Z 1). In 9 cases, information about reasons for conservative treatment were not available. Of 98 patients not undergoing surgery, endoscopic follow-up was available in 69 (70.4%). Mean ( SD) follow-up was 15 weeks ( 11). The FTRD clip had spontaneously dislodged in 45 patients (69%). In 3 patients (4%) a www.giejournal.org
Kuellmer et al
Endoscopic full-thickness resection for early colorectal cancer
P = .537 P < .001
A
80 60 40 20
60 40 20
re-resection of malignant polyp
non-lifting lesions
B
P = .471
60 40 20 0 T1, sm<1000
10-20 mm
≤ 9 mm
>T1, sm<1000
D
> 20 mm
P = .058
100
R0 resection [%]
R0 resection [%]
P = .708
0
0
80
C
P = .754
80
R0 resection [%]
R0 resection [%]
100
100
80 60 40 20 0 rectum
non-rectum
Figure 3. Microscopically complete resection (R0) rate in various conditions is shown. A, Group 1 versus group 2. B, R0 regarding lesion size. C, R0 in deepness of submucosal infiltration. D, R0 in a lesion location in the rectum versus the rest. P < .05 was considered significant. R0, Microscopically complete resection; T1, according to TNM classfication; sm, submucosal infiltration depth measured in micrometers.
recurrent lesion after previous R0 resection was diagnosed. In 2 of those, a second FTRD procedure was performed, and histology showed adenoma without evidence of malignancy. The third patient developed recurrent carcinoma with distant metastasis and underwent chemoradiation. In 1 case (1.4%), a residual tumor was found, but the patient refused further therapy.
TABLE 5. Outcome of the patients undergoing oncologic resection and endoscopic follow-up Entire cohort (n [ 156)
Outcome Oncologic resection, no. (%)
53 (34%)
Residual tumor found
17/44 (38.6%)
Thereof N1
5 (18.5%)
No residual tumor found
DISCUSSION We reported on EFTR with an over-the-scope device for incompletely resected malignant polyps and non-lifting malignant colorectal lesions without prior attempt at resection. Our results demonstrate high technical efficacy and safety of this novel endoscopic technique. The study underlines the potential of EFTR to exactly discriminate between high-risk versus low-risk tumors to aid decision for the optimal individual treatment strategy. To our best knowledge, this is currently the largest study on EFTR for early colorectal cancer. Our cohort included 156 patients with malignant colorectal tumors undergoing EFTR. Technical success was 92.3%. Comparing these results with the literature is difficult because studies on EFTR for colorectal carcinoma are very rare. Besides some case reports, there is only 1 www.giejournal.org
Endoscopic follow-up obtained,* no. (%) Median time (min, range) Clip position (in situ vs departed) Residual/recurrent lesion
27/44 (61.4%) 69/98 (70.4%) 12 (1-64) 20 (30.8%) vs 45 (69.2%) 4 (5.8%)
N1, Histologic evidence of lymphnode involvement in surgical resection specimen. *Referring to patients who did not undergo surgical resection.
case series of 6 patients undergoing EFTR for early rectal cancer.13-15 Other cases are reported in larger studies investigating EFTR for various indications.5-12 In these studies, data on success and efficacy usually were calculated for the whole cohort and not specifically for patients with malignant lesions. Reported technical success rates of EFTR for other indications (mainly non-lifting adenomas) range from 75% to 100%5-12 and are in line with our Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 7
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
data. The mean ( SD) procedure time of 42 minutes ( 26.6 minutes) in our study was also in the range of procedure times in other studies.10,12 Our data therefore suggest that the presence of malignancy does not negatively influence the feasibility of EFTR with the FTRD. The total R0 resection rate in our study was 71.8%. This is below the rate reported in the WALL-RESECT study for the total cohort (76.9%), but subgroup analysis of adenocarcinomas in the WALL-RESECT study showed a rate similar to ours (72%).10 Notably, in 29% (9/31) of patients with histology classified as incomplete, no residual tumor was found in the surgical specimen. Hence, in those cases the tumor margins were positive even though complete resection had been achieved. If these specimens were reclassified as R0, as done in another study,11 our overall R0-rate would rise to 78% (89% in group 1, 70% in group 2). In contrast to the WALL-RESECT study, we did not find a significant correlation of lesion size with R0-resection rate. One reason for this difference may be that lesion size has not been determined in a standardized way in our patient cohort. Moreover, the WALL-RESECT study had a different patient population including various indications for EFTR with only a minority being adenocarcinomas. We did observe a significant difference of the R0 resection rate between group 1 (repeat resection of malignant polyps, 87.5%) and group 2 (non-lifting lesions, 60.9%). This may be because group 1 mainly consisted of patients with low-risk tumors and small resection scars, whereas the majority of patients in group 2 proved to have advanced tumors. Data on R0 resection rates of standard endoscopic resection methods for T1 tumors are scarce. Asamaya et al16 showed a 92% R0 resection rate in a series of 37 patients. Ito et al17 analyzed local recurrences after endoscopic resection. R0 resection rates for EMR and endoscopic submucosal dissection (ESD) were as low as 28.6% and 67%, respectively. In a meta-analysis comparing EMR and ESD for resection of colon neoplasms, the R0 resection rate for EMR was 42.3% compared with 80.3% for ESD. This rate accounted for all lesions (adenoma/ adenocarcinoma) but was not further specified for adenocarcinoma alone. The analysis also reported that noncurative resection (8% pooled rate) was mainly (60%) due to invasive cancer.18 Whether cancer histology showed highrisk features within a complete resection or if the resection itself was incomplete was not specified.18 Hence, the R0 resection rate in our study is higher than reported rates of EMR and in the range of rates for ESD. We observed procedure-related adverse events in 14% of patientsd3.8% were classified as major adverse events, and surgical therapy due to perforation was required in 3% of patients. This indicates that colorectal EFTR for malignant indications is safe. The rate of major adverse events, including perforation, is in line with reported data on EFTR for other indications (0%-5%).6-12 In contrast, perforation rates for ESD in non-lifting lesions are reported to 8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
be as high as 15%.19 Perforation rates for EMR are significantly lower (2.5%), but the technique is generally not recommended for malignant and/or non-lifting lesions. Our data indicate that EFTR for colorectal cancer is feasible and safe. But what is its clinical significance? Current international guidelines recommend endoscopic resection of colorectal malignant tumors that are considered low risk.2-4 For those lesions, local treatment is sufficient because the risk of lymph node metastasis or local recurrence is very low.20 If complete resection of these lesions is achieved, the 5-year recurrence-free-survival rate is reported to be 98%.21 On the other hand, oncologic resection with lymph node dissection is the treatment of choice for early colorectal cancer with histology proven high-risk features. Surgery is effective, but more invasive, and it is associated with increased morbidity and mortality compared with endoscopic resection.22 Therefore, correct discrimination between low-risk and high-risk tumors is crucial to assigning patients to the optimal treatment strategy. In our study cohort, risk status was initially unclear in all patients and could be clarified in 99.4% of patients by means of EFTR. This underlines that reliable pretherapeutic discrimination between low-risk and high-risk T1-CRC usually is not possible. This classification is based on histologic criteria such as submucosal infiltration depth or infiltration of lymphatic vessels, which requires a thorough microscopic examination of the whole lesion and cannot be determined reliably by endoscopy with or without a biopsy. Our study cohort included 64 patients who had undergone incomplete resection of incidentally found malignant polyps (group 1). Those patients are often referred to surgery because local endoscopic repeat treatment usually is difficult because of scarring.23 EFTR with the FTRD allows resection of non-lifting, pretreated lesions as shown in multiple retrospective studies and 1 large, prospective study.10 In our study, 84.1% of patients were finally classified as low risk and were identified correctly not to be candidates for oncologic surgical resection. In other words, EFTR obviated the need for surgery in the majority of these patients and may therefore be the method of choice for this indication. The second subgroup (group 2) in our study included 92 patients with non-lifting lesions. The majority of the lesions had primarily been classified as non-malignant by the endoscopist, based on macroscopic criteria, and biopsies were not performed, whereas 19 (21%) patients had biopsy-proven carcinomas. The vast majority (83%) of lesions was finally classified as high risk, and surgical resection was recommended. However, a primary EFTR approach, in our opinion, is still justified for the following reasons: the 92 patients in this group represented only a small subgroup of all patients undergoing EFTR for non-lifting lesions. As previously shown in the WALLRESECT study, the majority of non-lifting lesions with indication for EFTR is not malignant.10 Similarly, we www.giejournal.org
Kuellmer et al
screened a total of 1234 patients undergoing EFTR with the FTRDdmost of them being non-lifting lesions. Of those, only 156 (12.6%) had malignant histology and met our inclusion criteria, and 87 (7%) were finally classified as high risk. Hence, the majority of patients in the total cohort had been successfully treated endoscopically without a requirement of surgical resection. Moreover, a recent study by Overwater et al24 showed that endoscopic resection of high-risk T1 colorectal carcinoma before surgical resection has no adverse effect on long-term oncologic outcomes. Therefore, EFTR may be considered primarily as a diagnostic tool for tissue acquisition to achieve correct risk stratification in patients with untreated lesions suspicious of T1 carcinoma. In 87 patients of our cohort, surgical resection was indicated because lesions were classified as high risk. However, only 53 patients actually underwent oncologic resection, mainly because they refused further treatment or were unfit for surgery because of comorbidities. These selected patients may benefit from a local endoscopic resection, because lymph-node involvement is present in only 1020% of high-risk cases.20 In a retrospective study of 6 patients undergoing EFTR for T1 high-risk lesions without further surgery, no patient had signs of local or distant recurrence after 12 months of follow-up.14 In our cohort, 24 of 33 patients (73%) with high-risk conditions and no surgical intervention underwent R0 resection. Of these, the available follow-up of 14 patients showed no signs of recurrence after 12 weeks. Identifying those patients with high-risk features that are curable by endoscopic resection would be tempting. There is growing evidence that number and function of tumor infiltrating lymphocytes influence the rate of lymphnode metastasis25 and prognosis26 in CRC. In our study we did not have access to tissue samples to address this issue. Further studies are needed to allow exact risk stratification of high-risk tumors in order to spare surgery for those selected patients. Our study has several limitations. First, it is a retrospective cohort study. We therefore cannot provide data on how many patients in total have been evaluated for EFTR, and treatment algorithms as well as endoscopic expertise may differ between participating centers. Second, the majority of data was derived from the online FTRD registry. Entering patients into the database was on a voluntary basis and no data monitoring was performed. Therefore, data were not always complete, and we cannot exclude inclusion bias. The registry data set contains only basic pretherapeutic data, and detailed information (eg, about endoscopic characteristics of lesions and standardized assessment of lesion size) is missing. Third, the follow-up period is relatively short, and we are not yet able to provide data on long-term outcome of patients. In conclusion, our study shows that EFTR for early CRC is feasible, effective, and safe. As a primarily diagnostic procedure for tissue acquisition, it allows exact histologic risk stratification in order to assign patients individually to the www.giejournal.org
Endoscopic full-thickness resection for early colorectal cancer
best treatment and avoid surgery for low-risk lesions. For patients with high-risk lesions unfit for surgery, it might also be a valuable option for local endoscopic treatment. However, prospective studies are required to further define indications for EFTR in malignant colorectal lesions and to evaluate long-term outcome of patients.
REFERENCES 1. Colorectal cancer: fact sheet 2018, The Global Cancer Observatory. Available at: http://gco.iarc.fr/today/data/factsheets/cancers/10_8_9Colorectum-fact-sheet.pdf. Accessed September 18, 2018. 2. ASGE Standards of Practice Committee; Fisher DA, Shergill AK, Early DS, et al. Role of endoscopy in the staging and management of colorectal cancer. Gastrointest Endosc 2013;78:8-12. Erratum in: Gastrointest Endosc 201378:559. 3. Ferlitsch M, Moss A, Hassan C, et al. Colorectal polypectomy and endoscopic mucosal resection (EMR): European Society of Gastrointestinal Endoscopy (ESGE) clinical guideline. Endoscopy 2017;49:270-97. 4. Tanaka S, Kashida H, Saito Y, et al. JGES guidelines for colorectal endoscopic submucosal dissection/endoscopic mucosal resection. Dig Endosc 2015;27:417-34. 5. Aepli P, Criblez D, Baumeler S, et al. Endoscopic full thickness resection (EFTR) of colorectal neoplasms with the Full Thickness Resection Device (FTRD): clinical experience from two tertiary referral centers in Switzerland. United European Gastroenterol J 2018;6:463-70. 6. Andrisani G, Pizzicannella M, Martino M, et al. Endoscopic full-thickness resection of superficial colorectal neoplasms using a new over-thescope clip system: a single-centre study. Dig Liver Dis 2017;49:1009-13. 7. Fähndrich M, Sandmann M. Endoscopic full-thickness resection for gastrointestinal lesions using the over-the-scope clip system: a case series. Endoscopy 2015;47:76-9. 8. Richter-Schrag HJ, Walker C, Thimme R, et al. Full thickness resection device (FTRD). Experience and outcome for benign neoplasms of the rectum and colon. [German]. Chirurg 2016;87:316-25. 9. Schmidt A, Bauerfeind P, Gubler C, et al. Endoscopic full-thickness resection in the colorectum with a novel over-the-scope device: first experience. Endoscopy 2015;47:719-25. 10. Schmidt A, Beyna T, Schumacher B, et al. Colonoscopic full-thickness resection using an over-the-scope device: a prospective multicentre study in various indications. Gut 2018;67:1280-9. 11. Valli PV, Mertens J, Bauerfeind P. Safe and successful resection of difficult GI lesions using a novel single-step full-thickness resection device (FTRD). Surg Endosc 2018;32:289-99. 12. Vitali F, Naegel A, Siebler J, et al. Endoscopic full-thickness resection with an over-the-scope clip device (FTRD) in the colorectum: results from a university tertiary referral center. Endosc Int Open 2018;6: E98-103. 13. Andrisani G, Pizzicannella M, Di Matteo FM. Endoscopic full-thickness resection of synchronous adenocarcinomas of the distal rectum. Case Rep Gastroenterol 2017;11:78-84. 14. Soriani P, Tontini GE, Neumann H, et al. Endoscopic full-thickness resection for T1 early rectal cancer: a case series and video report. Endosc Int Open 2017;5:E1081-6. 15. Wedi E, Orlandini B, Gromski M, et al. Full-thickness resection device for complex colorectal lesions in high-risk patients as a last-resort endoscopic treatment: initial clinical experience and review of the current literature. Clin Endosc 2018;51:103-8. 16. Asayama N, Oka S, Tanaka S, et al. Endoscopic submucosal dissection as total excisional biopsy for clinical T1 colorectal carcinoma. Digestion 2015;91:64-9. 17. Ito S, Hotta K, Imai K, et al. Treatment strategy for local recurrences after endoscopic resection of a colorectal neoplasm. Surg Endosc. Epub 2018 Jul 24.
Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 9
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
18. Hassan C, Repici A, Sharma P, et al. Efficacy and safety of endoscopic resection of large colorectal polyps: a systematic review and metaanalysis. Gut 2016;65:806-20. 19. Kuroki Y, Hoteya S, Mitani T, et al. Endoscopic submucosal dissection for residual/locally recurrent lesions after endoscopic therapy for colorectal tumors. J Gastroenterol Hepatol 2010;25:1747-53. 20. Kikuchi R, Takano M, Takagi K, et al. Management of early invasive colorectal cancer. Risk of recurrence and clinical guidelines. Dis Colon Rectum 1995;38:1286-95. 21. Yoda Y, Ikematsu H, Matsuda T, et al. A large-scale multicenter study of long-term outcomes after endoscopic resection for submucosal invasive colorectal cancer. Endoscopy 2013;45:718-24. 22. Ahlenstiel G, Hourigan LF, Brown G, et al; Australian Colonic Endoscopic Mucosal Resection (ACE) Study Group. Actual endoscopic versus predicted surgical mortality for treatment of advanced
10 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
23.
24.
25.
26.
mucosal neoplasia of the colon. Gastrointest Endosc 2014;80: 668-76. Ishiguro A, Uno Y, Ishiguro Y, et al. Correlation of lifting versus nonlifting and microscopic depth of invasion in early colorectal cancer. Gastrointest Endosc 1999;50:329-33. Overwater A, Kessels K, Elias SG, et al; Dutch T1 CRC Working Group. Endoscopic resection of high-risk T1 colorectal carcinoma prior to surgical resection has no adverse effect on long-term outcomes. Gut 2018;67:284-90. Däster S, Eppenberger-Castori S, Hirt C, et al. High frequency of CD8 positive lymphocyte infiltration correlates with lack of lymph node involvement in early rectal cancer. Dis Markers. Epub 2014 Dec 30. Ling A, Löfgren-Burström A, Larsson P, et al. TAP1 down-regulation elicits immune escape and poor prognosis in colorectal cancer. Oncoimmunology 2017;6:e1356143.
www.giejournal.org
Kuellmer et al
Endoscopic full-thickness resection for early colorectal cancer
APPENDIX 1. Full-thickness resection device study group* Klinikum Augsburg Klinikum Altenburger Land
Dr Andreas Probst
Klinikum Augsburg
Dr Michael Repp
Stiftungsklinik Weibenhorn
Dr Gregor Fitzel
Asklepiosklinik Langen
Dr Thomas Hansen Helios Klinikum Berlin Buch
Dr Christof Kurz
Städtisches Klinikum Braunschweig
Dr Lars Höpner
Ev. Krankenhaus Düsseldorf
APPENDIX 1. Continued
Dr Thorsten Beyna
Stadtkrankenhaus Schwabach
Dr Andreas Probst Dr Werner Schmidbaur Professor Dr Dominik Faust Dr Manfred Schmidt
Krankenhaus der barmherzigen Brüder Trier Uniklinik Tübingen
Dr Michael Knoll Professor Dr Martin Götz
Uniklinik Ulm
Professor Alexander Meining
Klinikum Darmstadt
Professor Dr Carl Schimanski
Donau Isar Klinikum Deggendorf
PD Dr Martin Caselitz
Petrus Krankenhaus Wuppertal
Professor Dr Andreas Erhardt
Dr Klaus Gutberlet
St. Bernhard Hospital Kamp-Lintfort
Dr Esmatollah Kasim
Josef- Hospital Delmenhorst Klinikum Dortmund St. Johannes Hospital Dortmund Malteser KH St. Anna Duisburg
Dr Martin Fähndrich Dr Kester Tüffers (inzwischen in anderem Haus) Dr Thomas Wiedbrauch Professor Dr Martin Wegener
Elisabeth KH Essen
Klinikum Weiden
Professor Dr Frank Kullmann Dr Sonja Pampuch
Nouvel Hôpital Civil Strasbourg
Dr Edris Wedi
*Special thanks to the additional authors who contributed cases for the study.
Dr David Albers
Universitätsklinikum Erlangen
Professor Dr Timo Rath
Universitätsklinikum Freiburg
Dr Henning Schwacha Professor Dr Richter Schrag Professor Dr Andreas Fischer
Klinikum GarmischPatenkirchen
Professor Dr Hans Dieter Allescher
Asklepios Klinik Lich
Dr Roland Fischer
AMEOS Klinik am Bürgerpark Bremerhaven
Dr Vera Knedeisen
St. Anna Hospital Herne Mathilden Hospital Herford
Dr Werner Hoffmann Dr Detlev Scholz Dr Adam Bielich
Klinikum Robert Koch Gehrden Hannover Universitätsmedizin Greifswald UKSH Campus Lübeck St. Vinzenz Hospital Köln Helios Klinikum Krefeld Städtisches Klinikum St. Georg Klinikum Ludwigsburg Luzerner Kantonspital
Professor Dr Jochen Wedemeyer Dr Anne Glitsch Professor Dr Klaus Fellermann (inzwischen in Freudenstadt) Dr Philipp Zervoulakos Dr Jürgen Heise Dr Ingo Wallstabe Professor Dr Karel Caca Dr Patrick Aepli
Universitätsklinikum Mannheim
Professor Dr Georg Kähler
Universitätsklinikum Marburg
Professor Dr Ulrike Denzer
Katholisches Klinikum Mainz
Dr Christopher Hofmann
Kliniken des Landkreises Neumarkt i.d. Opf.
Professor Claus Schäfer
www.giejournal.org
Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 10.e1
Endoscopic full-thickness resection for early colorectal cancer Armin Kuellmer, MD,1,* Julius Mueller,1,* Karel Caca, MD,2 Patrick Aepli, MD,3 David Albers, MD,4 Brigitte Schumacher, MD,4 Anne Glitsch, MD,5 Claus Schäfer,6 Ingo Wallstabe, MD,7 Christopher Hofmann, MD,8 Andreas Erhardt,9 Benjamin Meier, MD,2 Dominik Bettinger, MD,1 Robert Thimme, MD,1 Arthur Schmidt, MD1, the FTRD study group Freiburg, Germany
Background and Aims: Current international guidelines recommend endoscopic resection for T1 colorectal cancer (CRC) with low-risk histology features and oncologic resection for those at high risk of lymphatic metastasis. Exact risk stratification is therefore crucial to avoid under-treatment as well as over-treatment. Endoscopic full-thickness resection (EFTR) has shown to be effective for treatment of non-lifting benign lesions. In this multicenter, retrospective study we aimed to evaluate efficacy, safety, and clinical value of EFTR for early CRC. Methods: Records of 1234 patients undergoing EFTR for various indications at 96 centers were screened for eligibility. A total of 156 patients with histologic evidence of adenocarcinoma were identified. This cohort included 64 cases undergoing EFTR after incomplete resection of a malignant polyp (group 1) and 92 non-lifting lesions (group 2). Endpoints of the study were: technical success, R0-resection, adverse events, and successful discrimination of high-risk versus low-risk tumors. Results: Technical success was achieved in 144 out of 156 (92.3%). Mean procedural time was 42 minutes. R0 resection was achieved in 112 of 156 (71.8%). Subgroup analysis showed a R0 resection rate of 87.5% in Group 1 and 60.9% in Group 2 (p < .001). Severe procedure-related adverse events were recorded in 3.9% of patients. Discrimination between high-risk versus low-risk tumor was successful in 155 of 156 cases (99.3%). In Group 1, 84.1% were identified as low-risk lesions, whereas 16.3% in group 2 had low-risk features. In total, 53 patients (34%) underwent oncologic resection due to high-risk features whereas 98 patients (62%) were followed endoscopically. Conclusions: In early colorectal cancer, EFTR is technically feasible and safe. It allows exact histological risk stratification and can avoid surgery for low-risk lesions. Prospective studies are required to further define indications for EFTR in malignant colorectal lesions and to evaluate long-term outcome. (Gastrointest Endosc 2019;-:1-10.)
Abbreviations: CRC, colorectal cancer; EFTR, endoscopic full-thickness resection; ESD, endoscopic submucosal dissection; FTRD, full-thickness resection device; R0, microscopically complete resection; T1, Tumor stage according to TNM classification. DISCLOSURE: K. Caca received lecture fees and study grants from Ovesco Endoscopy. A. Schmidt received lecture fees from Ovesco Endoscopy. All other authors disclosed no financial relationships relevant to this publication. *Drs Kuellmer and Mueller contributed equally to this article. Copyright ª 2019 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 https://doi.org/10.1016/j.gie.2018.12.025 Received September 29, 2018. Accepted December 29, 2018.
www.giejournal.org
Current affiliations: Department of Medicine II, Medical CenterdUniversity of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg (1), Department of Gastroenterology and Oncology, Klinikum Ludwigsburg, Ludwigsburg, Germany (2), Gastroenterology and Hepatology Unit, Luzerner Kantonsspital, Lucerne, Switzerland (3), Department of Gastroenterology, Elisabeth-Krankenhaus Essen, Teaching Hospital of the University of Duisburg-Essen, Essen (4), Department of Surgery, University Hospital Greifswald, Greifswald, (5), Medical Clinic II, Klinikum Neumarkt, Neumarkt (6), Department for Gastroenterology, Hepatology, Diabetology und Endocrinology, Klinikum St. Georg gGmbH, Leipzig (7), Department for Gastroenterology, Katholisches Klinikum Mainz, Mainz (8), Department for Gastroenterology, Hepatology und Diabetology, Petrus-Krankenhaus, Wuppertal, Germany (9). Reprint requests: Arthur Schmidt, Department of Medicine II, Medical Center, Faculty of Medicine, University of Freiburg, Hugstetter Strasse 55, 79106 Freiburg, Germany.
Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 1
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
Colorectal cancer (CRC) is the third most commonly diagnosed cancer and the second leading cause of cancer death.1 In recent years, the incidence of CRC as well as cancer-associated deaths could be reduced significantly. This is mostly due to an improved screening strategy that allows detection and treatment in earlier stages. For early (T1) CRC with a low risk of lymphatic metastasis, current international guidelines recommend endoscopic treatment, whereas oncologic surgical resection is recommended for lesions with high-risk histology features.2-4 It is therefore crucial to correctly discriminate low-risk versus high-risk lesions to assign patients to the best treatment strategy. However, low-risk versus high-risk lesions are difficult to discriminate before resection because criteria are based on histology features. Moreover, T1 cancer is often found incidentally after resection of non-suspicious polyps, and further treatment after incomplete resection is not welldefined. Endoscopic full-thickness resection (EFTR) is an emerging technique for removal of non-lifting colorectal lesions but data on resection of malignant tumors are limited.5-12 Here we report on efficacy and safety of EFTR for staging and treatment of early CRC.
METHODS Study design and patients We conducted a retrospective observational study. Data of 1234 patients who had undergone EFTR with the fullthickness resection device (FTRD) system (Ovesco Endoscopy, Tübingen, Germany) at 96 centers were screened for eligibility. Inclusion criteria were: (1) repeat resection of incidentally found malignant polyps after incomplete endoscopic resection; (2) non-lifting lesions initially classified as benign with final histology of adenocarcinoma; and (3) non-lifting lesions with known malignant histology before EFTR. Exclusion criteria were (1) malignant histology other than adenocarcinoma (eg, neuroendocrine tumors); (2) no evidence of malignancy before or after EFTR; (3) lesions treated with combination of EFTR and other endoscopic resection techniques; and (4) incomplete data on histology results in the resected EFTR specimen. Study data were collected anonymously from the following sources (Fig. 1): the German Online FTRD Registry, which was initiated in September 2015 by Ovesco Endoscopy within a post-marketing clinical follow-up. It is accessible to all German endoscopists trained in full-thickness resection with the FTRD. Until July 2018, it was made up of data from 950 FTRD cases from 83 centers. We further screened 181 patients who had undergone EFTR within the WALL-RESECT study (NCT 02362126).10 Data were obtained from the coordinating study center in Ludwigsburg, Germany. To avoid the risk of doublecounting, all centers participating in the WALL-RESECT 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
study were thoroughly instructed not to enter data regarding these patients into the registry. During data collection, we confirmed this policy again with the responsible physicians at the participating centers. An additional 30 patients (103 screened) undergoing EFTR for CRC were identified from 7 endoscopic centers in Germany and Switzerland that were neither documented in the FTRD registry nor in the WALL-RESECT study. Centers with special expertise in full-thickness resection were contacted and asked for cases meeting the inclusion criteria. Written informed consent on the EFTR procedure was obtained in all cases. Informed consent was obtained from participants of the WALL-RESECT study. For all other patients, including those from the German Registry, no specific written consent had been obtained. Data acquisition and analysis for our study was approved by our institutional review board with the requirement of using only anonymized data.
Endpoints and definitions The primary endpoints were (1) technical success, defined as reaching the lesion with the FTRD, successful clip application, and macroscopically complete resection of the lesion with the integrated snare, and (2) R0 resection: complete histologic resection, defined as tumor-free lateral and deep resection margins. Cases with no evidence of neoplastic tissue in the EFTR specimen (after incomplete removal with standard methods) were also considered as R0. Secondary endpoints were (1) procedure-related adverse events. Mild adverse events were defined as those not requiring medical or repeated endoscopic intervention and not prolonging hospital admission. Moderate adverse events were defined as those requiring medical or repeated endoscopic intervention and/or prolonging hospital admission. Severe adverse events were defined as those requiring surgical therapy and/or potentially life threatening. Histology findings of EFTR specimens confirm the following: (1) histologically confirm full-thickness resection, (2) successful discrimination of low-risk versus high-risk tumors, (3) evidence of residual or recurrent tumors in endoscopic follow-up, (4) requirement for oncologic surgical resection, (5) evidence of residual tumor and lymph node involvement in the surgical resection specimen, T-stage and N-stage. According to international guidelines,2-4 patients were classified as low risk (in terms of risk for lymph node metastasis) if all of the following histology features were present: submucosal infiltration <1000 mm, no lymphovascular infiltration, no infiltration of blood vessels, well or moderate differentiation/grading, and microscopically complete resection (R0). Patients who did not meet all of these criteria were classified as high risk. www.giejournal.org
Kuellmer et al
Endoscopic full-thickness resection for early colorectal cancer
Patients screened n = 1234 Patients excluded, n = 1078 No adenocarcinoma, n = 770 NET/ NEC, n = 59 Hybrid technique, n = 6 Location other than colon/ rectum, n = 6 Incomplete information, n = 237
Patients enrolled n = 156
Indication
Re- resections (Group 1)
Non- lifting lesions (Group 2)
n = 64
“Low risk,” n = 53
n = 92
“High-risk,” n = 10
“High-risk,” n = 77
“Low- risk,” n = 15
-R0. but sm >1000, n = 4 -Rx and “high-risk,” n = 2 -Rx, otherwise low-risk n=1 -Rx, no further information, n=3
- R0, but sm >1000, n = 42 -Rx and “high- risk,” n = 25 -Rx, otherwise “low-risk,” n=5 - Rx, no further information, n=5
Surgery* n=9
3 month – F/U n = 37
Recurrent lesions n=2
Surgery* n = 44
3 month – F/U n=0
Recurrent/ residual lesions
3 month – F/U* * n = 19
3 month – F/U n = 13
Recurrent lesions n=1
Residual lesions n=1
n=0
End. therapy, n = 1 radiochemoth., n = 1
End. therapy n=1
Lost in F/U n=1
Figure 1. Patient enrollment, distribution in groups 1 and 2, and outcome are shown. *Surgery due to oncologic re-resection. ySeven patients refusing surgery. Three patients not further treated due to comorbidities. Four patients received endoscopic treatment without information about reasons. Five patients with no further information. Three patients were not treated further because of comorbidities. Four patients received endoscopic treatment, without information about reasons. In 5 patients, no further information was available. NET, Neuroendocrine tumor; NEC, neuroendocrine carcinoma; R0, microscopically complete resection; F/U, Endoscopic follow-up (Patients not undergoing surgery); Rx, microscopically incomplete resection; End, endoscopic; (Patients, not undergoing surgery).
www.giejournal.org
Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 3
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
Figure 2. Key steps of the endoscopic full-thickness resection procedure are shown. A, B, C, Repeat resection after an incomplete standard resection of a T1 carcinoma (group 1). A, The lesion is marked. B, The lesion is pulled into the over the scope–loaded cap. C, The lesion is resected after clip application. Histology revealed only scar tissue. The patient was classified as low risk. D, A non-lifting lesion in the cecum without prior therapy (group 2). E, The lesion was resected. F, Image of the full-thickness resection specimen. Histology showed a T1 tumor with no lymphovascular infiltration but deep submucosal infiltration and venous infiltration. Although a microscopically complete resection was achieved, the patient was classified as high risk and referred to surgery.
FTRD and the EFTR procedure
Data management and statistical analysis
For resection of lesions, the FTRD (Ovesco Endoscopy) was used in all cases. For description of the device, technique, and peri-interventional management, we referred to previous studies9-10 and Figure 2.
The study data were collected and analyzed at the University of Freiburg and were used exactly as presented by the contributing investigators who were responsible for the correctness of the data at their investigation sites. Continuous
4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
www.giejournal.org
Kuellmer et al
Endoscopic full-thickness resection for early colorectal cancer
TABLE 1. Baseline characteristics
Variable Age, mean ( SD), y
TABLE 2. Procedure data and adverse events Entire cohort (n [ 156) 72 9.6
Sex, no. (%) Female
55 (35.3%)
Male
101 (64.7%)
Indication for EFTR, no. (%) Repeat resection of malignant polyp (group 1) Non-lifting lesion (group 2)
64 (41%) 92 (59%) 19 (20.7%)
Initial classified as (benign) adenoma
73 (79.3%)
Location, no. (%) Cecum
11 (7.1%)
Ascending colon
17 (10.9%)
Transverse colon
18 (11.5%)
Descending colon
12 (7.69%)
Sigmoid colon
48 (30.8%)
Rectum
50 (32.1%) 20 (2-45)
SD, Standard deviation; EFTR, endoscopic full-thickness resection.
variables were reported as mean with standard deviation, whereas categorical variables were expressed as frequencies and percentages (in parentheses) unless stated otherwise. For categorical variables, differences were determined by using c2 tests or the Fisher exact test, as appropriate. P values < .05 were considered significant. No interim analysis was performed. Data collection was performed with Microsoft Excel 2016 for Mac Os (version 15.21; Microsoft, Redmond, Wash). Statistical analyses were performed with SPSS (version 24.0; IBM, New York, NY) and GraphPad Prism (version 6; GraphPad Software, San Diego, Calif).
RESULTS Patients A total of 1234 patients at 96 centers were screened for eligibility, and 156 patients were included in the study. Mean age was 72 ( 9.6) years. Indications for the FTRD intervention were repeat resection of malignant polyps (group 1, 41%) and non-lifting lesions (group 2, 59%). Group 2 made up 21% (n Z 19) of adenocarcinomas on which biopsies were performed; the rest (n Z 73, 79%) were initially classified as benign. Most lesions (62.9%) were located in the rectosigmoid colon. Median lesion size was 20 mm (range 2-45 mm). Patient and lesion characteristics are shown in Table 1.
Procedural data (entire cohort) Technical success was achieved in 144 of 156 (92.3%) patients (Table 2). Reasons for technical failure (n Z 12) www.giejournal.org
Procedure time, mean ( SD) min*
42.1 26.6
Technical success,y no. (%)
144 (92.3%)
Adverse events, total
21 (14.1%)
Mild adverse eventsz Moderate adverse eventsx
Known carcinoma
Lesion size, median (range), mm
Entire cohort (n [ 156)
Variable
7 (4.5%) 8 (5.1%)
Bleeding with successful endoscopic treatment
6/8 (75%)
Infection with a need for antibiotics
2/8 (25%)
Severe adverse eventsk
6 (3.9%)
Perforation (1x immediate, 5x secondary)
6/6 (100%)
Immediate with successful endoscopic closure
1/6 (16.7%)
Secondary with a need for surgery
5/6 (83.3%)*
SD, Standard deviation. *Total procedure time including colonoscopy. yLesion successfully reached with FTRD, successful clip application and macroscopically complete resection. zManaged conservatively without intervention. xNeed for endoscopic hemostasis or antibiotics. kNeed for surgery and/or potentially life-threatening condition.
were dysfunction of the snare (10/12) and incorrect application of the FTRD clip (2/12). Mean total procedure time (including colonoscopy with marking of the lesion) was 42 minutes (standard deviation [SD] 26.6). In 33 cases the procedure time was not reported.
Histology The entire cohort. In the entire cohort of patients (n Z 156), R0 resection was achieved in 112 cases (71.8%). Full-thickness resection was histologically confirmed in 135 cases (86.5%). Definitive discrimination between high-risk and low-risk lesions by histology examination of the EFTR specimen according to the definition given previously was possible in all cases (99.4%) except one. A total of 68 cases (43.9%) were histologically classified as low risk, and 87 cases (56.1%) were classified as high risk. In 8 cases of high-risk tumors, not all histology criteria were available for analysis. Most high-risk lesions (43/79, 54%) were resected completely, but the submucosal infiltration depth was >1000 mm. In 29 of 79 cases (34%), resection was incomplete, and histology showed additional high-risk criteria. In 3 of 79 cases (4%), deep submucosal and lymphovascular infiltration appeared synchronously. Six of 79 cases (8%) had low-risk features but were not R0 resected. Detailed results of the entire cohort and subgroup analysis are shown in Tables 3 and 4. Group 1 (repeat resection of malignant polyp, n [ 64). In this subgroup, the R0 resection rate was achieved in 56 of 64 cases (87.5%). Histology findings were scar tissue in 49 of 64 (77%), residual adenocarcinoma in 14 of 64 (22%), and adenoma with low-grade interaepithelial dysplasia (LGIEN) in 1 of 64 (2%). According to the definition given previously, 53 patients Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 5
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
TABLE 3. Histology features of EFTR specimens
TABLE 4. Histologic features of high-risk tumors
Entire cohort (n [ 156)
Group 1 (n [ 64)
Group 2 (n [ 92)
Histology
Full-thickness resection, no. (%)
135 (86.5%)
54 (84.4%)
81 (88%)
Microscopically complete resection, no. (%)
112 (71.8%)
56 (87.5%)
56 (60.9%)
Variable
Histology Scar, no. (%)
Entire cohort (n [ 79)
Group 1 (n [ 7)
Group 2 (n [72)
R0 but sm >1000 mm
46 (58%)
4 (57%)
42 (58%)
Rx and other high-risk features
27 (34%)
2 (29%)
25 (35%)
Rx but otherwise lowrisk features
6 (8%)
1 (15%)
5 (7%)
R0, Microscopically complete resection; Rx, microscopically incomplete resection.
49 (31.4%)
49 (76.6%)
Adenoma, no. (%)
1 (0.6%)
1 (1.6%)
Carcinoma, no. (%)
106 (68%)
14 (21.9%)
92 (100%)
Low risk, no. (%)
68/155 (43.9%)
53/63 (84.1%)
15/92 (16.3%)
High risk, no. (%)
87/155 (56.1%)
10/63 (15.9%)
77/92 (83.7%)
EFTR, Endoscopic full-thickness resection.
(84.1%) were classified as low risk, whereas 10 patients (16%) had high-risk features. In 7 of these, all histology criteria were available for further workup: 4 of 7 (57%) had been resected completely (R0) but showed deep submucosal infiltration. In 2 of 7 (29%), resection was incomplete, and other high-risk criteria were present. In 1 of 7 (15%), resection was incomplete, but other histology features were low risk (Table 4). Group 2 (non-lifting lesions, n [ 92). The R0 resection rate for group 2 was 60.9% (56/92). Adenocarcinoma was present in all specimens (100%). Fifteen patients (16.3%) were classified as low risk, and 77 (83.7%) had high-risk features. In the high-risk group, T1 carcinoma was found in 55 of 77 cases (71.4%), whereas T2 or T3 carcinoma was diagnosed in 22 patients (28.6%). Seventy-two of 77 cases were available for further work-up because all histology features were present. In this high-risk group, 58% (42/72) had been resected completely but showed deep submucosal infiltration. Thirtyfive percent (25/72) could not be resected completely and met other high-risk criteria, and 7% (5/72) showed low-risk criteria but histological resection was incomplete (Rx/R1). Risk factors for incomplete resection (entire cohort). The R0 resection rate was significantly higher in group 1 than in group 2 (87.5% vs 60.9%; P < .001). We did not find any significant differences for smaller versus larger lesions. Moreover, there was no significant correlation between depth of tumor infiltration and the R0 resection rate (P 0.471). There was a trend toward higher rates of R0 resection in lesions in the rectum versus other lesions (P Z .058). The results are shown in Figure 3.
Adverse events In total, 21 procedure-related adverse events occurred (14.1%) (Table 2). Seven of these could be managed conservatively without intervention and were classified as 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
mild adverse events (4.49%). Moderate adverse events such as bleeding with the need for endoscopic hemostasis (n Z 6) or local peritonitis treated with antibiotics (n Z 2) occurred in 8 patients (5.1%). Severe adverse events (need for surgery and/or potentially life-threatening condition) were recorded for 6 patients (3.9%). In total, 6 perforations occurred. One perforation was diagnosed immediately during EFTR and was related to incorrect application of the clip. This perforation was closed successfully endoscopically with an overthe-scope clip. Five secondary perforations were observed. Median time to diagnosis was 7 days (range 1-8 days). All of those cases required surgical intervention.
Clinical outcome (entire cohort) Data on outcome of patients undergoing surgery or endoscopic follow-up are shown in Table 5. A total of 53 of 156 patients (34%) underwent oncologic resection because of high-risk histology features. Of those, 45 were available for detailed analysis. Five patients (11%) required surgery because of histologically incomplete resection (Rx/ R1) but had otherwise low-risk features. Twenty-two patients (49%) had incomplete EFTR and at least 1 additional high-risk criteria. In 18 cases (40%), the lesion was resected completely with the FTRD system but showed other highrisk features in histology. In 44 of the 53 cases undergoing oncologic resection, information about the histology of the surgical resection specimen was available. Residual tumor was found in 27 cases (61%), whereas in 17 cases (39%) no residual carcinoma was found. Ninety-eight of 156 patients (62%) did not undergo surgical resection. Of those, 65 patients (66%) showed lowrisk features and did not require further therapy. A total of 33 patients (34%) had high-risk criteria, but surgery was not performed because of the following reasons: refusal by the patient (n Z 18), comorbidity (n Z 5), and death caused by coexisting disease (n Z 1). In 9 cases, information about reasons for conservative treatment were not available. Of 98 patients not undergoing surgery, endoscopic follow-up was available in 69 (70.4%). Mean ( SD) follow-up was 15 weeks ( 11). The FTRD clip had spontaneously dislodged in 45 patients (69%). In 3 patients (4%) a www.giejournal.org
Kuellmer et al
Endoscopic full-thickness resection for early colorectal cancer
P = .537 P < .001
A
80 60 40 20
60 40 20
re-resection of malignant polyp
non-lifting lesions
B
P = .471
60 40 20 0 T1, sm<1000
10-20 mm
≤ 9 mm
>T1, sm<1000
D
> 20 mm
P = .058
100
R0 resection [%]
R0 resection [%]
P = .708
0
0
80
C
P = .754
80
R0 resection [%]
R0 resection [%]
100
100
80 60 40 20 0 rectum
non-rectum
Figure 3. Microscopically complete resection (R0) rate in various conditions is shown. A, Group 1 versus group 2. B, R0 regarding lesion size. C, R0 in deepness of submucosal infiltration. D, R0 in a lesion location in the rectum versus the rest. P < .05 was considered significant. R0, Microscopically complete resection; T1, according to TNM classfication; sm, submucosal infiltration depth measured in micrometers.
recurrent lesion after previous R0 resection was diagnosed. In 2 of those, a second FTRD procedure was performed, and histology showed adenoma without evidence of malignancy. The third patient developed recurrent carcinoma with distant metastasis and underwent chemoradiation. In 1 case (1.4%), a residual tumor was found, but the patient refused further therapy.
TABLE 5. Outcome of the patients undergoing oncologic resection and endoscopic follow-up Entire cohort (n [ 156)
Outcome Oncologic resection, no. (%)
53 (34%)
Residual tumor found
17/44 (38.6%)
Thereof N1
5 (18.5%)
No residual tumor found
DISCUSSION We reported on EFTR with an over-the-scope device for incompletely resected malignant polyps and non-lifting malignant colorectal lesions without prior attempt at resection. Our results demonstrate high technical efficacy and safety of this novel endoscopic technique. The study underlines the potential of EFTR to exactly discriminate between high-risk versus low-risk tumors to aid decision for the optimal individual treatment strategy. To our best knowledge, this is currently the largest study on EFTR for early colorectal cancer. Our cohort included 156 patients with malignant colorectal tumors undergoing EFTR. Technical success was 92.3%. Comparing these results with the literature is difficult because studies on EFTR for colorectal carcinoma are very rare. Besides some case reports, there is only 1 www.giejournal.org
Endoscopic follow-up obtained,* no. (%) Median time (min, range) Clip position (in situ vs departed) Residual/recurrent lesion
27/44 (61.4%) 69/98 (70.4%) 12 (1-64) 20 (30.8%) vs 45 (69.2%) 4 (5.8%)
N1, Histologic evidence of lymphnode involvement in surgical resection specimen. *Referring to patients who did not undergo surgical resection.
case series of 6 patients undergoing EFTR for early rectal cancer.13-15 Other cases are reported in larger studies investigating EFTR for various indications.5-12 In these studies, data on success and efficacy usually were calculated for the whole cohort and not specifically for patients with malignant lesions. Reported technical success rates of EFTR for other indications (mainly non-lifting adenomas) range from 75% to 100%5-12 and are in line with our Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 7
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
data. The mean ( SD) procedure time of 42 minutes ( 26.6 minutes) in our study was also in the range of procedure times in other studies.10,12 Our data therefore suggest that the presence of malignancy does not negatively influence the feasibility of EFTR with the FTRD. The total R0 resection rate in our study was 71.8%. This is below the rate reported in the WALL-RESECT study for the total cohort (76.9%), but subgroup analysis of adenocarcinomas in the WALL-RESECT study showed a rate similar to ours (72%).10 Notably, in 29% (9/31) of patients with histology classified as incomplete, no residual tumor was found in the surgical specimen. Hence, in those cases the tumor margins were positive even though complete resection had been achieved. If these specimens were reclassified as R0, as done in another study,11 our overall R0-rate would rise to 78% (89% in group 1, 70% in group 2). In contrast to the WALL-RESECT study, we did not find a significant correlation of lesion size with R0-resection rate. One reason for this difference may be that lesion size has not been determined in a standardized way in our patient cohort. Moreover, the WALL-RESECT study had a different patient population including various indications for EFTR with only a minority being adenocarcinomas. We did observe a significant difference of the R0 resection rate between group 1 (repeat resection of malignant polyps, 87.5%) and group 2 (non-lifting lesions, 60.9%). This may be because group 1 mainly consisted of patients with low-risk tumors and small resection scars, whereas the majority of patients in group 2 proved to have advanced tumors. Data on R0 resection rates of standard endoscopic resection methods for T1 tumors are scarce. Asamaya et al16 showed a 92% R0 resection rate in a series of 37 patients. Ito et al17 analyzed local recurrences after endoscopic resection. R0 resection rates for EMR and endoscopic submucosal dissection (ESD) were as low as 28.6% and 67%, respectively. In a meta-analysis comparing EMR and ESD for resection of colon neoplasms, the R0 resection rate for EMR was 42.3% compared with 80.3% for ESD. This rate accounted for all lesions (adenoma/ adenocarcinoma) but was not further specified for adenocarcinoma alone. The analysis also reported that noncurative resection (8% pooled rate) was mainly (60%) due to invasive cancer.18 Whether cancer histology showed highrisk features within a complete resection or if the resection itself was incomplete was not specified.18 Hence, the R0 resection rate in our study is higher than reported rates of EMR and in the range of rates for ESD. We observed procedure-related adverse events in 14% of patientsd3.8% were classified as major adverse events, and surgical therapy due to perforation was required in 3% of patients. This indicates that colorectal EFTR for malignant indications is safe. The rate of major adverse events, including perforation, is in line with reported data on EFTR for other indications (0%-5%).6-12 In contrast, perforation rates for ESD in non-lifting lesions are reported to 8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
be as high as 15%.19 Perforation rates for EMR are significantly lower (2.5%), but the technique is generally not recommended for malignant and/or non-lifting lesions. Our data indicate that EFTR for colorectal cancer is feasible and safe. But what is its clinical significance? Current international guidelines recommend endoscopic resection of colorectal malignant tumors that are considered low risk.2-4 For those lesions, local treatment is sufficient because the risk of lymph node metastasis or local recurrence is very low.20 If complete resection of these lesions is achieved, the 5-year recurrence-free-survival rate is reported to be 98%.21 On the other hand, oncologic resection with lymph node dissection is the treatment of choice for early colorectal cancer with histology proven high-risk features. Surgery is effective, but more invasive, and it is associated with increased morbidity and mortality compared with endoscopic resection.22 Therefore, correct discrimination between low-risk and high-risk tumors is crucial to assigning patients to the optimal treatment strategy. In our study cohort, risk status was initially unclear in all patients and could be clarified in 99.4% of patients by means of EFTR. This underlines that reliable pretherapeutic discrimination between low-risk and high-risk T1-CRC usually is not possible. This classification is based on histologic criteria such as submucosal infiltration depth or infiltration of lymphatic vessels, which requires a thorough microscopic examination of the whole lesion and cannot be determined reliably by endoscopy with or without a biopsy. Our study cohort included 64 patients who had undergone incomplete resection of incidentally found malignant polyps (group 1). Those patients are often referred to surgery because local endoscopic repeat treatment usually is difficult because of scarring.23 EFTR with the FTRD allows resection of non-lifting, pretreated lesions as shown in multiple retrospective studies and 1 large, prospective study.10 In our study, 84.1% of patients were finally classified as low risk and were identified correctly not to be candidates for oncologic surgical resection. In other words, EFTR obviated the need for surgery in the majority of these patients and may therefore be the method of choice for this indication. The second subgroup (group 2) in our study included 92 patients with non-lifting lesions. The majority of the lesions had primarily been classified as non-malignant by the endoscopist, based on macroscopic criteria, and biopsies were not performed, whereas 19 (21%) patients had biopsy-proven carcinomas. The vast majority (83%) of lesions was finally classified as high risk, and surgical resection was recommended. However, a primary EFTR approach, in our opinion, is still justified for the following reasons: the 92 patients in this group represented only a small subgroup of all patients undergoing EFTR for non-lifting lesions. As previously shown in the WALLRESECT study, the majority of non-lifting lesions with indication for EFTR is not malignant.10 Similarly, we www.giejournal.org
Kuellmer et al
screened a total of 1234 patients undergoing EFTR with the FTRDdmost of them being non-lifting lesions. Of those, only 156 (12.6%) had malignant histology and met our inclusion criteria, and 87 (7%) were finally classified as high risk. Hence, the majority of patients in the total cohort had been successfully treated endoscopically without a requirement of surgical resection. Moreover, a recent study by Overwater et al24 showed that endoscopic resection of high-risk T1 colorectal carcinoma before surgical resection has no adverse effect on long-term oncologic outcomes. Therefore, EFTR may be considered primarily as a diagnostic tool for tissue acquisition to achieve correct risk stratification in patients with untreated lesions suspicious of T1 carcinoma. In 87 patients of our cohort, surgical resection was indicated because lesions were classified as high risk. However, only 53 patients actually underwent oncologic resection, mainly because they refused further treatment or were unfit for surgery because of comorbidities. These selected patients may benefit from a local endoscopic resection, because lymph-node involvement is present in only 1020% of high-risk cases.20 In a retrospective study of 6 patients undergoing EFTR for T1 high-risk lesions without further surgery, no patient had signs of local or distant recurrence after 12 months of follow-up.14 In our cohort, 24 of 33 patients (73%) with high-risk conditions and no surgical intervention underwent R0 resection. Of these, the available follow-up of 14 patients showed no signs of recurrence after 12 weeks. Identifying those patients with high-risk features that are curable by endoscopic resection would be tempting. There is growing evidence that number and function of tumor infiltrating lymphocytes influence the rate of lymphnode metastasis25 and prognosis26 in CRC. In our study we did not have access to tissue samples to address this issue. Further studies are needed to allow exact risk stratification of high-risk tumors in order to spare surgery for those selected patients. Our study has several limitations. First, it is a retrospective cohort study. We therefore cannot provide data on how many patients in total have been evaluated for EFTR, and treatment algorithms as well as endoscopic expertise may differ between participating centers. Second, the majority of data was derived from the online FTRD registry. Entering patients into the database was on a voluntary basis and no data monitoring was performed. Therefore, data were not always complete, and we cannot exclude inclusion bias. The registry data set contains only basic pretherapeutic data, and detailed information (eg, about endoscopic characteristics of lesions and standardized assessment of lesion size) is missing. Third, the follow-up period is relatively short, and we are not yet able to provide data on long-term outcome of patients. In conclusion, our study shows that EFTR for early CRC is feasible, effective, and safe. As a primarily diagnostic procedure for tissue acquisition, it allows exact histologic risk stratification in order to assign patients individually to the www.giejournal.org
Endoscopic full-thickness resection for early colorectal cancer
best treatment and avoid surgery for low-risk lesions. For patients with high-risk lesions unfit for surgery, it might also be a valuable option for local endoscopic treatment. However, prospective studies are required to further define indications for EFTR in malignant colorectal lesions and to evaluate long-term outcome of patients.
REFERENCES 1. Colorectal cancer: fact sheet 2018, The Global Cancer Observatory. Available at: http://gco.iarc.fr/today/data/factsheets/cancers/10_8_9Colorectum-fact-sheet.pdf. Accessed September 18, 2018. 2. ASGE Standards of Practice Committee; Fisher DA, Shergill AK, Early DS, et al. Role of endoscopy in the staging and management of colorectal cancer. Gastrointest Endosc 2013;78:8-12. Erratum in: Gastrointest Endosc 201378:559. 3. Ferlitsch M, Moss A, Hassan C, et al. Colorectal polypectomy and endoscopic mucosal resection (EMR): European Society of Gastrointestinal Endoscopy (ESGE) clinical guideline. Endoscopy 2017;49:270-97. 4. Tanaka S, Kashida H, Saito Y, et al. JGES guidelines for colorectal endoscopic submucosal dissection/endoscopic mucosal resection. Dig Endosc 2015;27:417-34. 5. Aepli P, Criblez D, Baumeler S, et al. Endoscopic full thickness resection (EFTR) of colorectal neoplasms with the Full Thickness Resection Device (FTRD): clinical experience from two tertiary referral centers in Switzerland. United European Gastroenterol J 2018;6:463-70. 6. Andrisani G, Pizzicannella M, Martino M, et al. Endoscopic full-thickness resection of superficial colorectal neoplasms using a new over-thescope clip system: a single-centre study. Dig Liver Dis 2017;49:1009-13. 7. Fähndrich M, Sandmann M. Endoscopic full-thickness resection for gastrointestinal lesions using the over-the-scope clip system: a case series. Endoscopy 2015;47:76-9. 8. Richter-Schrag HJ, Walker C, Thimme R, et al. Full thickness resection device (FTRD). Experience and outcome for benign neoplasms of the rectum and colon. [German]. Chirurg 2016;87:316-25. 9. Schmidt A, Bauerfeind P, Gubler C, et al. Endoscopic full-thickness resection in the colorectum with a novel over-the-scope device: first experience. Endoscopy 2015;47:719-25. 10. Schmidt A, Beyna T, Schumacher B, et al. Colonoscopic full-thickness resection using an over-the-scope device: a prospective multicentre study in various indications. Gut 2018;67:1280-9. 11. Valli PV, Mertens J, Bauerfeind P. Safe and successful resection of difficult GI lesions using a novel single-step full-thickness resection device (FTRD). Surg Endosc 2018;32:289-99. 12. Vitali F, Naegel A, Siebler J, et al. Endoscopic full-thickness resection with an over-the-scope clip device (FTRD) in the colorectum: results from a university tertiary referral center. Endosc Int Open 2018;6: E98-103. 13. Andrisani G, Pizzicannella M, Di Matteo FM. Endoscopic full-thickness resection of synchronous adenocarcinomas of the distal rectum. Case Rep Gastroenterol 2017;11:78-84. 14. Soriani P, Tontini GE, Neumann H, et al. Endoscopic full-thickness resection for T1 early rectal cancer: a case series and video report. Endosc Int Open 2017;5:E1081-6. 15. Wedi E, Orlandini B, Gromski M, et al. Full-thickness resection device for complex colorectal lesions in high-risk patients as a last-resort endoscopic treatment: initial clinical experience and review of the current literature. Clin Endosc 2018;51:103-8. 16. Asayama N, Oka S, Tanaka S, et al. Endoscopic submucosal dissection as total excisional biopsy for clinical T1 colorectal carcinoma. Digestion 2015;91:64-9. 17. Ito S, Hotta K, Imai K, et al. Treatment strategy for local recurrences after endoscopic resection of a colorectal neoplasm. Surg Endosc. Epub 2018 Jul 24.
Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 9
Endoscopic full-thickness resection for early colorectal cancer
Kuellmer et al
18. Hassan C, Repici A, Sharma P, et al. Efficacy and safety of endoscopic resection of large colorectal polyps: a systematic review and metaanalysis. Gut 2016;65:806-20. 19. Kuroki Y, Hoteya S, Mitani T, et al. Endoscopic submucosal dissection for residual/locally recurrent lesions after endoscopic therapy for colorectal tumors. J Gastroenterol Hepatol 2010;25:1747-53. 20. Kikuchi R, Takano M, Takagi K, et al. Management of early invasive colorectal cancer. Risk of recurrence and clinical guidelines. Dis Colon Rectum 1995;38:1286-95. 21. Yoda Y, Ikematsu H, Matsuda T, et al. A large-scale multicenter study of long-term outcomes after endoscopic resection for submucosal invasive colorectal cancer. Endoscopy 2013;45:718-24. 22. Ahlenstiel G, Hourigan LF, Brown G, et al; Australian Colonic Endoscopic Mucosal Resection (ACE) Study Group. Actual endoscopic versus predicted surgical mortality for treatment of advanced
10 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2019
23.
24.
25.
26.
mucosal neoplasia of the colon. Gastrointest Endosc 2014;80: 668-76. Ishiguro A, Uno Y, Ishiguro Y, et al. Correlation of lifting versus nonlifting and microscopic depth of invasion in early colorectal cancer. Gastrointest Endosc 1999;50:329-33. Overwater A, Kessels K, Elias SG, et al; Dutch T1 CRC Working Group. Endoscopic resection of high-risk T1 colorectal carcinoma prior to surgical resection has no adverse effect on long-term outcomes. Gut 2018;67:284-90. Däster S, Eppenberger-Castori S, Hirt C, et al. High frequency of CD8 positive lymphocyte infiltration correlates with lack of lymph node involvement in early rectal cancer. Dis Markers. Epub 2014 Dec 30. Ling A, Löfgren-Burström A, Larsson P, et al. TAP1 down-regulation elicits immune escape and poor prognosis in colorectal cancer. Oncoimmunology 2017;6:e1356143.
www.giejournal.org
Kuellmer et al
Endoscopic full-thickness resection for early colorectal cancer
APPENDIX 1. Full-thickness resection device study group* Klinikum Augsburg Klinikum Altenburger Land
Dr Andreas Probst
Klinikum Augsburg
Dr Michael Repp
Stiftungsklinik Weibenhorn
Dr Gregor Fitzel
Asklepiosklinik Langen
Dr Thomas Hansen Helios Klinikum Berlin Buch
Dr Christof Kurz
Städtisches Klinikum Braunschweig
Dr Lars Höpner
Ev. Krankenhaus Düsseldorf
APPENDIX 1. Continued
Dr Thorsten Beyna
Stadtkrankenhaus Schwabach
Dr Andreas Probst Dr Werner Schmidbaur Professor Dr Dominik Faust Dr Manfred Schmidt
Krankenhaus der barmherzigen Brüder Trier Uniklinik Tübingen
Dr Michael Knoll Professor Dr Martin Götz
Uniklinik Ulm
Professor Alexander Meining
Klinikum Darmstadt
Professor Dr Carl Schimanski
Donau Isar Klinikum Deggendorf
PD Dr Martin Caselitz
Petrus Krankenhaus Wuppertal
Professor Dr Andreas Erhardt
Dr Klaus Gutberlet
St. Bernhard Hospital Kamp-Lintfort
Dr Esmatollah Kasim
Josef- Hospital Delmenhorst Klinikum Dortmund St. Johannes Hospital Dortmund Malteser KH St. Anna Duisburg
Dr Martin Fähndrich Dr Kester Tüffers (inzwischen in anderem Haus) Dr Thomas Wiedbrauch Professor Dr Martin Wegener
Elisabeth KH Essen
Klinikum Weiden
Professor Dr Frank Kullmann Dr Sonja Pampuch
Nouvel Hôpital Civil Strasbourg
Dr Edris Wedi
*Special thanks to the additional authors who contributed cases for the study.
Dr David Albers
Universitätsklinikum Erlangen
Professor Dr Timo Rath
Universitätsklinikum Freiburg
Dr Henning Schwacha Professor Dr Richter Schrag Professor Dr Andreas Fischer
Klinikum GarmischPatenkirchen
Professor Dr Hans Dieter Allescher
Asklepios Klinik Lich
Dr Roland Fischer
AMEOS Klinik am Bürgerpark Bremerhaven
Dr Vera Knedeisen
St. Anna Hospital Herne Mathilden Hospital Herford
Dr Werner Hoffmann Dr Detlev Scholz Dr Adam Bielich
Klinikum Robert Koch Gehrden Hannover Universitätsmedizin Greifswald UKSH Campus Lübeck St. Vinzenz Hospital Köln Helios Klinikum Krefeld Städtisches Klinikum St. Georg Klinikum Ludwigsburg Luzerner Kantonspital
Professor Dr Jochen Wedemeyer Dr Anne Glitsch Professor Dr Klaus Fellermann (inzwischen in Freudenstadt) Dr Philipp Zervoulakos Dr Jürgen Heise Dr Ingo Wallstabe Professor Dr Karel Caca Dr Patrick Aepli
Universitätsklinikum Mannheim
Professor Dr Georg Kähler
Universitätsklinikum Marburg
Professor Dr Ulrike Denzer
Katholisches Klinikum Mainz
Dr Christopher Hofmann
Kliniken des Landkreises Neumarkt i.d. Opf.
Professor Claus Schäfer
www.giejournal.org
Volume
-,
No.
-
: 2019 GASTROINTESTINAL ENDOSCOPY 10.e1