Endoscopic treatment outcome of oesophageal gastrointestinal stromal tumours

Abstract

BACKGROUND

There are few clinicopathologic characteristics and clinical results for oesophageal gastrointestinal stromal tumours (GISTs). Thus, the objective of this study was to identify the clinicopathologic characteristics and clinical results of oesophageal GISTs.

AIM

To investigate endoscopic treatment effective of oesophageal GISTs.

METHODS

It was retrospective research that collected 32 patients with oesophageal GISTs treated by endoscopic resection (ER) between January 2012 and January 2023 in two Hospital. Clinicopathologic, endoscopic records, and follow-up data were collected and analysed.

RESULTS

Thirty-one patients underwent en bloc resection and 24 (75.0%) lesions underwent R0 resection. The size of GISTs was 2.12 ± 1.88 cm. The overall complication rate was 25.0%, including hydrothorax and post-endoscopic submucosal dissection electrocoagulation syndrome. The mean mitotic index was 3.34 ± 5.04 (median, 1.50; range, 1.00-4.00). Eighteen (56.3%), 6 (18.8%), 2 (6.3%), and 6 (18.8%) patients were identified as very low, low, intermediate, and high risk, respectively. Three patients developed recurrence after a median follow-up of 64.69 ± 33.13 months. The 5-year overall survival rate was 100%, and the disease-free survival rate was 90.6%.

CONCLUSION

ER is safe and effective for patients with low-risk oesophageal GISTs. Early detection of oesophageal GISTs is essential to achieve a favourable prognosis.

Key Words: Gastrointestinal stromal tumours; Oesophageal; Endoscopic resection; Treatment outcome; Survival

Core Tip: Oesophageal gastrointestinal stromal tumours are extremely rare. Endoscopic resection proves to be both safe and effective for patients with low-risk. Early detection plays a critical role in determining the prognosis of these tumours.

INTRODUCTION

Gastrointestinal stromal tumours (GISTs) are most commonly found in the stomach, whereas oesophageal GISTs are relatively rare[1]. Given the rarity of oesophageal GISTs, clinicopathologic data pertaining to oesophageal GISTs are restricted and only few case reports are available[2].

Indeed, optimal surgical management recommendations are lacking due to the rarity of oesophageal GISTs[3]. It has been reported that esophagectomy is advised for GISTs larger than 9 cm, while enucleation of oesophageal GISTs is acceptable for smaller tumours (2-5 cm)[2]. The clinical results of these two operations have been identified to be comparable assuming suitable patient selection[3-7]. With the development of endoscopic resection (ER), complete removal can now be achieved via endoscopy[8], but the efficacy still requires further research for confirmation.

The outcomes of GIST patients vary depending on tumour size, tumour location, and mitotic activity[9]. Many GIST risk stratification criteria have been developed based on these factors to assess the prognosis of GISTs[10-13]. However, the efficiency of these models in predicting the prognosis of oesophageal GISTs was unclear, as few oesophageal GISTs were included in their development[3]. Feng et al[9] reported that oesophageal GISTs differ greatly from gastric GIST in terms of clinicopathologic features.

The number of published reports involving ER of oesophageal GISTs have increased in the past decade; however, these studies were generally based on small sample sizes with limited long-term follow-up data[14,15]. The objective of our retrospective study was to investigate the effective of ER for oesophageal GISTs on oncologic outcomes using multidisciplinary treatment.

MATERIALS AND METHODS

Patients

We collected patients with localized, primary oesophageal GISTs that could be verified histologically. Patients was from Zhongshan Hospital Fudan University and Xuhui District Central Hospital from January 2012 to January 2023. The Committee of Medical Ethics (Zhongshan Hospital Fudan University) approved this study. The criteria for inclusion were as follows: (1) > 18 years of age; (2) Pathologic diagnosis of oesophageal (E-GIST); and (3) No recurrent or metastasis before treatment. The criteria for exclusion were as follows: (1) Patients who were lost to follow-up; and (2) Existence of other malignant tumours.

Finally, thirty-two patients who had regular follow-up were enrolled.

Endoscopic procedures

Before the surgery, all patients underwent a 24-hour fast. The patient was under general anaesthesia to the ER. The procedure was conducted using a standard, single-channel gastroscope (GIF-Q260J; Olympus Medical Systems Corp., Tokyo, Japan). The tip of the endoscope was capped with a transparent cap (D-201-11802; Olympus). Firstly, the mixed normal saline was injected at 3-5 cm away from oral orientation of GIST to create a fluid cushion. Then, hook or hybrid knife was used to create a 2 cm longitudinal mucosal incision at the top of the fluid cushion, entering into the submucosa. A submucosal tunnel was carefully created, and the tumour was resected within the tunnel as previously described (Figure 1)[16].

Figure 1 Processes of Submucosal tunneling endoscopic resection for oesophageal gastrointestinal stromal tumours. A: Endoscopic view of submucosal tumours; B: The tumour could be observed directly in submucosal tunnel; C: The submucosal tumour dissected endoscopically; D: Endoscopic view of the submucosal tunnel after the tumour was removed; E: The mucosal entry closed using endoscopic clips; F: The tumour.

Pathology

Every oesophageal GIST specimen was routinely subjected to a histologic investigation in our hospital's pathology department. Using the modified National Institutes of Health (NIH) criteria to evaluate the malignancy risk of GIST by calculating maximum tumour size and the number of mitoses per 50 high-power fields (HPF)[12]. To support the diagnosis of a GIST, immunohistochemistry was carried by using a set of antibodies, such as against CD34, CD117 (c-Kit), and so on.

A multidisciplinary team, including endoscopist, pathologist, and surgeon, was assembled to develop a treatment strategy (imatinib adjuvant therapy, scheduled follow-up evaluation, or surgery) for patients with positive surgical margins. Scheduled follow-up was arranged for patients who were not receiving imatinib adjuvant therapy.

Main outcomes were identified as en bloc resection, R0 resection, recurrence, and residual. En bloc resection was termed as the tumour resected in one piece. R0, R1, and R2 resection were termed as negative lateral and vertical margins, microscopic residual tumour, and macroscopic residual tumour basing on en bloc resection, respectively[17]. ER0 was termed as no residual tumour under endoscopic observation, and ER1 was termed as having residual tumour under endoscopic observation[18].

Data collection

Clinicopathologic data and endoscopic outcomes were collected from electronic medical records in our hospital. Tumour size was reported by pathologist measuring the largest diameter of the tumour specimen. The GISTs were classified into high-risk, intermediate, low, and very low groups according to the modified NIH criteria[12]. Due to nature of retrospective collection, completeness of data is limited.

Patients had scheduled follow-up at 1, 3, 6, and 12 months after resection with endoscopy and/or endoscopic ultrasounds to monitor tumour residual or recurrence. The updated of follow-up was in January 2023.

Statistical analysis

The statistical analysis was performed by SPSS 21.0 program (IBM Corp., Armonk, NY, United States). We used Fisher's exact test or χ² test to compare classified variables. We used analysis of variance or Student’s t-test to calculate continuous variables. Kaplan-Meier survival curve was drawn by GraphPad Prism.

RESULTS

Clinicopathologic and demographic features of the patients

Table 1 provided a detailed description of the clinical features. Thirty-two patients with oesophageal GISTs were identified and included. Oesophageal GISTs were more common in men (56.3%) and the median age was 57.78 ± 7.39 years. Of the 32 patients, 87.5% were > 50 years of age. The majority of oesophageal GISTs were detected incidentally [28 (87.5%)], followed by abdominal pain [3 (9.4%)] and obstruction [1 (3.2%)]. Oesophageal GISTs were most frequently located in the lower oesophagus [20 (62.5%)], middle oesophagus [10 (31.3%)], and upper oesophagus [2 (6.3%)].

Table 1 Demographic and clinicopathological data of 32 patients with oesophageal gastrointestinal stromal tumour, n (%).

Parameter	Variable
Age, mean ± SD (year)	57.78 ± 7.39
< 50	4 (12.5)
≥ 50	28 (87.5)
Sex
Male	18 (56.3)
Female	14 (43.8)
Localization
Upper	2 (6.3)
Middle	10 (31.3)
Lower	20 (62.5)
Tumour size (cm)	2.12 ± 1.88 (median: 1.50, range: 0.80-2.75)
Mitotic rate
Mean (per 50 HPF, ± SD)	3.34 ± 5.04
Median (per 50 HPF, range)	1.50 (1.00-4.00)
NIH risk category
High	6 (18.8)
Intermediate	2 (6.3)
Low	6 (18.8)
Very Low	18 (56.3)
Symptoms
Asymptomatic (discovered incidentally)	28 (87.5)
Abdominal pain	3 (9.4)
Obstruction	1 (3.1)
Operation time, mean (± SD), minute	52.63 ± 43.29
Hospital stay duration, mean (± SD), day	4.22 ± 2.21
Surgical resection
En bloc resection	31 (96.9)
ER0 resection	32 (100)
R0 resection	24 (75.0)
R1 resection	8 (25.0)
Tumour growth pattern
Intraluminal growth	29 (90.6)
Extraluminal growth	3 (9.4)
Complication	8 (25.0)
Hydrothorax	4 (12.5)
PEECS	5 (15.6)

Tumour size was showed as mean ± SD and Median with quartiles. NIH: The National Institutes of Health; PEECS: Post-endoscopic submucosal dissection electrocoagulation syndrome; HPF: High-power field; ER: Endoscopic resection.

In total, thirty-one (96.9%) patients underwent an en bloc resection, of whom 32 (100%) and 24 (75%) underwent ER0 and R0 resections, respectively. The mean operative time was 52.63 ± 43.29 minutes; The mean hospital stay duration was 4.22 ± 2.21 days. Overall, eight (25.0%) patients had complications, including 4 (12.5%) patients with hydrothoraces and 5 (15.6%) patients with post-endoscopic submucosal dissection electrocoagulation syndrome (PEECS). All patients with complications recovered fully following conservative treatment.

The mean tumour size was 2.12 ± 1.88 (median: 1.50, range: 0.80-2.75) cm. The mean mitotic index was 3.34 ± 5.04 (median, 1.50; range, 1.00-4.00). According to the modified NIH criteria, 6 (18.8%), 2 (6.3%), 6 (18.8%), and 18 (56.3%) patients were identified as high, intermediate, low, and very low risk, respectively. Two patients received imatinib therapy.

Survival

Oesophageal GIST survival data were analysed and were presented in Table 2. Three patients had recurrences. The mean follow-up was 64.69 ± 33.13 months (median, 65.50; range, 37.75-93.00). Table 3 summarized the detailed features of patients who experienced recurrences. The 5-year overall survival (OS) and disease-free survival (DFS) rate were 100% (32/32) and 90.6% (29/32), respectively. The survival curves were showed in Figure 2.

Figure 2 Kaplan-Meier analysis of overall survival and disease-free survival of patients treated endoscopic resection. A: Kaplan-Meier analysis of overall survival; B: Kaplan-Meier analysis of disease-free survival.

Table 2 Survival data of 32 cases of oesophageal gastrointestinal stromal tumour, n (%).

Parameter	Variable
Treatment
Imatinib	2 (6.3)
Wait and see	30 (93.7)
Follow-up time
Mean (mean ± SD)	64.69 ± 33.13
Median (month, range)	65.50 (37.75-93.00)
Survival data
Recurrence	3
GISTs-related deaths	0
Survival rates (%)
1-/3-/5-year OS	100/100/100
1-/3-/5-year DFS	100/90.6/90.6

GISTs: Gastrointestinal stromal tumours; OS: Overall survival; DFS: Disease-free survival.

Table 3 Tumour characteristics of recurrence.

Patient	Location	Size, cm	Mitotic rates	Growth pattern	Margin status	Classification of tumour	Connection to MP layer	Follow-up time, months	Imatinib time, months	Reoperation	Survive
1	Middle	7	1	Extraluminal	R1	High risk	Yes	33	No	No	Yes
2	Lower	2	0	Intraluminal	R0	Very low risk	Yes	107	No	No	Yes
3	Lower	1	7	Intraluminal	R0	Low risk	Yes	40	No	No	Yes

MP: Muscularis propria.

DISCUSSION

Oesophageal GIST was a very rare subtype of GIST with little demographic and clinicopathologic characteristic[19], although there are some studies that have focused on oesophageal GISTs based on European and American populations[20,21]. Therefore, we evaluated demographic and clinical outcome of 32 oesophageal GIST patients.

In our study, the median duration of follow-up was 64.69 ± 33.13 months. Overall, three patients had recurrences. The 5-year OS and DFS was 100% and 90.6%, respectively. The mean size of oesophageal GISTs was 2.12 ± 1.88 cm. The lower oesophagus was the most frequently location, followed by the middle oesophagus, then upper oesophagus. It was consistent with the aforementioned research[21]. This may be attributed to the concentration of Cajal cells in the lower oesophagus, as GISTs originated from these cells.

Many researches demonstrated that oesophageal GIST had a higher propensity for malignancy, with an increased risk of metastases and/or tumour recurrence, as well as a high mortality rate[6,9,19]. In contrast, our study showed that patients with oesophageal GISTs can also have a good prognosis. Specifically, the 5-year OS and DFS rates were 100% and 90.6%, respectively. The mean follow-up was 64.69 ± 33.13 months (median, 65.50; range, 37.75-93.00), and there were no oesophageal GIST-related mortalities. Poor outcomes in patients with oesophageal GIST were likely linked to much greater mitotic rates and larger tumour size. It was identified that mitotic index > 5/50 HPFs and larger tumour size were related to poorer prognosis[9]. Briggler et al[22] reported that metastasis was linked to a mitotic count > 5/50 HPFs in the Mayo Clinic cohort and to larger tumour size in the National Cancer Data Base. The four Mayo Clinic patients with low-risk without imatinib treatment or surgical resection had favorable outcome. Another study also showed that the 5-year OS rate of patients with the lesions size ≤ 5 cm on computed tomography scan was 92%, 90% of them undergoing enucleation[23]. In the current study the mean tumour size was 2.12 ± 1.88 cm. The mean mitotic index was 3.34 ± 5.04 (median, 1.50; range, 1.00-4.00). The most common risk criteria were very low and low risk. These findings suggested that patients having minor tumour sizes and limited mitotic referencing to low risk may have a good prognosis.

We are of the opinion that oesophageal GISTs were shown to be malignant in other studies because oesophageal GISTs were diagnosed late rather than inherent malignancy of the tumour. The oesophagus GIST and g-GIST shared a common basic pathophysiology, which oesophageal GISTs also develop from interstitial pacemaker cells of Cajal. It has been reported that the distribution of mitotic indices between oesophageal and g-GISTs are similar[9,24]. According to previous studies, 75% of patients with oesophageal GISTs present with bleeding (10%), weight loss (20%), and dysphagia (51%)[19]. Prior to the onset of symptoms, oesophageal GISTs grow steadily over time. One case was reported in which the oesophageal GIST grow from 1.8 cm to 10 cm within 2 years, and the oesophageal GIST volume doubling time was assessed to be 3.1 months[25]. When patients present with symptoms, oesophageal GISTs are often large with increased malignant potential. In our study, 87.5% GISTs were detected incidentally and the mean size was 2.12 ± 1.88 cm, which is significantly smaller than reported in the literature[21]. This finding also accounts, at least in part, for the good prognosis in our oesophageal GIST patients. This finding clearly reflects the importance of early diagnosis and treatment.

It was accepted that oesophageal GISTs, like other GISTs, should be managed by surgery[23]; however, the optimal surgical method has not been established[14]. In our research, the total complication rate was 25%, and all patients recovered with conservative treatment. The complication was hydrothorax and PEECS, which are non-fatal complications (Clavien-Dindo < 3b). As for esophagectomy, it was reported Clavien-Dindo ≥ 3b was 10.3%[21]. They also need to take reconstruction. Another biggest advantage of ER is that it preserves the structural integrity of the esophagus, allowing patients to have a higher quality of life. What’s more, the recurrence rate after ER is 9.4% which was lower than the surgery (16.5%)[26]. The generally accepted principle for surgical resection of GIST was that it should be completely resected with negative margins[27]. However, recent studies have suggested that R1 resection may also be considered feasible. A multicenter retrospective study[21] on oesophageal GISTs found that 83.3% of patients who underwent enucleation had R1 resection, and the study indicated that R1 resection was not a prognostic risk factor. Even the recurrence rate of enucleation group was lower than surgical group. Some studies also have found that R1 resection with gastric GIST may not have significant effects on recurrence[28,29]. Our R0 resection rate 75.0%, respectively. The recurrence rate of R1 resection was 12.5% which was lower than the other published oesophageal GIST (18.2%-25.5%)[9,19,21]. We hypothesized that endoscopic R1 resection maybe not the main factor for oesophageal GIST recurrence. We are of the opinion that the margin of endoscopic R1 resection was different from surgical R1 resection. Tumour microstructures would be more clearly identified by endoscopy than our unaided eye, which allow endoscopists to remove all visible tumour without residue. And thermal damage to the pseudo-capsule frequently resulted in an R1 margin. Additionally, it could destroy the tumour cells when tumour edge was cauterized. We found that all ER succeed in ER0 resection, having a 100% 5-year OS. As for the reason for R1 resection, we found that the size of oesophageal GIST with R1 was 4.20 ± 2.50 (median: 4.00, range: 2.00-6.00) cm, which was bigger than the total size of oesophageal GISTs. It was proved that large tumour size was associated with R1 resection[30]. There was a minor submucosal room in the operating space and a narrow esophageal lumen, which became increasingly difficult to operate when the tumour was large. GISTs are often fragile increasing the difficulty of ER. It was proved that neoadjuvant imatinib treatment effectively reduced the size of GIST to facilitate R0 resection[14]. It may be the direction of ER of larger oesophageal GIST in the future.

Enucleation was the preferred approach for treating smaller E-GISTs (2-5 cm)[31,32]. Compared to surgical operation, ER focuses on minimizing invasiveness and promoting faster recovery[33]. The selection of the appropriate endoscopic technique was an important topic worthy of exploration (some ER for E-GIST was listed in Supplementary Table 1). We recommended submucosal tunneling ER (STER) as the preferred ER strategy. STER facilitated complete removal of the lesion in the esophagus while preserving the integrity of the esophageal mucosa, effectively reducing postoperative strictures and gas-related complications[34]. Due to the limitations of the tunnel, STER is suitable for the resection of GISTs smaller than 4 cm[32]. Endoscopic full-thickness resection (EFTR) was employed to resect lesions originating from the deep MP, especially large lesions that cannot be excised by STER due to limited tunneling space.

There were also some limitations. Firstly, it was a two-institution design, which restrict the generalizability of findings to a broader population. The sample size and patient demographics may not fully represent the diversity of patients seen in population-based studies. The results may be influenced by institutional factors such as specific surgical techniques, expertise, and patient selection, which may not be applicable to other settings. To address these limitations, future researches should adopt a multi-center design to increase the diversity of patient populations and enhance the generalizability of the findings. Future research could also incorporate a prospective design to minimize biases and enhance the accuracy of data collection. Secondly, the number of oesophageal GIST sample was insufficient, resulting in sampling errors. Thirdly, there was a lack of data on surgical treatment of oesophageal GIST, and it was unable to establish the appropriate criteria for surgical or ER.

CONCLUSION

ER is safe and effective in patients with low-risk oesophageal GISTs. Early detection of oesophageal GISTs is crucial for prognosis.