Esophageal GCTs are the most common GCTs in the gastrointestinal tract, accounting for approximately 2–3% of all GCTs [3]. It can occur at any age, but it is more likely to occur in 40–60 years old [7]. Our study reconfirmed this result. Previous studies showed that the frequency of esophageal GCTs was higher in females than in males, which is not consistent with a female/male ratio of 1:1.2 (10/12) in our study [7]. Most tumors are solitary and located in the middle and lower esophagus [1, 16]. All cases in our study were solitary, and 86.4% (19/22) of lesions were located in the middle and lower esophagus. However, esophageal GCTs can appear as multiple tumors, and multi-organs simultaneously or asynchronously [15, 17].
Various methods have been used to diagnose esophageal GCTs, but preoperative diagnosis is still difficult. In our study, of the 12 patients who underwent CT examination, only 1 patient (8.3%) was found with esophageal thickening, while the rest were normal. The typical endoscopic appearance of esophageal GCTs shows a white-to-yellow, smooth-surfaced lesion with normal overlying mucosa [1, 6, 15, 16, 18]. Similar to previous reports, 14 lesions (63.6%) exhibited a typical yellowish appearance, but this feature may cause the misdiagnosis of lipoma, particularly in hyperechoic lesions on EUS. Besides, although superficial erosion or ulceration of esophageal GCTs is rare, further EUS differentiation between GCTs and gastrointestinal stromal tumors (GISTs) is needed [19].
Recently, EUS has improved the preoperative diagnosis of esophageal GCTs. The typical EUS pattern of esophageal GCTs appears as a hypoechoic, mildly inhomogeneous, and smooth-edged lesions located in mucosa or submucosa [6, 15]. But, 6 cases of esophageal GCTs derived from the MP layer have been reported in previous studies and 2 were reported in our study, suggesting that GCTs do not originate merely in mucosa and submucosa [6, 7, 12, 15]. In addition, hyperechoic lesions have also been gradually reported (7 cases in our study). In previous studies, most leiomyomas also showed hypoechoic images originated from muscularis mucosa or the MP layer. But the brightness of GCTs on EUS was similar to that of the submucosal layer, and it was significantly higher in comparison to leiomyomas, which can largely help in differential diagnosis of esophageal GCTs and leiomyomas [18].
Pathological diagnosis is still the gold standard. Forceps biopsy or EUS-FNA could provide histopathological information, which may help distinguish esophageal SMTs and guide the treatment strategy. However, the accuracy of them for the diagnosis of esophageal GCTs is also low because the specimen obtained was not adequate. In addition, forceps biopsy and EUS-FNA may increase the risk of tumor capsule damage and the difficulty of R0 resection [8]. Therefore, forceps biopsy and EUS-FNA was not performed in the patients in our study.
Adequate tumor specimens are a precondition for a correct pathological diagnosis. Pathologically, the tumor cells are separated by fibers into distinct nests or clusters, consisting of spherical or polygonal large cells with abundant eosinophilic granules cytoplasm and small, round, centrally located nucleus[6, 8]. In addition, schwannomas derived from Schwann cells, present as spindle tumor cells, sometimes with a lymphoid cuff [20]. If pathological characteristics are not enough for the diagnosis, IHC staining will be necessary. According to previous studies, some positive stainings have also been reported, such as S-100, periodic acid-Schiff (PAS), NSE, SOX-10, CD68, CD56 and nestin [7, 16]. Negative markers, which include CD34, CD117, SMA, CK and desmin, have also been reported [7, 16]. Conversely, esophageal leiomyomas are positive for SMA and desmin, and negative for S-100; GISTs are positive for CD117 and CD34, and negative for S-100 [7, 21].
According to previous studies, approximately 1–2% of esophageal GCTs are malignant [4–6]. Malignant GCTs usually present as tumor cell necrosis, tumor cell spindling, large nucleoli, increased mitotic activity (> 2 per 10 high power fields), increased nuclear size, and nuclear pleomorphism [4]. Meanwhile, Ki-67 positive cells (> 10%) and p53 positive cells (> 50%) were highly associated with malignancy [4]. In addition, remote metastasis after ER is reliable clinical evidence for the diagnosis of malignancy, however it remains unclear whether infiltrative features are credible evidence [6, 22]. None of the 22 patients in our study had postoperative pathology suggestive of malignancy, and there was no recurrence or metastasis during follow-up. This may suggest that esophageal GCTs have a lower malignant potential than other organs, consistent with some series of studies on ER for esophageal GCTs [8, 11, 15].
Management of esophageal GCTs includes surgical treatment, follow-up, and endoscopic treatment. Surgical treatment for esophageal GCTs, including open surgery and thoracoscopic surgery, is the most invasive approach with the high complications. Nowadays, it is only performed when malignancy is suspected or ER is contraindicated [1, 6]. EUS follow-up and ER become the preferred recommendation. For incidental findings, asymptomatic and relatively small (≤ 10 mm) esophageal GCTs, the patient should be involved in the decision to follow up or resect [6, 22]. However, long-term follow-up may increase the financial and psychological burden on the patient and may lead to missed diagnosis of malignancy, so ER is usually more acceptable. For symptomatic or large (> 10 mm), ER should be chosen [6, 22]. In the past, dehydrated alcohol injection, YAG laser ablation and argon plasma coagulation have been used for the endoscopic treatment of GCTs [22, 23]. But they have been rarely used since they cannot provide a complete removal of the lesion and a definitive tissue diagnosis. Currently, various ER techniques have been used in the treatment of esophageal GCTs, such as EMR, ESD, ESE, STER and EFTR. Appropriate approach of ER should be selected according to the origin and size of the lesion.
In previous studies, EMR appears safe and effective for treating esophageal GCT limited in the mucosal and submucosal layers and ≤ 10 mm in size [7, 14–16]. And ESD is also used in the treatment of GCTs located in the mucosal or submucosal layers, especially for those > 10 mm because of its high en bloc resection and R0 resection [8, 11]. These were demonstrated again in our study. However, as esophageal GCTs derived from the MP layer are gradually being reported, ER of them has become an inevitable topic. The literature over the past 20 years reported 6 cases of esophageal GCTs derived from the MP layer, of which 2 were treated surgically, 1 was treated with STER, 1 was treated with EFTR, 1 was followed up, and the treatment approach in 1 case was unclear [6, 7, 12, 13, 15]. STER is inspired by peroral endoscopic myotomy and natural orifice transluminal endoscopic surgery and has been gradually applied for treating esophageal SMTs derived from the MP layer. One systematic review and meta-analysis study proved the long-term efficacy and safety of STER for upper gastrointestinal SMTs, and all of STER-related complications can be managed conservatively [24]. In previous studies, four cases of STER treated esophageal GCTs were reported, three of which originated in the submucosa and one in the MP layer [6, 12]. No tumor recurrence or metastasis was observed during the long-term follow-up. Besides, EFTR has also been reported for the treatment of esophageal SMTs derived from the MP layer [13, 25]. Although safety and efficacy have not been systematically evaluated, EFTR offers a novel option for the treatment of esophageal GCTs derived from the MP layer. Without doubt, more cases are needed to evaluate the long-term prognosis of STER and EFTR for esophageal GCTs derived from the MP layer.
Of 2 esophageal GCTs derived from the MP layer in our study, 1 lesion underwent ESD, and 1 lesion underwent ESE. ESE is adapted from ESD for the treatment of upper gastrointestinal SMTs derived from the MP layer [26]. The ESE longitudinally incises the mucosa and bluntly separates the complete tumor, which makes the ESE procedure simple and quick. ESE has been applied to the esophageal, cardial and gastric lesions, but the risk of bleeding and perforation is higher in the esophagus compared to other locations, due to the narrow lumen and thin wall without a serosal layer. As techniques for handing perforations advances, perforation is no longer an obstacle to performing ESE, and meanwhile, ESE will mature to allow fewer perforations. One retrospective study found that both STER and ESE were safe and effective techniques for the treatment of esophageal SMTs derived from the MP layer and have similar overall complications [27]. Besides, ESE is easier to operate than STER and 95% of esophageal GCTs are less than 20 mm in size [16]. So ESE has great potential in treating esophageal GCTs. In our study, one patient who underwent ESE had no intraoperative perforation or delayed bleeding during hospitalization and no metastasis or recurrence during 84-months follow-up. This is the first report to assess the safety, efficacy and long-term prognosis of ESE in esophageal GCTs.
There were some limitations in our study. First, this was a retrospective single-center study with a limited number of patients included. Second, we identified patients from the pathology database, which may have a selection bias. The patients we finally identified did not include patients who had surgery treatment or follow-up, so we did not compare clinical outcomes of surgery treatment or follow-up with ER.
In conclusion, esophageal GCTs are rare, and typical endoscopic features can improve the preoperative diagnosis of esophageal GCTs. Histopathology and IHC are necessary for diagnosis and distinguish between benign and malignant. ER is safe and effective for management of esophageal GCTs. Clinically, the appropriate approach of ER should be selected according to the origin and size of the lesion.