Outcomes and precautions of endoscopic submucosal dissection for undifferentiated-type early gastric cancer

Abstract

Objectives:

Since the development of techniques for endoscopic submucosal dissection (ESD), the indication range of endoscopic resection (ER) has been extended in early gastric cancer (EGC) treatment. For undifferentiated-type EGC, tumors with an intramucosal depth of invasion, no ulceration and a diameter of 20 mm or less were included in the expanded indications for ER in the Japanese Gastric Cancer Treatment Guidelines 2010. Nonetheless, because of difficulty in detecting lesions that meet the criteria for ER, the number of endoscopically resected cases of undifferentiated-type EGC is less than that of differentiated-type EGC.

Methods:

We retrospectively investigated the outcomes of ESD in 38 patients with 40 lesions of EGC in which the dominant pathological type was confirmed to be undifferentiated (signet ring cell carcinoma, poorly differentiated adenocarcinoma, mucinous adenocarcinoma) on histological examination of resected specimens.

Results:

Margin involvement and submucosal infiltration were common noncurative factors. Precise evaluation of the area and depth of lesions is a problem to be solved. Among a total of five patients with involved or uncertain horizontal margins, one of two patients who underwent additional surgery had residual cancer, and one of three patients who were observed had recurrence.

Conclusions:

Undifferentiated-type EGC with a positive horizontal margin may relapse after ESD. It is therefore essential to precisely evaluate the area of the lesion and to perform resection with an adequate safety margin to decrease the risk of recurrence.

Introduction

The development of techniques for endoscopic submucosal dissection (ESD) has considerably reduced the technical limitations of endoscopic resection (ER) in patients with early gastric cancer (EGC) [Ono et al. 2008; Takeuchi et al. 2007; Gotada et al. 1999, 2006; Oda et al. 2005; Yamamoto et al. 2003; Oyama et al. 2002]. Nonetheless, because ER entails local resection of the primary lesion without lymph node dissection, the indication range of ER has been restricted to EGC with a negligible risk of lymph node metastasis [Rembacken et al. 2001; Soetikno et al. 2003]. In undifferentiated-type EGC, tumors with an intramucosal depth of invasion, no ulceration and a diameter of 20 mm or less have been reported to have a very low risk of lymph node metastasis [Gotoda et al. 2000] and were thus included in the expanded indications for ER in the Japanese Gastric Cancer Treatment Guidelines 2010 [Japanese Gastric Cancer Association, 2011].

Because of difficulty in detecting lesions that meet the criteria for ER [Chonan et al. 1996; Nishimata et al. 1996; Ohyama et al. 1996; Fuchigami et al. 1996], the number of endoscopically resected cases of undifferentiated-type EGC is less than that of differentiated-type EGC [Yamamoto et al. 2010, 2012; Yokoi et al. 2009; Kang et al. 2010]. Moreover, even if the risk of lymph node metastasis is very low in patients who meet the expanded indications for ER and undergo surgery, it remains unclear whether the outcomes of patients who undergo ER of undifferentiated-type EGCs that meet the expanded indications for ER are equivalent to those who undergo surgery in terms of survival or local control.

Patients and methods

The study group comprised 38 patients with 40 EGC lesions treated by 39 sessions of ESD in Kanagawa Cancer Center Hospital between July 2005 and December 2013. In all patients, the dominant pathological type was confirmed to be undifferentiated (signet ring cell carcinoma, poorly differentiated adenocarcinoma, mucinous adenocarcinoma) on histological examination of endoscopically dissected specimens.

The pretreatment status, results of ESD, additional treatment and outcomes were evaluated retrospectively in all patients. Pretreatment status included the age and sex of the patients and the location and macroscopic type of the tumors. The following variables were recorded to assess the results of ESD: size of dissected specimen; presence or absence of an incision in the tumor site; complications; and pathological findings. As for the results of ESD, the curability of each endoscopically dissected specimen was evaluated according to the Japanese Gastric Cancer Treatment Guidelines 2010 [Japanese Gastric Cancer Association, 2011]. In patients who underwent additional surgery, the presence or absence of local residual cancer at the site of the endoscopically dissected lesion and the presence or absence of lymph node metastasis were examined. Outcomes were evaluated on the basis of overall survival (OS), which was measured from the date of ESD to the date of death or the last follow-up visit.

Results

Of the 38 patients in this study, the median age was 69.5 years (range: 40–88). A total of 25 patients (65.8%) were male and 13 (34.2%) were female. Of the 40 treated tumors, the most common location was the middle portion of the stomach (M)/greater curvature (Gre) and the most common macroscopic appearance was superficial depressed type (0-IIc, 85.0%). Overall, 92.5% of the tumors were depressed types [0-IIc, superficial protruding type (0-IIa) + IIc] (Table 1). A total of 39 of the 40 lesions were endoscopically dissected en bloc and the other was dissected fractionally in two pieces. During 39 sessions of ESD, postoperative hemorrhage occurred 3 times (7.7%) and intraoperative perforation occurred once (2.6%) as complications (Table 2). None of the patients with complications required emergency surgery.

Table 1.

Clinical findings of patients and tumors.

Total number of patients	n = 38
Age (years)
median	69.5
mean	68.9
range	40–88
Sex
male	25	(65.8%)
female	13	(34.2%)
Total number of tumors	n = 40
Location
U	2	(5.0%)
M	23	(57.5%)
L	14	(35.0%)
remnant stomach*	1	(2.5%)
Cross-sectional parts
Less	10	(25.0%)
Gre	17	(42.5%)
Ant	6	(15.0%)
Post	7	(17.5%)
Macroscopic type
0-IIa	1	(2.5%)
0-IIb	2	(5.0%)
0-IIc	34	(85.0%)
0-IIa+IIc	3	(7.5%)

^*After surgery for duodenal ulcer.

U, upper third of stomach; M, middle third of stomach; L, lower third of stomach; Less, lesser curvature; Gre, greater curvature; Ant, anterior wall of stomach; Post, posterior wall of stomach.

Table 2.

Results of endoscopic submucosal dissection (ESD).

Total number of specimens	n = 39
Specimen size (mm)
median	40
range	19–75
⩽30 mm	4	(10.3%)
31–50 mm	28	(71.8%)
>50 mm	7	(17.9%)
Total number of tumors	n = 40
Incision in tumor site
absent	37	(92.5%)
present	3	(7.5%)
Margin involvement
not involved	31	(77.5%)
involved/unknown	9	(22.5%)
En bloc resection	39	(97.5%)
Complete en bloc resection	30	(75.0%)
Complications (total ESD n = 39)
bleeding	3	(7.7%)
perforation	1	(2.6%)

The pathological findings of the endoscopically dissected tumors are summarized in Table 3. The median tumor size was 17 mm (range: 2–36), and 11 (27.5%) of the 40 tumors were larger than 20 mm. The depth of invasion was mucosal (M) for 30 lesions (75.0%), submucosal (SM) for 9 (22.5%) and uncertain for 1 (2.5%). Capillary invasion was present or unknown in 4 (10.0%) lesions, all of which had SM invasion. The dissection margin was involved or the margin status was unknown in 9 (22.5%) of the 40 tumors, including 3 lesions in which the margin status was unclear because an incision had been made in the tumor site. The ESD treatment was evaluated to be pathologically curative for 16 (40.0%) of the 40 endoscopically dissected tumors and noncurative for 24 (60.0%). Among the tumors that underwent noncurative ESD, histological examination of the dissected specimens revealed that the frequencies of tumors associated with noncurative factors such as a size exceeding 20 mm, a depth of invasion of SM or uncertain, intramural ulceration and margin involvement were similar (22.5–27.5%). Among the 24 noncuratively dissected tumors, 15 were evaluated to be out of the indication range before ESD. The other nine cases were evaluated to meet the criteria for the expanded indications before ESD. The noncurative factors are summarized in Table 4. Among 3 noncurative factors that were assessable before ESD (size >20 mm, SM depth of invasion, presence of intramural ulceration), misdiagnosis of the depth of invasion was most common (36.0%) for lesions evaluated to meet the expanded indications before ESD. However, margin involvement was positive or uncertain in 9 (22.5%) of the 40 dissected specimens.

Table 3.

Pathological findings.

Total number of tumors	n = 40
Tumor size (mm)
median	17
range	2–36
⩽20 mm	29	(72.5%)
>20 mm	11	(27.5%)
Depth
M	30	(75.0%)
SM	9	(22.5%)
uncertain	1	(2.5%)
Intratumoral ulceration
UL (–)	31	(77.5%)
UL (+)	9	(22.5%)
Capillary invasion
absent	36	(90.0%)
present/unknown	4	(10.0%)
Margin involvement
not involved	31	(77.5%)
involved/unknown	9	(22.5%)
HM1/HMX	5	(12.5%)
VM1/VMX	5	(12.5%)
Curability
expanded indication* (n = 25)	16/25	(64.0%)
out of indication* (n = 15)	0/15	(0.0%)
total	16/40	(40.0%)

^*Pre-ESD evaluation.

ESD, endoscopic submucosal dissection; HM1, horizontal margin is involved; HMX, horizontal margin involvement is unknown; M, mucosal; SM, submucosal; UL(–), ulcer or ulcer scar is absent; UL(+), ulcer or ulcer scar is present, VM1, vertical margin is involved; VM1, vertical margin involvement is unknown.

Table 4.

Non-curative factors.

Size over 20 mm	11/40	(27.5%)
expanded indication*	3/25	(12.0%)
out of indication*	8/15	(53.3%)
Depth SM/uncertain	10/40	(25.0%)
expanded indication*	9/25	(36.0%)
out of indication*	1/15	(6.7%)
Intramural ulceration	9/40	(22.5%)
expanded indication*	3/25	(12.0%)
out of indication*	6/15	(40.0%)
Capillary invasion$	4/40	(10.0%)
Margin involved/unknown	9/40	(22.5%)
HM+/X	5/40	(12.5%)
VM+/X	5/40	(12.5%)

^*Pre-ESD evaluation.

^$All cases positive for capillary invasion had an invasion depth of SM.

ESD, endoscopic submucosal dissection; HM+/X, horizontal margin involved/unknown; SM, submucosal; VM+/X, vertical margin involved/unknown.

Additional surgery was performed in 11 of the 38 patients (Table 5). Among the patients who underwent noncurative ESD, those with positive capillary invasion and a depth of invasion of SM or uncertain tended to require additional surgery. Local remnant cancer was found in 1 of the 11 patients who underwent additional surgery. None of the 11 patients had lymph node metastasis (Table 6).

Table 5.

Additional treatment after endoscopic submucosal dissection (ESD).

Total number of patients	n = 38
Additional surgery
done	11	(28.9%)
not done	27	(71.1%)
	Surgery	Observation
curative ESD	1 (6.3%)	15 (93.7%)
noncurative ESD	10 (45.5%)	12 (54.5%)

Table 6.

Pathological findings of surgical specimens.

Number of patients undergoing surgery	n = 11
Pathologically residual tumor
negative	10	(90.9%)
positive*	1	(9.1%)
Lymph node metastasis
negative	11	(100.0%)
positive	0	(0.0%)

^*Noncurative ESD (36 mm 0-IIc with pathological evidence of ulceration, HM1) at primary site.

ESD, endoscopic submucosal dissection; HM1, Horizontal margin involvement.

The long-term outcomes after ESD are summarized in Table 7. At the time of writing, all 11 patients who underwent additional surgery after ESD (surgery group) were alive after a median follow up of 1358 days (range: 547–2694). Among the 27 patients who were observed (observation group), 1 of the 12 patients who underwent noncurative ESD had recurrence 5 months after ESD. None of the 15 patients in the observation group who underwent curative ESD had recurrence after a median follow up of 1120 days (range: 477–3130). In the observation group, 1 patient who underwent curative ESD and 1 who underwent noncurative ESD died of other diseases, as shown in Figure 1.

Table 7.

Long-term outcomes after endoscopic submucosal dissection (ESD).

	Median follow-up period after ESD	Recurrence	Death from primary disease	Death from other disease
Surgery group (n = 11)	1358 (547–2694)	0	0	0
Observation group (n = 27)	750 (184–3130)	1  (3.7%)	0	2  (7.4%)
curative ESD (n = 15)	1120 (477–3130)	0	0	1$ (6.7%)
noncurative ESD (n = 12)	570 (184–1960)	1* (8.3%)	0	1‡ (8.3%)

^*Local recurrence 5 months after ESD.

^$Death 1479 days after ESD.

^‡Death from esophageal cancer 492 days after ESD.

Figure 1.

Kaplan–Meier overall survival estimate.

Group A (n = 15): group of observation without additional surgery after curative ESD.

Group B (n = 12): group of observation without additional surgery after noncurative ESD.

Group C (n = 11): group of surgery after curative or noncurative ESD.

ESD, endoscopic submucosal dissection.

Discussion

Undifferentiated-type EGCs are far less common than differentiated-type EGCs. The study group comprised only 38 patients with 40 undifferentiated-type EGC lesions treated by 39 sessions of ESD in Kanagawa Cancer Center Hospital between July 2005 and December 2013. In the same time frame, a total of 1027 sessions of ESD were performed for EGCs, indicating that the rate of undifferentiated-type EGSs treated by ESD was as few as 3.8% of total ESD-treated EGCs as shown in Figure 2. Nevertheless, the number of undifferentiated-type EGCs treated by ESD in our hospital has been increasing owing to the minimal invasiveness of ESD.

Figure 2.

Number of endoscopic submucosal dissection (ESD) sessions for early gastric cancers (EGCs) at Kanagawa Cancer Center Hospital, 2005–2013.

Overall 7.7% of the patients had post-treatment bleeding and 2.6% had perforation during ESD as complications; however, all patients recovered fully without any sequelae, despite the fact that 13 of the 38 patients were 75 years of age or older. ESD was thus confirmed to be minimally invasive, even in very elderly patients. Although the minimal invasiveness of ESD is undoubtedly attractive, the outcomes of ESD in patients with undifferentiated-type EGC remain to be fully evaluated. In the present study, we endoscopically dissected lesions that met the expanded indications for ER as well as those that did not and evaluated outcomes after ESD.

In our study, the curability of ESD was only 40%. This low rate was attributed to the fact that 15 (37.5%) of the 40 lesions were evaluated to be out of the indication range before ESD. Among the 25 lesions that met the expanded indications before treatment, endoscopic dissection was curative for 16 lesions (64.0%). In other words, 36.0% of the lesions considered to meet the expanded indications before treatment underwent pathologically noncurative resection by ESD because of inaccurate pretreatment evaluations or technical problems, such as making an incision in the tumor site. Among factors responsible for the noncurative ESD of lesions that met the expanded indications before treatment, margin involvement and an invasion depth of SM were common in our study. Our results indicate that precise evaluation of the area and depth of lesions is an important problem that remains to be solved. Moreover, the vertical margin was involved in three cases because an incision was made in the tumor site. Technical improvements are therefore also required. A total of 12 out of 22 noncuratively dissected patients were observed without subsequent treatment with gastrectomy. The reason of not receiving additional surgery was the patient’s refusal in seven cases and the attending physician’s decision not to add an operation on top of poor physical status or underlying disease in five cases.

Many studies have reported that it is difficult to determine the demarcation line of the lesion and to delineate the normal surrounding area on endoscopic examination of undifferentiated-type EGC. One of the reasons is lateral extension of cancer cells beneath a noncancerous mucosa [Doyama et al. 2013; Kakushima et al. 2011]; thus, the importance of pretreatment ‘around-the-lesion’ biopsy has been reported [Doyama et al. 2013; Hirasawa et al. 2013]. In the present study, a four-point biopsy was not performed around the lesion before ESD in three of the five cases with a positive horizontal margin, showing the importance of pretreatment around-the-lesion biopsy. Another problem is that there are some undifferentiated-type EGC lesions which are diagnosed as differentiated type by a pretreatment pathological diagnosis. We therefore recommend the following precautions: (1) ‘around-the-lesion biopsy’ such as four-point biopsy should be performed before ESD, not only for undifferentiated-type but also for differentiated-type EGC, especially in patients with depressed-type lesions, which accounts for most undifferentiated-type lesions; and (2) the endoscopic dissection should be extensive enough to ensure an adequate safety margin.

Because both local recurrence as well as residual cancer were associated with positive horizontal margins on pathological diagnosis of the ESD specimens, it is important to recognize that undifferentiated-type EGC with a tumor-positive horizontal margin after ESD carries an increased risk of recurrence. We strongly recommend additional surgery in patients who have predominantly undifferentiated-type EGC (signet ring cell carcinoma, poorly differentiated adenocarcinoma, mucinous adenocarcinoma) with positive or uncertain horizontal margins as confirmed by histopathological examination of endoscopically resected specimens.

However, there was a high incidence of multiple malignancies, as previously reported [Seo et al. 2010]. We must recognize that patients with undifferentiated-type EGC have high risk malignant disease and must therefore be carefully followed up.

Long-term outcomes of ESD for undifferentiated EGC, without additional surgery, in published papers are compared in Table 8. Although the follow-up period varies in each report, all cases with recurrence were noncurative ESD patients and the rate were 0–11.1% in those patients. None of them died of gastric cancer but metachronous gastric cancers were seen in 0–33.3%, indicating the importance of intensive follow up.

Table 8.

Comparison of long-term outcomes of endoscopic submucosal dissection (ESD) alone for undifferentiated early gastric cancer (ECG).

Reference	Number followed*	Follow-up period(months)	Recurrence rate	Metachronous EGC	Disease specific survival rate	Overall survival rate	Published
Kang et al. [2010]	17 curative	13 (2–39)	0	no description	100%	no description	March 2010
	26 noncurative		0		100%
Okada et al. [2012]	78 curative	36.0 (6–92)	0	1.3%	100%	100%	February 2012
	7 noncurative	33.0 (26–86)	0	0	100%	85.7%
Kim et al. [2013]	20 curative	34 (7–81)	0	no description	100%	95.0%	August 2013
	36 noncurative		11.1% (local)		100%	97.2%
Abe et al. [2013]	46 curative	76.4 (7–159)	0	11.4%	100%	5yr, 93.0%	September 2013
	14 noncurative		7.1% (local)	no description	100%	5yr, 82.5%
Oda et al. [2014]	56 curative	⩾5 years	0	no description	100%	89.3%	March 2014
Suzuki et al. [2015]	43 curative	5 years	0	20.9%	100%	92.9%	January 2015
Present study	15 curative	36.8 (15.7–102.9)	0	0	100%	93.3%
	12 noncurative	18.7 (6.0–64.4)	8.3% (local)	33.3%	100%	91.7%

^*Number of patients followed without additional surgery, after curative or noncurative ESD.

In conclusion, undifferentiated-type EGC with a tumor-positive horizontal margin after ESD has a high risk of recurrence. Therefore, it is important to precisely evaluate the area of the lesion and to perform extensive ER to ensure an adequate safety margin.