Outcomes for endoscopic submucosal dissection of pathologically staged T1b esophageal cancer: a multicenter study

Abstract

Background and Aims:

The outcomes of endoscopic submucosal dissection (ESD) for T1b esophageal cancer (EC) and its recurrence rates remain unclear in the West. Using a multicenter cohort, we evaluated technical outcomes and recurrence rates of ESD in the treatment of pathologically staged T1b EC.

Methods:

We included patients who underwent ESD of T1b EC at 7 academic tertiary referral centers in the United States (n = 6) and Brazil (n = 1). We analyzed demographic, procedural, and histopathologic characteristics and follow-up data. Time-to-event analysis was performed to evaluate recurrence rates.

Results:

Sixty-six patients with pathologically staged T1b EC after ESD were included in the study. A preprocedure staging EUS was available in 54 patients and was Tis/T1a in 27 patients (50%) and T1b in 27 patients (50%). En-bloc resection rate was 92.4% (61/66) and R0 resection rate was 54.5% (36/66). Forty-nine of 66 patients (74.2%) did not undergo surgery immediately after resection and went on to surveillance. Ten patients had ESD resection within the curative criteria, and no recurrences were seen in a 13-month (range, 3–18.5) follow-up period in these patients. Ten of 39 patients (25.6%) with noncurative resections had residual/recurrent disease. Of the 10 patients with noncurative resection, local recurrence alone was seen in 5 patients (12.8%) and metastatic recurrence in 5 patients (12.8%). On univariate analysis, R1 resection had a higher risk of recurrent disease (hazard ratio, 6.25; 95% confidence interval, 1.29–30.36; P = .023).

Conclusions:

EUS staging of T1b EC has poor accuracy, and a staging ESD should be considered in these patients. ESD R0 resection rates were low in T1b EC, and R1 resection was associated with recurrent disease. Patients with noncurative ESD resection of T1b EC who cannot undergo surgery should be surveyed closely, because recurrent disease was seen in 25% of these patients.

There has been a dramatic increase in the incidence of esophageal adenocarcinoma (EAC) over the past few decades. The traditional treatment for esophageal cancer (EC) includes esophagectomy with or without chemoradiation therapy (CRT).^1,2 Despite modern advancements in surgical techniques, the operative mortality of esophagectomy remains around 3.1% and continues to carry significant morbidity (33%).³ Additionally, surgery may not be a viable option for many elderly patients with comorbid disease in whom EC tends to occur.⁴

For early EC (T1a), endoscopic resection has largely replaced surgery as the main therapy and has been demonstrated to be safe and effective.^5,6 Submucosal EC (T1b) lies within a watershed area with an increased risk of lymph node metastases ranging from 7.5% to 45%, but most of these cancers may not have lymph node involvement.⁷ For nonoperative candidates with T1b EC, there are limited therapeutic options. Although CRT and brachytherapy are used, no substantial published data are available on the outcomes of these treatments for T1b EAC.

Because endoscopic submucosal dissection (ESD) has been shown to achieve complete resection with low recurrence rates in many GI tumors, there has been growing interest in the use of ESD for T1b EC.^8,9 Recent studies have suggested that ESD of early Barrett’s neoplasia may achieve higher complete remission of dysplasia rates and lower recurrence rates than EMR.^10,11 Little is known about the technical outcomes of ESD for T1b EC and its recurrence rates.¹² Thus, we chose to perform a multicenter retrospective cohort study to assess the technical outcomes and recurrence rates of ESD in the treatment of pathologically staged T1b EC.

METHODS

This was a retrospective cohort study of patients who underwent ESD for pathologically confirmed T1b EC at 7 academic tertiary referral centers in the United States (n = 6) and Brazil (n = 1) between January 1, 2013 and June 1, 2021. Institutional review board approval from each participating center was obtained. Information was gathered from prospectively managed databases at each institution. Study data were collected and managed using REDCap electronic data capture system hosted at the Cleveland Clinic.^13,14

Inclusion criteria

Patients were included in the study if they underwent an ESD resection for pathologically staged T1b EAC or esophageal squamous cell carcinoma (ESCC). Based on the European Society of Gastrointestinal Endoscopy ESD guidelines published in 2015, many proceduralists adopted their recommendations of performing ESD if T1b EC was suspected to allow for complete histopathologic analysis.

Exclusion criteria

Patients were excluded if they had a pathologic diagnosis other than T1b cancer (eg, high-grade dysplasia, T1a or T2 cancer). Patients who underwent esophagectomy immediately after ESD were included in the evaluation of technical outcomes but were excluded from the recurrence analysis.

Procedures

All procedures were performed by experienced advanced endoscopists. Radial echoendoscopes were used in all preprocedure EUS staging. The ESD procedure was performed using an endoscope with a distal attachment and CO² insufflation in all cases. The basic steps of ESD have been previously described.¹⁵ The exact ESD technique used was at the discretion of the endoscopist. Circumferential marking around the lesion was made using an ESD knife or argon plasma coagulation. A viscous lifting agent was injected to expand the submucosa. Mucosal and submucosal dissection was performed with ESD knives. Traction was also used in selected cases and achieved using a ProdiGI traction wire (ERD-TW20, ERD-TW35; Medtronic, Minneapolis, Minn, USA) or clip-line traction.^16–18 After resection, if the submucosal (SM) SM defect was larger than two-thirds of the circumference of the esophageal mucosa, in some patients, at the discretion of the endoscopist, triamcinolone was injected into the submucosa and/or a course of oral fluticasone was given for 4 to 8 weeks to prevent stricture development.¹⁹

Histopathologic analysis

Before formalin fixation, ESD specimens were minimally stretched to their preresection size and pinned to wax or a corkboard. Resected specimen size was recorded as the measured dimension of the specimen after resection. The tissue specimens were sliced into sections (2–3 mm thick) and analyzed. D2–40 staining was not routinely performed to identify lymphovascular invasion (LVI). Submucosal depth of invasion was collected when available.

Endpoints

Primary outcomes were local and metastatic recurrence. Local recurrence was defined as pathologically confirmed EC on any follow-up surveillance endoscopy at or in close proximity to the ESD scar. Metastatic recurrence was defined as radiologic or EUS appearance typical of metastatic disease or biopsy sample–proven metastases on any follow-up examination. Metastases were divided into regional lymph nodes metastases and distant metastases (cancer spread to distant organs or distant lymph nodes).

The secondary endpoint was technical efficacy in terms of en-bloc and R0 resection rates. En-bloc resection was recorded when the target lesion was resected as a single specimen. The absence of cancer at the lateral and deep margins was defined as R0 resection. Curative resection for EAC was R0 resection with the absence of LVI, absence of poorly differentiated or undifferentiated tumor, and restriction to the superficial submucosa (<500 μm for EAC). Curative resection for ESCC was R0 resection with an absence of LVI, absence of poorly differentiated or undifferentiated pathology, and restriction to the superficial mucosa (m2).^20,21 Because inclusion criteria of the study were T1b cancers, no ESCC was deemed as curative resection.

For adverse events, major early (intraprocedural) bleeding was defined as bleeding requiring premature termination of the procedure, transfusion of red blood cells, or not amenable to management by endoscopic intervention. Major delayed bleeding was defined as clinical signs of bleeding, with a hemoglobin drop > 2 g/dL, after completion of the procedure. Perforation was defined as a visible defect in the esophageal wall that allowed visualization into the mediastinum or signs of mediastinitis with postinterventional imaging proving extravasation of contrast medium. Stricture was defined as a post-treatment luminal narrowing that could not be traversed with a standard upper GI endoscope.

Follow-up

After ESD, EAC patients were recommended to have ablation therapy of any remaining Barrett’s esophagus. Because Western guidelines were not available for surveillance after ESD resection of T1b EC during the study period, the frequency of surveillance endoscopy and cross-sectional imaging depended on institutional protocols, with most first surveillances occurring 3 months after the initial procedure. Follow-up examinations included endoscopy with biopsy sampling of suspicious areas and radiologic cross-sectional imaging.

Statistical analysis

Data are presented as medians and quartiles [25th, 75th] for non-normally distributed continuous variables and frequency (percentage) for categorical variables. The nonparametric Kruskal-Wallis test was used to assess differences in continuous variables. The χ² test and Fisher exact test were used to compare categorical variables as appropriate. Unadjusted univariate binary logistic regression was used to assess the association between each clinical factor and the outcome, en-bloc resection, R0 resection, and curative resection. The unadjusted univariate analysis with a Cox proportional hazard model with a cloglog link was used to determine the association between each factor with recurrence. The Kaplan-Meier survival curve with 95% confidence intervals was conducted. Analyses were performed using R (version 3.6.2; Vienna, Austria) and SAS (SAS Institute, Cary, NC, USA) software, and P < .05 was considered statistically significant.

RESULTS

Patient and procedure characteristics

Sixty-six patients underwent ESD for pathologically staged T1b EC. Patient demographics and lesion characteristics are presented in Table 1. The number of procedures from each center was as follows: Cleveland Clinic, 19; University of Florida, 13; University of Sao Paulo, 11; Brigham and Women’s Hospital, 7; Mayo Clinic Arizona, 8; Parkview Cancer Center, 5; and University Hospitals, 3.

TABLE 1.

Summary of patient demographics and lesion characteristics (n = 66)

Characteristics	No. of cases	Value
Age at time of procedure, y	66	71.0 [65.0–76.0]
Body mass index, kg/m2	53	26.1 [22.6–30.1]
Age-adjusted Charlson comorbidity index	66	4.0 [3.0–6.0]
Gender	66
Female		12 (18.2)
Male		54 (81.8)
Ethnicity, white	49	35 (71.4)
History of smoking	54	36 (66.7)
Presence of hiatal hernia	55	31 (56.4)
Presence of esophageal varices	54	2 (3.7)
Prior endoscopic therapy	66	14 (21.2)
Previous radiofrequency ablation		11 (16.6)
Previous EMR		3 (4.5)
Location of esophageal lesion	65
Proximal part (upper third)		4 (6.2)
Middle part (middle third)		8 (12.3)
Distal part (lower third)		27 (41.5)
Gastroesophageal junction		26 (40.0)
Tumor histology	66
Adenocarcinoma		49 (74.2)
Squamous cell carcinoma		17 (25.8)
Histologic grade
G1 (well differentiated)		22 (33.3)
G2 (moderately differentiated)		31 (43.9)
G3 (poorly differentiated)		13 (19.7)
Preoperative EUS stage	54
Tis		5 (9.3)
T1a		22 (40.7)
T1b		27 (50.0)
Postoperative pathologic stage T1b	66	66 (100.0)
Lesion diameter, mm	65	25.0 [19.0–35.0]
Percentage of circumference involved by the lesion	48	33.0 [30.0–50.0]
Submucosal invasion, μm	39
<500		22 (56.4)
≥500		17 (43.6)
Lymphovascular invasion	66	22 (33.3)
Submucosal invasion, μm	39	300.0 [200.0–1100.0]

Values are median [p25-p75] or n (%).

Seventy-four percent of patients (49/66) had EAC and 25.8% (17/49) had ESCC. Twenty-one percent of patients (14/66) had prior endoscopic therapy, of which 17% (11/66) had undergone radiofrequency ablation and 5% (3/66) had undergone prior EMR. The preprocedure EUS stage was available in 81.8% of patients (54/66) and was found to be Tis in 9.3% (5/54), T1a in 40.7% (22/54), and T1b in 50% (27/54).

The median preoperative lesion diameter was 25.0 mm [range 19.0–35.0]. The median procedure time for ESD was 91 minutes [66.0–127.0] (Table 2). The median follow-up time for all patients was 20 months [12–23]. Submucosal depth of invasion was reported in 59% of patients (39/66). Median SM depth of invasion was 300 μm [200.0–1100.0], and 43.6% of patients (17/39) had ≥500 μm invasion and 56.4% (22/39) had <500 μm invasion. En-bloc resection rate was 92.4% (61/66), R0 resection rate was 54.5% (36/66), and curative resection rate was 15.2% (10/56) (Table 2). During the follow-up period, all-cause mortality was 13.6% (9/66). No deaths were related to the ESD procedure.

TABLE 2.

Summary of procedural characteristics (n = 66)

Characteristics	No. of cases	Value
Type of anesthesia	66
General anesthesia		56 (84.8)
Monitored anesthesia care		10 (15.2)
En-bloc resection		61 (92.4)
R0 resection		36 (54.5)
Curative resection		10 (15.2)
Steroid use	66
Injected steroid only		5 (7.6)
Swallowed steroid only		3 (4.5)
Both injected and swallowed steroid		1 (1.5)
Perforation	66	2 (3.0)
Major delayed bleeding	66	2 (3.0)
Stricture development	66	7 (10.6)
Procedure duration, min	59	91.0 [66.0–127.0]
Length of hospital stay, days	66	.00 [.00–1.00]

Values are median [p25-p75] or n (%).

Patients who immediately went on to surgery

Twenty-six percent of patients (17/66) underwent esophagectomy immediately after ESD for noncurative resection (Fig. 1). Seventy-one percent of these patients (12/17) underwent surgery alone, and 29.4% of patients (5/17) received surgery with chemotherapy and/or radiation therapy. Sixty-five percent of patients (11/17) had positive margins (R1 resection), 44% (4/9) had SM invasion ≥500 μm, 24% (4/10) had poorly differentiated tumors, and 59% of patients (10/17) had LVI on post-ESD pathology. Postesophagectomy pathology reports were available for review in only 52.9% of patients (9/17). Of these 9 of patients, none had positive lymph nodes at the time of surgery. One patient with long-segment Barrett’s esophagus (C6M8) had metachronous cancer discovered on esophagectomy specimen, and 1 patient with R1 ESD resection had 2 mm of residual tumor in the submucosa noted on the esophagectomy specimen.

Figure 1.

Flowchart diagram of the study. EC, Esophageal cancer; ESD, endoscopic submucosal dissection.

Patients who went on to surveillance

Seventy-four percent of patients (49/66) did not undergo surgery immediately after resection and went on to surveillance. Median patient age was 71 years [65–76], and 75.5% were men with a median tumor size of 26 mm [20–35]. LVI was present in 24.5% of patients (12/49), and 16.3% (8/49) had poorly differentiated tumors. Median depth of SM invasion was 305 μm [200–1100], and the en-bloc resection rate was 98% and R1 resection rate 38.8% (19/49). Twenty percent (10/49) had curative resections. Of the 10 patients with curative resection, 5 had documented SM invasion depth and 5 had a description of minimal or minute submucosal invasion without reporting micrometer depth of invasion and were deemed as within curative criteria by the study center. No patient with curative resection received adjuvant therapy, and no recurrence was seen in these patients in a median follow-up of 13.5 months [3–18.5] (Table 3).

TABLE 3.

Follow-up and recurrence

	Total
Factor	No. of cases	Value
Recurrent disease	49	10 (20.4)
Follow-up time to recurrence, mo		31 [23–26]
Follow-up time, mo		20 [21–23]
Recurrence on first surveillance endoscopy/imaging	10	5 (50)
Recurrence in curative resection patients	10	0 (0)
Recurrence in noncurative resection patients	39	10 (25.6)
Local recurrence alone	49	5 (10.2)
Metastatic recurrence	49	5 (10.2)
Regional lymph node metastases only		2 (4.0)
Distant metastases		3 (6.1)
All-cause mortality	66	9 (13.6)

Values are median [p25-p75] or n (%).

Of 39 patients with noncurative resections (EAC, 64.1% [25/39]; ESCC, 35.9% [14/39]), median age was 70 years [63–74], 74.4% were men, and median tumor size was 30 mm [20–37.50]. Median depth of SM invasion was 310 μm [175–1350]. Of these 39 patients, 19 (48.7%) had R1 resection, 12 (30.8%) had LVI, 8 (20.5%) had poorly differentiated tumors, and 12 (41.3%) had SM invasion >500 μm. Of the 39 patients with noncurative resections who went on to surveillance with a median follow-up of 17 months [9–24], 44% (17/39) received adjuvant CRT and/or radiation therapy, 3% (1/39) received chemotherapy alone, 15% (6/39) received radiation alone, and 33% (10/39) received combined CRT. Twenty-six percent of patients (10/39) with noncurative resections developed recurrence (EAC, 50% [5/10]; ESCC, 50% [5/10]). Five of these patients developed recurrence on the first surveillance endoscopy/imaging.

The median time to recurrence was 13.5 months [6–23]. Local recurrence alone was seen in 12.8% of patients (5/39) and metastatic recurrence also in 12.8% (5/39). Of the 5 patients who developed metastatic disease, 40% (2/5) had metastases to regional lymph nodes alone and 60% (3/5) had distant metastases. Of the 5 patients with local recurrence, 2 received CRT, 2 received additional radiation alone, and 1 died before receiving additional treatment. Three of 5 patients with local recurrence had cross-sectional imaging, which confirmed the disease was limited to the esophagus. Table 4 presents the clinical characteristics and pathology of patients who went on to develop metastatic recurrence.

TABLE 4.

Pathology and clinical characteristics of patients who went on to develop metastatic recurrence

Age (y)	Gender	Esophageal cancer type	R0	High-risk pathologic features	Depth of submucosal invasion (μm)	Adjuvant therapy	Type of recurrence	Time to recurrence (mo)
74	Female	EAC	No	VM+, LVI+, poorly differentiated+	N/A	Radiation therapy	Distant metastases (brain, liver)	17
66	Male	EAC	No	VM+, LVI−, poorly differentiated+	1200	Chemotherapy	Distant metastases (liver, peritoneal)	13
57	Male	EAC	No	VM+, LVI−, moderately differentiated−	2000	CRT	Lymph nodes	4
81	Male	EAC	No	VM+, LVI−, moderately differentiated−	N/A	CRT	Lymph nodes	14
82	Male	EAC	No	VM+, LVI−, moderately differentiated−	2000	Surgery (liver metastases found at time of surgery)	Distant metastases (liver)	2

LVI, Lymphovascular invasion; EAC, esophageal adenocarcinoma; CRT, chemoradiation therapy; VM, vertical margin; N/A, not applicable.

Adverse events

Fourteen percent of patients (9/66) developed adverse events. No patients had early bleeding, 3% (2/66) had delayed bleeding, 3% (2/66) had perforations, and 11% (7/66) developed strictures during the follow-up period. Both patients with delayed bleeding were admitted to the intensive care unit and received blood transfusions and endoscopic intervention. Injected or oral steroids were used in 13.6% of patients (9/66) who underwent ESD to prevent strictures.

DISCUSSION

The main finding of our study was that the R0 resection rate for ESD of T1b EC was lower (54.5%) than that reported in contemporary series of ESD for T1a EC. We found that R1 resection of T1b EC was associated with residual/recurrent disease (hazard ratio, 6.25; 95% confidence interval, 1.29–30.36; P = .023). In our series, no local or metastatic recurrence occurred after ESD resection of T1b EAC within the curative criteria. In patients with noncurative resection of T1b EC, irrespective of adjuvant therapy, residual/recurrent disease was seen in 25% of patients.

In our study, the rate of R0 resection was 54.5% (36/66), which is lower than that reported in contemporary ESD series on high-grade dysplasia and T1a EC (73%–88%).^10,22–24 This lower R0 resection rate is consistent with a recent study on noncurative esophageal ESD where only 41.1% of patients (7/17) with T1b EAC had R0 resection.^12,25 The challenge of obtaining negative margins in T1b EC is also seen with EMR, as in a study investigating EMR of T1b EAC positive deep margins were seen in 57% of patients (42/73).²⁶ The difficulty in obtaining negative deep margins in T1b EC is likely because the dissection plane would have to be deep enough to achieve a border of normal submucosa beneath the submucosal extension of the tumor. In addition, submucosal tumors are associated with submucosal fibrosis, making this even more challenging to achieve.^27,28

On univariate analysis, R1 resection had a higher risk of residual/recurrent disease (local and metastatic disease) (hazard ratio, 6.25; 95% confidence interval, 1.29–30.36; P = .023) (Table 5). A recent European study showed similar results with a 29% (10/34) residual/recurrence rate in patients undergoing R1 resection of EAC.^25,29 This reinforces the principle that good oncologic resection with negative margins should be the goal of ESD. Achieving R0 resection appears to be more challenging in T1b esophageal tumors than superficial lesions, and efforts should be made to improve this. We did notice on a recent case done with a traction wire that R0 resection was achieved despite SM invasion of 3000 μm. Our study was not powered or designed to look at the effect of traction devices on R0 resection, but further studies should be done to evaluate this. Retraction methods (eg, endoscopic submucosal tunneling) and traction devices (eg, clip-line traction and traction wire) may allow for better exposure of the deep submucosa and associated fibrosis and should be evaluated to see if they can help achieve higher R0 resection rates in T1b EC (Fig. 2).

TABLE 5.

Univariate analysis with discrete time hazard model on recurrence

Factor	Hazard ratio (95% confidence interval)	P value
Submucosal invasion, ≥500 μm vs <500 μm	NA	NA
Lymphovascular invasion, yes vs no	.95 (.12–7.76)	.96
Histologic grade G2 (moderately differentiated) vs G1 (well differentiated)	3.386 (.33–34.49)	.30
Histologic grade G3 (poorly differentiated) vs G1 (well differentiated)	5.505 (.43–70.66)	.19
Did the patient receive the recommended follow-up radiofrequency ablation? yes vs no	NA	NA
Preoperative maximal diameter of the lesion (1-unit increment)	.97 (.92–1.03)	.31
Maximal length of largest specimen (mm) (1-unit increment)	.97 (.93–1.02)	.215
Histology of resected specimen, squamous cell carcinoma vs adenocarcinoma	.654 (.151–2.835)	.57
R0 resection, yes vs no	.16 (.03–.78)	.023
R1 resection, yes vs no	6.253 (1.29–30.36)	.023
Adjuvant chemotherapy, yes vs no	2.475 (.657–9.323)	.181
Adjuvant radiotherapy, yes vs no	1.576 (.422–5.887)	.499
Adjuvant chemotherapy and/or radiotherapy, yes vs no	2.326 (.581–9.315]	.233

NA, not applicable.

Figure 2.

A and B, Endoscopic submucosal tunnel dissection of a tumor in the gastroesophageal junction showing en-bloc (C) and R0 resection (D) with 3689 μm depth of submucosal invasion.

In our study, no patient with ESD resection of T1b EAC within the curative criteria developed local or metastatic recurrence. This supports the current American Gastroenterological Association and European Society of Gastrointestinal Endoscopy guidelines of including <500 μm SM invasion within the curative criteria.^20,21

For patients with noncurative resection of T1b EC, 25.6% (10/39) developed recurrence. Recurrences occurred in patients who received chemotherapy and radiation after ESD, so adjuvant CRT did not confer total prevention of disease recurrence. Larger, prospective studies are needed to see if adjuvant CRT reduces the risk of recurrent noncurative ESD resection T1b EC. The recurrences that occurred after noncurative resection of EC were local in 50% of patients (5/10) but, more alarmingly, were metastatic in 50% of patients (5/10) (Table 4). An 82-year old man developed metastatic recurrence 2 months after ESD. He did have cross-sectional imaging (positron emission tomography-CT) before ESD that did not detect any metastatic disease; however, given the short time frame to metastatic recurrence, we suspect there may have been undetected metastatic disease at the time of ESD. This risk for the metastatic disease after ESD resection of T1b EC supports the use of cross-sectional imaging in surveillance, as suggested in the recent American Gastroenterological Association update on surveillance after pathologically curative ESD of early GI neoplasia.³⁰

Although the recurrence outcomes of noncurative ESD of T1b cancers are suboptimal, select patients with suspected T1bN0 tumors are likely still appropriate candidates for ESD resection. EUS is known to have poor accuracy in differentiating T1a and T1b tumors. In a series on gastroesophageal junction tumors, 50% of patients with pathologic T1a tumors had EUS staging of T1b or higher.³¹ Consistent with this, in this study, 50% of patients in this study with T1b EC had a preprocedure EUS stage of Tis or T1a. These data suggest that ESD, similar to EMR, can play an important role in staging in EC, with further therapy based on resected pathology.¹⁰

In addition, patients with T1b EC who are nonsurgical candidates have limited therapeutic options. Although CRT and brachytherapy are used, little data are available on outcomes for these therapies in T1b EAC, and future studies are needed to better understand these options (Table 6). Of patients with noncurative ESD resection of T1b EAC who did not undergo surgery, 74.3% (29/39) had no recurrence during the study period. Adjuvant CRT is often used to reduce recurrence risk in this setting, although the data are limited because nearly all available studies were conducted in Japan in ESCC, not EAC.³²

TABLE 6.

Available literature on ESD of T1b EC

	Current report	Van Munster et al (2021)25	Othman et al (2021)12	Namikawa et al (2021)33	Hsu et al (2021)34	Al-Kaabi et al (2021)35	Zhang et al (2019)36
Location	USA and Brazil	Netherlands	USA	Japan	Taiwan	Europe	China
Patient diagnosis	T1b EAC and ESCC	BE, T1a, T1b EAC	T1a and T1b EAC	T1a and T1b submucosa ESCC	High-grade dysplasia and T1a and T1b ESCC	T1a, T1b, T2 ESCC undergoing ESD after chemoradiation therapy	T1a, T1b ESCC
Study design	Multicenter retrospective	Multicenter retrospective study	Single-center retrospective study	Single-site retrospective	Single-site retrospective	Multicenter retrospective	Single-site retrospective cohort study
No. of T1b patients	66	67	17	10	40	8	82
Median size, mm [p25-p75]	25.0 [19.0–35.0].	N/A*	15* [13.5–20]	24.5* [3–70]	32.8 ± 16.9*,† (10–100)	N/A*	26* [20–36]
En-bloc resection rate, %	92.4	99	100	N/A*	N/A*	N/A*	N/A*
R0 resection rate, %	54.5	50	41.2	N/A*	N/A*	N/A*	N/A*
Overall survival	86.36% (20-mo follow-up)	N/A*	N/A*	N/A*	Calculated 10-y survival >85% for >sm1 with adjuvant therapy	N/A*	90.9% (21-mo follow-up)

EAC, Esophageal adenocarcinoma; BE, Barrett’s esophagus; ESCC, esophageal squamous cell carcinoma; ESD, endoscopic submucosal dissection; N/A, not applicable.

^*Results included entire study population including T1a and T1b outcomes and did not separately present outcomes for T1b.

^†This value is mean ± standard deviation (range).

This study has several strengths. The multicenter population included diverse geographic locations, making our findings more generalizable to ESD practices throughout the West. Our data focused on both short-term technical outcomes of the procedure and longer-term outcomes to identify both local and metastatic recurrence.

There are some limitations to our study. The overall patient population was small, but it was a pure group of pathologic T1b cancers. ESD for T1b cancers is uncommon, requiring the combination of multicenter data to obtain the sample size for this study. The findings of our study help give guidance to current clinical practice and help in the design of future prospective studies. Its retrospective design can result in the use of various techniques and different follow-up strategies. Given the retrospective nature of the study and the definition of local recurrence including tumor in close proximity to the ESD scar, there is a chance a metachronous lesion in close proximity to the ESD scar would be included as local recurrence. Depth of submucosal invasion was not reported at all centers routinely and thus was not available in all cases. Given the difference in the outcomes of curative ESD resection and noncurative ESD resection, we hope practices change and submucosal depth of invasion is more routinely measured and reported. Also, if patients did not undergo esophagectomy at the center performing ESD, which is not uncommon given the referral nature of ESD, we did not have lymph node status on these patients at the time of surgery.

In conclusion, our study shows ESD R0 resection rates are low in T1b EC, and R1 resection is associated with recurrence. Patients undergoing curative ESD of T1b EC <500 μm experience good outcomes. Conversely, both local and metastatic recurrence are seen after noncurative ESD resection of T1b EC, and patients who cannot undergo surgery should be closely surveyed both endoscopically and with cross-sectional imaging. Further studies are needed to see if retraction methods and devices can improve R0 resection in T1b EC and what role adjuvant CRT may have in reducing recurrences

Abbreviations:

CRT

chemoradiation therapy

EAC

esophageal adenocarcinoma

EC

esophageal cancer

ESCC

esophageal squamous cell carcinoma

ESD

endoscopic submucosal dissection

LVI

lymphovascular invasion