Introduction
Endoscopic submucosal dissection (ESD) is the treatment of choice for patients with
superficial gastric neoplasms with no or low risk of lymph nodes metastasis [1 ]
[2 ]. It permits en bloc resection of lesions regardless of their size, thus allowing accurate and reliable
histological assessment of resection margins and final staging to determine the curability
of the endoscopic resection. Endoscopic surveillance is recommended after gastric
ESD to inspect the healing process, rule out residual or recurrent tumor and screen
for metachronous lesions.
After a curative ESD, the postoperative scar typically appears homogeneous and whitish,
often with converging folds. Anecdotally we have noted at post-ESD endoscopic examination
that some patients develop irregular scars, with nodules some of which can be large
and protruding. Based on our observation, we named this entity polypoid nodule scar
(PNS) [3 ]
[4 ]. In our limited initial experience, biopsy specimens taken from these scars have
demonstrated regenerative mucosa with hyperplastic changes without recurrent neoplastic
cells. Presence of PNS creates a clinical dilemma because it may obscure neoplastic
recurrence, particularly considering that many of PNS-related nodules can be quite
large and protruding. Furthermore, in our initial series we noted a significant variation
in incidence of PNS from 1.7 % in the Japanese center up to 13.3 % in the Brazilian
center [4 ]. Therefore, we conducted a multicenter study to further investigate on a larger
scale incidence and clinical implications of PNS after curative gastric ESD.
Patients and methods
This was a multicenter retrospective case series involving five high-volume ESD centers
(1 in Brazil and 4 in Japan) as summarized in [Table 1 ]. The study was carried out in accordance with the Helsinki Declaration, and institutional
review board (IRB) approvals were obtained in each of the participating institutions.
All patients provided procedure informed consent. Informed consent for study participation
was waived because of the retrospective study design. Clinical information from 2003
to 2015 was extracted retrospectively from the endoscopy database of each institution.
Data were obtained by means of endoscopic and histological reports review, and assessment
of endoscopic images of the post-ESD scar.
Table 1
Endoscopic submucosal dissection procedures distribution among five participating
institutions and incidence of polypoid nodule scar, according to region of stomach.
ESD procedures
Center 1
Alfa Institute of Gastroenterology, Belo Horizonte – Brazil
Center 2
Osaka Medical Center for Cancer and Cardiovascular Diseases
Center 3
Kobe University School of Medicine
Center 4
National Cancer Center Hospital, Tokyo
Center 5
Japan Community Healthcare Organization Osaka Hospital
Total number of gastric ESD, n
79
343
241
1290
322
ESD in distal stomach, n (%)
49 (62.0 %)
158 (46.0 %)
69 (28.6 %)
495 (38.4 %)
141 (43.7 %)
ESD in proximal stomach, n (%)
30 (38.0 %)
185 (54.0 %)
172 (71.3 %)
795 (61.6 %)
181 (56.3 %)
Total number of PNS cases, n (%)
9 (11.4 %)
6 (1.7 %)
5 (2.0 %)
2 (0.15 %)
6 (1.8 %)
PNS in distal stomach, n (%)
9 (18.3 %)
6 (3.8 %)
5 (7.24 %)
2 (0.4 %)
6 (4.25 %)
ESD, endoscopic submucosal dissection; PNS, polypoid nodule scar; n, number
Inclusion and exclusion criteria
Eligibility for ESD was assessed by means of white-light endoscopy, image-enhanced
endoscopy (IEE), magnifying endoscopy, indigo carmine spraying and endoscopic ultrasound
(in selected cases, at the discretion of the endoscopist). Patients were included
in the study when all of the following were present: 1. successful en bloc ESD procedure with confirmatory histology of R0 and curative resection [1 ]; 2. postoperative endoscopic examination with identification of a protruded polypoid
nodule scar corresponding to the site where the ESD was undertaken; and 3. biopsies
of the PNS with histological assessment demonstrating hyperplastic or regenerative
tissue. Before data collection, each investigator involved in the study reviewed and
agreed on the typical endoscopic features of PNS previously described [4 ]. Experienced gastrointestinal pathologists from each center reviewed PNS biopsies.
In addition, endoscopic images of PNS cases from each Institution were sent to the
leading investigator (VA) aiming to check for interobserver agreement. Data were examined
for patient demographics, histologic findings, Helicobacter pylori status, postoperative medical management, clinical outcome and endoscopy follow-up
findings. Patients lost to follow-up that did not undergo at least one post-ESD examination
were excluded from this analysis. H. pylori infection was assessed by a combination of urea test, histological assessment of
gastric samples or serum antibody.
ESD procedure
The ESD technique has been described in detail elsewhere [5 ]. Briefly, markings were placed at least 2 mm beyond the borders of the lesion after
careful endoscopic assessment by chromoendoscopy and/or magnifying endoscopy (ME).
Viscous solution or 0.9 % saline solution was used for submucosal (SM) injection at
the discretion of the endoscopist. ESD was undertaken with specialized knives such
as Flush-Knife Ball Tipped (Fujifilm Co., Japan) or ceramic-ball insulated tip knife
(IT knife, Olympus Co., Japan), according to the preference of the endoscopist. Mucosal
incision was undertaken around the tumor in a circumferential or semi-circumferential
manner. SM dissection was performed in the deep submucosa with identification and
hemostasis of the penetrating vessels. After complete tumor resection, the ulcer site
was assessed and visible vessels were coagulated with a hemostatic forceps. The specimen
was stretched and fixed in a Styrofoam plate, immersed in 10 % formalin solution and
sent to the pathology department.
Histological assessment and definitions
After being embedded in 10 % paraffin, the specimens were cut into 2-mm slices and
stained with hematoxylin and eosin. Additional immunochemistry studies with D2 – 40
and CD34 were carried out for lymphatic and vascular invasion assessment, at the discretion
of the pathologist. Tumor size, depth of invasion, lymphatic and vascular invasion,
grade of differentiation, and resection margins were histopathologically examined
[6 ]. En bloc resection was defined as complete removal of the visible tumor including the markings
into one non-fragmented piece [2 ]. R0 resection was defined histologically as complete tumor removal with both lateral
and deep margins free of neoplastic cells. Endoscopic resection was considered curative
when the pathology report demonstrated adenoma with low- or high-grade dysplasia,
well or moderately differentiated adenocarcinoma, depth of invasion restricted to
mucosa or superficial submucosa (SM1), with negative lateral and deep margins and
no lymphovascular invasion [2 ]
[3 ]
[7 ]. ESD was considered non-curative when any one of the following criteria were present
[2 ]
[3 ]
[5 ]: undifferentiated cancer greater than 2 cm, deep submucosal tumor invasion (SM2),
positive lateral or horizontal margins, and lymphovascular invasion. Patients with
non-curative resection were not included in this study. PNS was defined as a protuberant
polypoid-appearing nodule situated in the post-ESD scar site, with or without converging
folds and with histological assessment of forceps biopsies demonstrating only regenerative
or hyperplastic tissue growth without any neoplastic tissue, confirmed by experienced
gastrointestinal pathologists.
Post-procedure care
All patients were hospitalized for postoperative observation. Intravenous or oral
proton pump inhibitors (PPI) were administered to all patients during the first postoperative
days followed by an 8-week course of oral PPI after hospital discharge. Second-look
endoscopy was performed at the discretion of the endoscopist. If the ESD procedure
was considered curative, the first follow-up endoscopy was scheduled usually at 3
and 12 months, and annually thereafter. The ESD scar was inspected carefully for any
abnormality such as residual tumor or polypoid nodule growth utilizing IEE or ME,
and multiple forceps biopsies were performed.
Statistical analysis
The primary outcome measure was assessment and characterization of post-ESD scars
at follow-up endoscopy. Descriptive statistical analysis of the data, including percentage,
median and 95 % confidence interval, were calculated. Categorical data were expressed
as frequencies and percentages.
Results
During the study period, 2865 patients from five centers underwent gastric ESD. A
total of 590 cases were excluded from analysis due to non-curative resection or absence
of postoperative follow-up. Therefore, 2275 patients that underwent curative gastric
ESD and received postoperative endoscopic follow-up entered the study. Among this
population, 28 patients (18 men/10 women) developed PNS for an overall occurrence
of 1.2 % [95 % confidence interval: 0.8 – 1.8 %]. The earliest appearance of PNS was
noted 3 months after ESD and the latest after 18 months (median: 6 months). The proportion
of PNS varied between the participating institutions and is summarized in [Table 1 ].
[Table 2 ] demonstrates the clinicopathological characteristics of patients with PNS. Tumor
size ranged from 8 mm to 55 mm (median 20 mm). Postoperative medical treatment included
PPI for all patients during 8 weeks (either standard or double dose at the discretion
of the endoscopist) and hospital admission for a minimum of 48 hours. Moreover, Center
#1 administered sucralfate postoperatively for 2 weeks and Center #3 added sucralfate
and sodium alginate for 1 week and rebapimide for 8 weeks. A total of 14 of 28 PNS
patients (50 %) tested positive for H. pylori . All patients that developed PNS had primary neoplastic lesions located in the distal
stomach (27 antrum/1 incisure). PNS was not observed in any patient with tumors located
in the body, fundus or cardia. Considering only lesions situated in the distal stomach,
the proportion of PNS cases was 3.1 % [95 % confidence interval: 2.0 – 4.4 %] (28
out of 912).
Table 2
Clinicopathological characteristics of patients with polypoid nodule scar.
Sex (male/female), n
18/10
Age (y), median (range)
73 (54 – 80)
Tumor size (mm), median (range)
20 (08 – 55)
Tumor location (distal stomach), n (%)
7 (25.0 %)
2 (7.0 %)
12 (43.0 %)
6 (21.5 %)
1 (3.5 %)
Tumor histology, n (%)
18 (64.3 %)
1 (3.5 %)
5 (17.9 %)
4 (14.3 %)
PNS disappearance over time, n (%)
5 (17.9 %)
H. pylori infection, n (%)
14 (50 %)
ESD, endoscopic submucosal dissection; PNS, polypoid nodule scar; n, number; H. pylori , Helicobacter pylori ; y, years; mm, millimeters
To assess the possibility of learning curve effect on occurrence of PNS, we divided
the study period into two phases (2003 – 2010 and 2011 – 2016). In 6 of 28 patients
(21.4 %), ESD was undertaken in the early period of experience from the operators
(2003 to 2010). The great majority of cases occurred after 2011, when all operators
had already overcome the learning curve for gastric ESD (first 40 cases).
All participating patients have been followed periodically with endoscopy and no neoplastic
recurrence was identified in any of the PNS patients (follow-up range: 3 to 192 months,
median:
41 months). In five cases (17.9 %) PNS disappeared over time (range: 10 – 30 months,
median: 24
months). In the remaining 23 cases, PNS remained present at follow-up endoscopic examination
without significant variations in size or shape, although occasionally some decrease
in size was
noted. [Fig. 1a ], [Fig. 1b ], [Fig. 1c ], [Fig. 1d ], [Fig. 1e ] and [Fig. 1f ] are illustrative of a PNS case with long-term follow-up.
Fig. 1 Illustrative case of polypoid nodule scar (PNS) after endoscopic submucosal dissection
(ESD) from Kobe University Hospital, Kobe, Japan. a Endoscopic view of defect after enbloc resection of a lesion situated in the pre-pyloric area. b Ulcerated scar 1 month after ESD. c Small nodule noted at ESD scar 1 year after ESD. d Protuberant nodule (PNS) observed 2 years after ESD. e PNS remains unchanged 3 years after ESD. f Histology from nodule biopsy demonstrates hyperplastic polypoid neoformation, showing
elongation, tortuosity and dilation of the gastric foveolae, and stroma with fibrosis
and inflammatory cells.
Discussion
The healing process of the post-ESD defect is still not well understood. The majority
of patients that undergo curative enbloc endoscopic resection of superficial gastric neoplasms develop homogeneous and flat
epithelized scar covered by a regular-appearing mucosa with correspondent fibrosis.
However anecdotally we noted, in two different institutions, that certain patients
would develop aberrant and protuberant polypoid nodules in the post-ESD scar. Biopsy
of these nodules demonstrated regenerative and hyperplastic tissue growth, and we
named this entity polypoid nodule scar (PNS) [3 ]
[4 ]. Presence of PNS may have direct clinical implications because it may render evaluation
of post-ESD neoplastic recurrence difficult. To better define the incidence and clinical
significance of PNS we conducted this large multicenter study including five high-volume
ESD expert centers (4 in Asia and 1 in Brazil). Our main findings are that PNS occurred
after curative gastric ESD in close to 1.2 % of the cases. Importantly, PNS occurred
only after ESD performed for antral lesions. If we consider only ESD procedures carried
out for tumors located in the antrum, the proportion of PNS cases was 3.1 %. Importantly,
in no patient with PNS neoplastic changes were detected and the lesions remain stable
on surveillance endoscopy but in approximately 18 %, PNS disappeared with time.
In this study we noted a wide variation in tumor size and ESD defect leading to development
of PNS, and likewise the size of PNS also presented some variation. The smallest tumor
treated by ESD that developed PNS measured only 8 mm, and the largest measured 55 mm.
Apparently tumor size or post-ESD ulcer size are not important factors in development
of PNS.
In our initial observation we did not detect a clear association between H. pylori status and development of PNS, since 57 % of the patients with PNS tested positive
for H. pylori . Now our current much larger investigation reconfirms that it is unlikely that H. pylori status influences occurrence of PNS, since half of our patients were H. pylori- positive. The effect of H. pylori eradication therapy on PNS remains to be elucidated in future studies.
The precise mechanism that explains why some individuals develop PNS and others do
not is still poorly understood. Similar aberrant PNS also have been noted to occur
after healing of benign gastric ulcers located in the antrum in up to 6 % of cases
[8 ], leading to unnecessary gastric resection in the past [9 ]. Other investigators have observed development of protruded granulation tissue in
healing ulcer after medical treatment either with histamine-2 receptor antagonist
(H2RA) [10 ] or PPI [11 ]
[12 ]. Interestingly, the protruded granulation tissue after benign ulcer healing may
disappear or remain unchanged for a long time [11 ]
[13 ] Histological findings of the polypoid nodule at benign ulcer scar similarly are
indicative of hyperplastic regenerative mucosa on fibrotic tissue [9 ].
Of note in this study, the proportion of PNS cases varied from 0.15 % to 11.4 % among
the participating institutions, although in three Japanese centers the appearance
was around 2 %. The reason for this wide variation may be related to differences in
ESD technique, total number of cases and operator experience, postoperative care or
even ethnic difference. To assess the influence of a learning curve effect on the
appearance of PNS, we divided the study period into two phases (2003 – 2010 and 2011 – 2016).
Only in six patients (21.4 %) ESD was undertaken in the early period (2003 to 2010).
The great majority of cases occurred after 2011, when all operators had already overcome
the learning curve for gastric ESD (first 40 cases). It is important to note that
the total number of PNS cases included in this study is still small and most likely
insufficient to allow a reliable statistical analysis. Future studies with larger
numbers of patients will be needed to address possible predisposing factors to PNS
appearance, such as ESD technique, type of knife or operator’s experience.
The great majority of patients undergoing gastric ESD are routinely treated postoperatively
with PPI to accelerate the healing process, a clinical management adopted universally
and in the five centers involved in this study [14 ]. However, there was some variation among the postoperative medical management, with
Center 1 administering sucralfate for 2 weeks and Center 3 adding sucralfate and sodium
alginate for 1 week and rebapimide for 8 weeks. PPI accelerates ulcer healing mainly
due to potent gastric acid secretion inhibition. However, PPI also increases cyclooxigenase-2
(COX-2) expression and prostaglandin E synthases in the ulcerated mucosa [15 ]. COX-2-generated prostaglandin E2 stimulates expression of growth factors such as
vascular endothelial growth factor (VEGF) [16 ], hepatocyte growth factor (HGF) [17 ], and basic fibroblast growth factor (bFGF) [18 ] in the mucosa. This accelerated mucosal repair and angiogenesis may contribute to
nodular overgrowth of the regenerative mucosa. The influence of other medications,
such as sucralfate, sodium alginate or rebapimide on the scarring process of ESD defects,
need to be addressed in future studies.
A unique universal characteristic of PNS is that this aberrant scar only occurs after
ESD performed in the distal stomach (antrum or incisure). No PNS was observed after
resection of lesions located in the gastric body, fundus or cardia. Likewise, so far,
we have not observed this finding after esophageal or colorectal ESD. The reason for
this phenomenon is unclear. We postulate that gastric peristalsis and/or bile reflux
may enhance development of PNS in the antrum. Moreover, the submucosal layer in the
antrum is thicker; therefore, inflammatory or regenerative reaction in the submucosa
may be more obvious in the antrum than in the corpus of fundus.
Our study was not without limitations. It was a retrospective analysis of prospectively
collected data. Although the endoscopic database of each center participating in this
study was organized and continuously updated, some cases of PNS may have been missed
during data extraction. Furthermore, only patients that returned for follow-up were
enrolled, therefore other potential cases of PNS may have been missed. Finally, there
might be overlooked cases of PNS that were not characterized by the expert endoscopist,
or even referred to surgery or endoscopic intervention, due to fear of recurrence
or malignancy, particularly in the early period of the study when ESD was not yet
a well-established procedure. These limitations suggest that the precise incidence
rate for PNS may be even greater than the rate of 1.2 % observed in this multicenter
study, and a prospective study is needed to further investigate the incidence rate
of PNS and its relationship with postoperative medical management, H. pylori status/eradication therapy and patient/tumor characteristics.
Conclusion
In conclusion, this large multicenter study demonstrates that after curative ESD,
PNS develops only after removal of lesions in the distal stomach in about 3.1 % of
cases. Endoscopists performing ESD should be aware that PNS may develop and although
it can have an alarming appearance, it is a benign condition. Based on our data, patients
with PNS can be reassured and followed with annual endoscopic surveillance to search
for metachronous lesions and to examine and biopsy the ESD scar.
