Keywords Endoscopy Upper GI Tract - Endoscopic resection (ESD, EMRc, ...) - Dilation, injection,
stenting - Barrett's and adenocarcinoma
Introduction
Endoscopic submucosal dissection (ESD) is a well-established treatment for superficial
esophageal squamous cell carcinoma (ESCC), enabling en-bloc resection and accurate
pathological evaluation [1 ]
[2 ]
[3 ]. Current guidelines recommend ESD for cT1a-EP/LPM lesions up to 50 mm in length.
Particularly for circumferential lesions, this size limit is emphasized due to the
high risk of postoperative esophageal stricture (ES) [4 ].
However, in clinical practice, whole circumferential ESD (WC-ESD) is frequently performed
beyond these criteria. A recent Japanese multicenter survey reported that more than
half the institutions had performed WC-ESD under extended indications, including lesions
> 50 mm or with suspected submucosal invasion [5 ]. These real-world practices reflect a need for evidence regarding the feasibility
and safety of WC-ESD beyond guideline limitations.
Notably, the current recommendation regarding the 50-mm threshold is largely based
on limited evidence, such as the single-center study by Miwata et al., which included
only 19 patients undergoing circumferential ESD [6 ], and long-term outcomes after WC-ESD for larger lesions remain unclear.
Therefore, this study aimed to evaluate short-term safety and long-term outcomes of
WC-ESD, focusing on lesions > 50 mm in length, and to assess risk of esophageal stricture
(ES) and its management in a real-world cohort.
Patients and methods
Study design and population
This retrospective observational study was conducted in a single center at the Keiyukai
Second Hospital, Hokkaido, Japan. We retrospectively analyzed 766 consecutive patients
who underwent ESD for superficial cESCC (cT1N0M0) between January 2009 and June 2019.
Exclusion criteria were described as follows: patients who had undergone prior chemoradiotherapy
for ESCC, those with multiple ESCC lesions treated in a single ESD session, or patients
who had two or more ESD procedures for ESCC during the study period.
To evaluate clinical outcomes of WC-ESD with a longitudinal extension > 50 mm, patients
were classified into two groups: WC-ESD with a longitudinal extension > 50 mm (> 50
mm group) and ≤ 50 mm (≤ 50 mm group). Each group was further categorized into pT1a-EP/LPM
and pT1a-MM/pT1b (SM1/SM2) subgroups based on post-ESD histopathological findings
for prognostic analysis. Primary analysis excluded patients who initiated surgery
or chemoradiotherapy (CRT) within 90 days after ESD when estimating short‑term stricture
incidence. In the sensitivity analysis, all patients were included, counting events
that occurred before surgery/CRT initiation and censoring at that time. Long-term
outcome analysis (overall survival [OS]/disease-specific survival [DSS]/recurrence-free
survival [RFS]) included all patients, with surgery/CRT considered part of real-world
management. In addition, descriptive data from non-WC-ESD cases were included as supplementary
information to aid interpretation of patient backgrounds (Supplementary Table 1 ).
Information about the study was posted on the hospital website, allowing participants
to opt out. Patients who did not opt out of the study were considered to have provided
tacit consent to participate. The study protocol was reviewed and approved by the
Review Board of Keiyukai Second Hospital (approval date: May 8, 2021). The study adhered
to the ethical principles outlined in the Declaration of Helsinki [7 ].
Indication for ESD
Extent of the lesions was determined using image-enhanced endoscopy and Lugol’s chromoendoscopy.
Invasion depth of ESCC was estimated using white-light non-magnifying endoscopy to
assess changes on the lesion surface. Narrow band imaging magnifying endoscopy was
used to further evaluate abnormal microvessels according to the Japan Esophageal Society
classification.
Computed tomography (CT) and endoscopic ultrasound (EUS) were used to rule out lymph
node metastasis, and positron emission tomography/CT was performed as necessary. As
recommended by the guidelines, we performed ESD for clinically diagnosed T1a-EP/LPM
[4 ].
ESD procedures
Sedation was achieved using midazolam (5–10 mg), pethidine (0–50 mg), and
dexmedetomidine hydrochloride (loading dose of 6 μg/kg over 10 minutes, maintenance
at
0.2–0.7 μg/kg/h, with adjustments based on patient condition. Electrosurgical units
used
were ICC-200 (Intelligent Cut and Coagulation; Erbe, Tübingen, Germany) and VIO3 (Erbe).
The
circumference of the lesion was first marked using spotty electrocautery with a needle
knife
1 mm from the tumor border and confirmed by Lugol-unstained regions. Normal saline
mixed
with glycerin (ranging from 10 to 100 mL, with an average of 40 mL) and epinephrine
(0.005%)
was injected into the lesion. Glycerin concentration was 10% of the weight/volume
(Chugai
Pharmaceutical Co., Ltd., Tokyo, Japan). The marked area was incised along its perimeter
using a hook knife (Olympus, Tokyo, Japan). Submucosal dissection of the lesion was
performed after mucosal incision. Two expert endoscopists (HT and SO) performed all
ESD
procedures. In circumferential ESD cases, the clip-with-line method [8 ] and endoscopic submucosal tunnel dissection [9 ] were utilized when appropriate.
Definition of WC-ESD
WC-ESD was defined as a complete circumferential esophageal mucosal defect following
ESD ([Fig. 1 ]). Patients with remaining esophageal mucosa were classified as non-WC-ESD [10 ].
Fig. 1 A complete circumferential esophageal mucosal defect following ESD.
Steroid treatment
To prevent ES after ESD, patients with mucosal defects involving more than three-quarters
of the esophageal circumference were immediately injected endoscopically with triamcinolone
acetonide (TA) (Kenacort-A, 40 mg/mL, Bristol-Myers Squibb, Anagni, Italy) (Supplementary Fig. 1 ). TA was diluted with 0.9% NaCl to a final concentration of 10 mg/mL. Aliquots (0.5
mL) were injected at the base of the artificial ulcer using a 25G, 3-mm needle (TOP
Corporation, Tokyo, Japan). Injections were initiated at the distal edge of the ulcer
base and evenly spaced at 10-mm intervals toward the proximal edge, taking care to
avoid injury to the muscularis propria [11 ]. Oral prednisolone was started at 30 mg/day on the second day post-ESD and continued
with a gradually tapering prednisolone dose (30, 25, 20, 15, 10, and 5 mg/day). Systemic
steroid administration was discontinued after 12 weeks. At the time when we started
WC-ESD at our hospital, even in cases of WC-ESD, there were some cases in which we
did not perform steroid injections and oral steroids. Furthermore, we adopted oral
steroids at our hospital in 2018. Regarding local steroid injections, we perform them
on the ulcer base on the day of WC-ESD, but not during subsequent dilation.
ES
ES was defined as inability to pass a standard diagnostic gastroscope (which had a
diameter of 9.8 to 10.2 mm; GIF-H260, GIF-H290Z, or GIF-HQ290; Olympus Medical Systems,
Tokyo, Japan) through the stenotic segment. In addition to endoscopic findings, clinical
symptoms such as dysphagia were also taken into account when diagnosing ES, if documented
in the medical records.
Dilation therapy
Dilation therapy was administered to the patients with dysphagia resulting from ES.
Prophylactic dilation was also performed in patients who underwent WC-ESD (Supplementary Fig. 1 ). The interval between dilations ranged from 1 to 4 weeks. Dilation was repeated
until dysphagia symptoms resolved and a gastroscope (8.9–9.9 mm) could pass without
resistance 4 weeks after the final session. Dilation therapy was conducted in the
outpatient department using a silicone bougie (Maloney, Medovations, Milwaukee, Wisconsin,
United States) or a balloon dilator (CRE Fixed Wire 15 mm/16.5 mm/18 mm, Boston Scientific,
Massachusetts, United States). We defined refractory strictures as “cases that underwent
dilatation six times or more.”
Pathological examination
Resected specimens were sectioned perpendicularly at 2-mm intervals and examined by
a board-certified gastrointestinal pathologist. Tumor size, histological type, differentiation,
lymphovascular invasion, and lateral and deep margins were evaluated according to
the 11th Edition of the Japanese Classification of Esophageal Cancer [12 ]. Complete resection (R0) was defined as en-bloc resection with pathological tumor-free
lateral and vertical margins. Histopathological analysis confirmed that curative resection
revealed pT1a-EP/LPM with no lymphovascular invasion.
Follow-up protocol and additional treatment
After ESD, patients with pT1a-EP/LPM ESCC underwent regular endoscopic surveillance.
The first follow-up endoscopy was conducted 6 months after ESD to assess mucosal healing,
with subsequent endoscopic surveillance every 6 months. For patients with pT1a-MM
and T1b ESCC, with or without additional treatment, follow-up included measurement
of tumor markers, endoscopy and CT every 3 to 6 months for 5 years.
We recommend additional treatment with surgery or CRT (50.4 Gy external-beam irradiation
with 5-fluorouracil and cisplatin) for patients with pT1b ESCC, positive lymphovascular
invasion, or positive/unclear VM (VM1/VMX). Some patients were monitored without additional
treatment based on their refusal or overall condition.
Outcome
The primary endpoint was incidence of short-term adverse events (AEs), particularly
ES, in the > 50-mm and ≤ 50-mm groups. This was assessed in the primary analysis,
which excluded patients who initiated surgery or CRT within 90 days after ESD. A sensitivity
analysis was also performed, including all patients and counting events that occurred
before surgery/CRT initiation. Additional short-term outcomes included number of dilatations
and use of steroid therapy.
Secondary endpoints focused on long-term outcomes, including OS, DSS, and RFS in the
> 50-mm and ≤ 50-mm groups. Both groups were further stratified into the pT1a-EP/LPM
and pT1a-MM/pT1b(SM1/SM2) subgroups.
OS was defined as time from the date of ESD to date of death from any cause other
than esophageal cancer. DSS was defined as time from ESD to death, specifically from
date of primary ESCC. Recurrence sites were categorized as local lymph nodes or distant
organs.
Statistical analysis
All statistical analyses were performed using SPSS Statistics version 22 (IBM Japan,
Tokyo, Japan). For the primary analysis of short-term outcomes, patients who underwent
surgery or CRT within 90 days after ESD were excluded when estimating incidence of
post-ESD stricture. In sensitivity analyses, all patients were included, counting
events occurring before surgery/CRT and censoring at the date of surgery/CRT initiation.
For long-term outcomes, all patients were included, and surgery/CRT was regarded as
part of real-world management.
Categorical variables were compared using the chi-square test or Fisher’s exact test,
and continuous variables using the Mann-Whitney U test. Kaplan-Meier curves with log-rank
tests were used for time-to-event analyses. Multivariate logistic regression was performed
to identify factors associated with stricture and refractory stricture, adjusting
for clinically relevant covariates. A two-sided P < 0.05 was considered statistically significant.
Results
Baseline characteristics
In total, 766 patients who underwent ESD for ESCC were included in this study. The
exclusion criteria included 47 patients who had received prior CRT, 62 patients with
multiple ESCC lesions resected in one ESD session, and 85 patients with two or more
ESD treatments for ESCC. After applying these criteria, 67 patients were categorized
into the WC-ESD group and 505 into the non-WC-ESD group ([Fig. 2 ]). In the WC-ESD group, which was divided into two groups based on longitudinal extension
longer than 50 mm or less, the > 50-mm group consisted of 28 patients and the ≤ 50-mm
group consisted of 39 patients. Among the cases that underwent additional treatment
with surgery or CRT, there were six cases in the > 50-mm group and nine cases in the
≤ 50-mm group. For short-term outcomes, including ES, the analysis was be conducted
in the two groups (> 50-mm group: 28/≤ 50-mm group: 39) excluding cases that underwent
additional treatment (surgery or CRT). Regarding long-term outcomes, we evaluated
the two groups (> 50-mm group: 28/≤ 50-mm group: 39) including cases that underwent
additional treatment (surgery or CRT). Median follow-up period was 61 months (interquartile
range [IQR], 18–135), and was significantly shorter in the > 50-mm group than in the
≤ 50-mm group (64 [18–56] vs. 85 [26–146] months, P = 0.0244). The follow-up rate at the last observation was 85% at the 5-year follow-up.
Fig. 2 Patient flow chart of the study.
Clinicopathological characteristics of patients in the > 50-mm and ≤ 50-mm groups
are summarized in [Table 1 ]. There was no significant difference in age between the two groups. The lesions
were predominantly located in the middle thoracic esophagus in both groups, with 0-IIc
being the most common macroscopic type.
Table 1 Baseline patient and lesion characteristics and pathological outcomes of WC-ESD (<
50-mm group and > 50-mm group).
Total (%)
(n = 67)
Axis length > 50 mm (n = 28)
Axis length ≤ 50 mm (n = 39)
P value
WC-ESD, whole circumferential endoscopic submucosal dissection.
* Significant at P < 0.05.
† Mann-Whitney's U-test.
‡ Chi-square test.
§ Fisher's exact test.
Age at endoscopic treatment, median (IQR), years
71 (48–89)
73.5 (48–84)
70 (54–89)
0.494†
Sex
0.217§*
55 (82)
21 (75)
34 (87)
12 (18)
7 (25)
5 (13)
Tumor location
0.958§*
3 (5)
1 (4)
2 (5)
12 (18)
5 (18)
7 (18)
36 (53)
16 (57)
20 (51)
16 (24)
6 (21)
10 (26)
Macroscopic type
0.590§*
1 (2)
0 (0)
1 (3)
8 (12)
2 (7)
6 (15)
37 (55)
17 (61)
20 (51)
21 (31)
9 (32)
12 (31)
Size of tumor, median (IQR), mm
65 (21–111)
71 (57–111)
57 (26–92)
0.0002†*
Depth of invasion
46 (69)
20 (71)
26 (67)
0.679§*
21 (31)
8 (29)
13 (33)
0.517§*
28 (42)
11 (39)
17 (44)
18 (27)
9 (32)
9 (23)
6 (9)
1 (4)
5 (13)
15 (22)
7 (25)
8 (20)
Lymphatic invasion (%)
10 (15)
4 (14)
6 (15)
1.000‡
Venous invasion (%)
5 (8)
1 (4)
4 (10)
0.391‡
Long-term outcomes
For pT1a-EP/LPM cases (> 50-mm group 39%; 11/28, ≤ 50-mm group 44%; 17/39),
5-year OS was 100% (95% CI, not estimable), DSS was 100% (95% CI, not estimable),
and RFS
was 100% (95% CI, NE) in both groups ([Fig. 3 ]
a-c ). In pT1a-MM and pT1b (SM1/SM2) cases (> 50-mm group
61%; 17/28), ≤ 50-mm group 56%; 22/39), 5-year OS was lower in the > 50-mm group than
in
the ≤ 50-mm group 68.9% (95% CI 38.8–99.2) vs 100% (95% CI-NE), respectively ([Fig. 4 ]
a ), and there were no significant differences in 5-year DSS
(100% vs 100%; 95% CI- NE and 95% CI-NE), respectively, and 5-year RFS between the
two
groups (90.9% and 83.3% and 95% CI 51.3- 99.4 and 95% CI 36.9- 97.7, respectively)
([Fig. 4 ]
b,c ). The proportions of patients who required additional
treatment in the > 50-mm and ≤ 50-mm groups were 21% (6/28) and 26% (10/39), respectively
(P = 0.777).
Fig. 3
a Kaplan-Meier curves of OS for patients with pT1a-EP/LPM disease in the > 50-mm and
≤ 50-mm groups. b Kaplan-Meier curves of DSS for patients with pT1a-EP/LPM disease in the > 50-mm and
≤ 50-mm groups. c Kaplan-Meier curves of RFS for patients with pT1a-EP/LPM disease in the > 50-mm and
≤ 50-mm groups.
Fig. 4
a Kaplan-Meier curves of OS for patients with pT1a-MM and pT1b disease in the > 50-mm
and ≤ 50-mm groups. b Kaplan-Meier curves of DSS for patients with pT1a-MM and pT1b disease in the > 50-mm
and ≤ 50-mm groups. c Kaplan-Meier curves of RFS for patients with pT1a-MM and pT1b disease in the > 50-mm
and ≤ 50-mm groups.
The clinical course of the 67 patients who underwent WC-ESD is shown in [Fig. 5 ]. No procedure-related deaths occurred in the WC-ESD group. However, one patient
(2%) died of primary ESCC and seven died of unrelated causes.
Four patients in the WC-ESD group (6%, 4/67) experienced nodal recurrence; three had
pT1b, and one had pT1a-MM with vascular invasion. Among them, two underwent CRT and
one patient who declined additional treatment died of ESCC-related causes (Supplementary Table 2 ).
Fig. 5 Clinical progression of WC-ESD with a longitudinal extension of > 50-mm cases. AT,
additional treatment; Chemo, Chemotherapy.
Procedure characteristics and adverse events
Procedure characteristics and AEs in the > 50-mm and ≤ 50-mm groups are detailed in
[Table 2 ]. En-bloc resection was achieved in 100% of the cases in both groups. R0 resection
rates were 95% in the > 50-mm group and 87% in the ≤ 50-mm group, respectively. The
curative resection rate was not significantly different between the > 50-mm group
and the ≤ 50-mm group (86% vs. 77%; P = 0.488).
Frequency of AEs was not significantly different between the > 50-mm group (18%, 4/22
patients) and the ≤ 50-mm group (13%, 4/13 patients; P = 0.708). These included ES. No cases of delayed bleeding and intraoperative perforation
were observed in either group.
Results of ES treatment are presented in [Table 2 ]. The proportion of patients in the > 50-mm group who required dilations was higher
than in the ≤ 50-mm group, with no significant difference (96% vs. 83%, P = 0.226). Median number of dilations was not significantly different between the
> 50-mm group and the ≤ 50-mm group (8 vs. 7.5) (IQR 0–39 vs. 0–53, respectively;
P = 0.205). One patient in the > 50-mm group developed a perforation during dilation
and required surgery. In the sensitivity analysis including patients who underwent
surgery or CRT, stricture occurred in six of 28 patients (21%) in the > 50-mm group
and six of 39 patients (15%) in the ≤ 50-mm group (OR 1.50; 95% CI 0.43–5.25; P = 0.538), consistent with the primary analysis. Local injection of steroids for preventing
post-ESD strictures for prophylaxis was administered to 96% of the > 50-mm group and
87% of the ≤ 50-mm group. Two patients who received oral steroids were in the > 50-mm
group patient (9% vs. 0%, P = 0.174).
Table 2 Procedure characteristics and adverse events of WC-ESD and outcomes related to treatment
for stricture.
Total (%) (n = 52)
Axis length > 50 mm (n = 22)
Axis length ≤ 50 mm (n = 30)
P value
IQR, interquartile range.
* Significant at P < 0.05.
† Mann-Whitney's U-test.
‡ chi-square test.
§ Fisher's exact test.
En-bloc resection (%)
52 (100)
22 (100)
30 (100)
-
R0 resection (%)
47 (90)
21 (96)
26 (87)
0.381§
Curative resection (%)
42 (81)
19 (86)
23 (77)
0.488§
Adverse events, n (%)
8 (15)
4 (13)
4 (18)
0.708§
Delayed bleeding, n (%)
0 (0)
0 (0)
0 (0)
-
Esophageal stricture, n (%)
8 (15)
4 (13)
4 (18)
0.708§
Intraoperative perforation, n (%)
0 (0)
0 (0)
0 (0)
-
Dilation, n (%)
46 (89)
21 (96)
25 (83)
0.226§*
Number of dilations, median (IQR), times
8 (0–53)
8 (0–39)
7.5 (0–53)
0.205§*
Complications after dilation (perforation), n (%)
1 (2)
1 (5)
0 (0)
0.423§
Refractory stricture, n(%)
35 (67)
16 (73)
19 (63)
0.473§*
Steroid, n (%)
47 (90)
21 (96)
26 (87)
0.381§*
Injection, n (%)
47 (90)
21 (96)
26 (87)
0.381§*
Oral, n (%)
2 (4)
2 (9)
0 (0)
0.174§*
Discussion
This retrospective study evaluated clinical safety and long-term outcomes of WC-ESD
for ESCC with a long axis length > 50 mm compared with WC-ESD for lesions ≤ 50 mm.
No significant differences were observed between the two groups in the incidence of
AEs, including ES, or in 5-year OS, DSS, and RFS. The proportion of patients requiring
endoscopic dilation was numerically higher in the > 50-mm group than in the ≤ 50-mm
group (93% vs. 77%, P = 0.0825), whereas median number of dilations was comparable between the groups.
These results provide realistic evidence demonstrating the feasibility of WC-ESD beyond
the limitations of current guidelines.
Current Japanese guidelines recommend WC-ESD for circumferential ESCC measuring ≤
50 mm in length, primarily due to the high risk of postoperative strictures [6 ]. Nevertheless, multicenter surveys in Japan have shown that WC-ESD is often performed
for larger lesions when technically feasible [5 ]. Most previous studies have focused on short-term safety in guideline-compliant
cases [13 ]
[14 ] and long-term prognosis of patients who had received WC-ESD [15 ]. Furthermore, the guidelines provide no detailed recommendations for WC-ESD in other
ESCC presentations, such as circumferential lesions exceeding 50 mm or those with
submucosal invasion [4 ]. Therefore, our study addresses this evidence gap by providing real-world data on
both short- and long-term outcomes of WC-ESD for lesions > 50 mm and demonstrating
that preventive strategies that may mitigate associated risks. Data from the non-WC-ESD
group were presented as supplementary information to aid interpretation.
Five-year OS for pT1a-MM and pT1b appeared numerically lower in the > 50-mm group
than in the ≤ 50-mm group (68.9% vs. 100%, P = 0.07) and all deaths in these subgroups were from non-esophageal cancer causes.
Median follow-up duration was 61 months (range, 8–148 months), with an 85.0% follow-up
rate at 5 years. This follow-up rate reflects the challenges inherent in retrospective,
real-world studies, in which the decreasing number of patients in long-term observation
is common, but survival analysis appropriately accounted for censoring, and there
was no significant difference in baseline characteristics between patients lost to
follow-up and those retained, supporting the validity of our survival outcomes.
Curative resection rates in this study (68% in the > 50-mm group and 59% in the ≤
50-mm group) were lower than expected. This may be due to the difficulty in accurately
predicting depth of invasion in circumferential or large lesions, leading to underestimation
of submucosal invasion even in lesions ≤ 50 mm. Some lesions initially judged as mucosal
or muscularis mucosae were found to have submucosal invasion on histological analysis,
leading to non-curative resection. This diagnostic limitation is consistent with previous
observations in large, flat esophageal lesions [16 ] and underscores the need for improved preoperative staging techniques, such as high-frequency
EUS or advanced imaging algorithms.
We found no significant differences in AEs, including ES, between the > 50-mm and
≤ 50-mm groups (13% vs. 18%). In a sensitivity analysis including patients who underwent
surgery or CRT, stricture incidence remained similar (21% vs. 15%). The lower stricture
rate compared with previous reports is likely attributable to prophylactic balloon
dilation and steroid therapy, which reduced incidence of refractory strictures from
25% to 7%. These findings suggest that WC-ESD for lesions > 50 mm can be performed
safely when such preventive measures are applied, although further improvements in
stricture prevention remain necessary [17 ]
[18 ]
[19 ].
This study has several limitations. First, its retrospective single-center design
and relatively small sample size may limit generalizability of the findings. Second,
the procedures were performed by only two endoscopists, which might have influenced
treatment outcomes. Third, the imbalance in the number of cases between the > 50-mm
and ≤ 50-mm groups complicates direct comparisons. Lastly, the balloon dilation data
included both preventive and therapeutic procedures, making it difficult to precisely
estimate the number of dilations required for stricture treatment. Large-scale registries
or pooled analyses of real-world data could provide more robust evidence regarding
safety and efficacy of WC-ESD for extensive esophageal lesions.
Conclusions
In this large single-center cohort, WC-ESD for superficial ESCC in pT1a-EP/LPM case
with tumor length > 50 mm showed no significant increase in AEs, including esophageal
stricture, and the number of endoscopic dilations required was not substantially different
from that for lesions < 50 mm when prophylactic steroid therapy and preventive dilation
were implemented. In addition, long-term OS, DSS, and RFS were not substantially different
between the two groups. These findings provide real-world evidence supporting the
feasibility of WC-ESD beyond current guideline limits, while underscoring the need
for ongoing strategies to reduce the high stricture rate.
Bibliographical Record
