Natural History of Gastric Subepithelial Lesions Measuring 1 to 2 cm
Article information
Abstract
Objectives
Management strategies for subepithelial lesions (SELs) measuring 1–2 cm remain undefined in most guidelines due to limited data. We aimed to establish a practical clinical approach by analyzing the long-term outcomes of SELs in this size range.
Methods
We retrospectively reviewed cases of gastric SELs measuring 1–2 cm between 2011 and 2023. We reviewed esophagogastroduodenoscopy and endoscopic ultrasound images, alongside data from abdominal computed tomography, biopsy results, and long-term follow-up records.
Results
Gastric SELs were initially identified in 648 patients. Of these, 88 patients (mean age, 57.1±10.6 years; 60% women) had lesions measuring 1–2 cm and underwent further evaluation or surveillance. The mean lesion size determined by endoscopy was 13.7±3.3 mm. Among the 52 patients with histological confirmation, 24 (46%) were diagnosed with gastrointestinal stromal tumors (GISTs), whereas the remainder had benign lesions. Most gastric SELs were located in the fundus and upper body. Pathological analysis indicated that the majority of GISTs were low-risk. Three GISTs were stratified as moderate-to-high risk; notably, one measured 0.9 cm upon pathological examination. The median surveillance period for the 28 patients without histological confirmation was 55 months (range, 43–115 months). Two of these patients underwent endoscopic or wedge resection due to tumor enlargement; both lesions were histologically confirmed as low-risk.
Conclusions
Approximately half of the gastric SELs measuring 1–2 cm were GISTs, representing an intermediate- to high-risk tumors. Given that GISTs in the upper stomach were associated with elevated risk even when smaller than 2 cm, we recommend strict short-term surveillance for such lesions.
INTRODUCTION
Gastrointestinal subepithelial lesions (SELs) have been increasingly reported [1]. This rise in detection is likely attributable to endoscopists’ heightened awareness of SELs and the widespread adoption of endoscopic ultrasound (EUS) [2-4]. However, recent studies indicate that the incidence is not negligible [5-7], and given that tumors with malignant potential—such as gastrointestinal stromal tumors (GISTs), neuroendocrine tumors, lymphomas, and glomerular tumors—may be present, accurate identification is clinically significant.
While endoscopic observation, EUS, and computed tomography (CT) aid in characterization, histological confirmation is often definitive and necessary. Although standard biopsy may yield diagnostic tissue in rare cases, pathological confirmation typically requires bite-on-bite biopsy, incision biopsy, or fine-needle aspiration/biopsy (FNA/B), as these lesions originate from the subepithelium [8,9].
Generally, tissue biopsy is not required for lesions smaller than 1 cm or for stable tumors that have shown no size progression over a prolonged period. However, definitive characterization is necessary for tumors that are indeterminate or demonstrate growth. Lesions measuring between 1 and 2 cm often fall into this indeterminate category. Although guidelines for SELs larger than 2 cm are well-established [10], clear consensus is lacking for those measuring 1–2 cm. Furthermore, ambiguity remains regarding the suitability of observation alone for SELs in this size range. Therefore, we conducted this study to characterize SELs measuring 1–2 cm, a group that represents a clinical “gray zone.”
METHODS
Subjects
We identified patients initially diagnosed with gastric SELs during esophagogastroduodenoscopy (EGD) at a university hospital between 2011 and 2023. Based on EGD records, we included only patients with gastric SELs measuring 1–2 cm who underwent further examination or surveillance EGD. Lesion size was estimated by comparing the SEL to fully opened biopsy forceps (approximately 6–7 mm). Patients with a prior history of gastric SELs were excluded. Data regarding EGD, EUS, abdominal CT, and pathology were retrospectively retrieved from medical records. This study was approved by the Institutional Review Board (IRB) of The Catholic University of Korea Incheon St. Mary’s Hospital (IRB No. OC21RISI0071). Due to the retrospective nature of the study, the IRB approved the study protocol under the condition that all data were anonymized, and the need for informed consent was addressed in accordance with the institutional guidelines.
Outcomes and statistical analysis
The primary outcome was the confirmation of malignancy or malignant potential in SELs measuring 1–2 cm requiring treatment. Gastric SEL location and size were assessed using EGD records. Layer of origin, echogenicity, composition, EUS size, and EUS impression were extracted from medical records. Additionally, abdominal CT scans were reviewed. Tissue acquisition methods, success rates, tumor resection details, and surveillance EGD results were analyzed. Continuous variables (e.g., age, size) were expressed as mean±standard deviation, whereas categorical variables were presented as numbers and percentages. All statistical analyses were performed using IBM SPSS Statistics for Windows, version 25.0 (IBM Corp.).
RESULTS
Demographic and clinical characteristics of SELs
Between 2011 and 2023, we identified 648 patients diagnosed with gastric SELs based on EGD records. Of these, 88 patients with lesions measuring 1–2 cm underwent further evaluation or surveillance EGD. Among the 88 patients, 35 (40%) were men, with a mean age of 57.1±10.6 years. The mean lesion size determined by endoscopy was 13.7±3.3 mm (Table 1).
Demographic characteristics for gastric subepithelial lesions
Diagnostic evaluations of SELs
Of the 88 subjects, 58 received EUS. Lesions originated from the second layer in 5 patients, the third layer in 16, the fourth layer in 36, and the fifth layer in 1; the layer of origin was unspecified in 1 case. Based on EUS findings, 23 patients were diagnosed with leiomyoma, 15 with GIST, 6 with ectopic pancreas, and 5 with lipoma (Fig. 1). EUS impressions were reported for 49 patients; for the remaining nine, retrospective image review yielded no definitive findings. A comparison of EUS findings with the final histological diagnosis of GISTs (n=24) revealed that EUS impressions matched the diagnosis in only 56% of cases (10 of 18), demonstrating moderate diagnostic accuracy in this size range.
Flowchart of gastric subepithelial lesions. SELs, subepithelial lesions; EUS, endoscopic ultrasound; GIST, gastrointestinal stromal tumor; CT, computed tomography; FNA/B, fine-needle aspiration/biopsy.
Abdominal CT was performed on 66 patients. In 26 of them, no specific lesions were observed in the gastric cavity. Twenty records reported nonspecific “gastric SELs,” 16 suspected GISTs, three identified lipomas, and one described a polypoid lesion.
Tissue samples were collected from 60 patients. Fifty-one had their tissues examined by various methods, and nine had them removed directly by endoscopic or surgical excision. Tissue examination methods included general biopsy (n=8), biteon-bite biopsy (n=20), incisional biopsy (n=15), open biopsy (n=6), and FNA/B (n=2).
Success rates of various biopsy techniques
Standard biopsy was successful in 5 of 8 patients, with the lesions diagnosed as gastritis cystica profunda, lipoma, and heterotopic pancreas, respectively (Table 2, Fig. 2). Diagnostic yield was 90% (18/20) for bite-on-bite biopsy. Incisional biopsy was successful in 13 of 15 patients (87%), and unroofing biopsy was successful in 83%. Both patients who underwent FNA/B had successful tissue acquisition. Statistical comparison of diagnostic yields across biopsy techniques revealed no significant difference (p=0.128) (Table 2).
Tissue acquisition methods for gastric subepithelial lesions
Histological examination results of subepithelial lesions measuring 1–2 cm. Bx., biopsy; GIST, gastrointestinal stromal tumor; GCP, gastritis cystica profunda; FNA/B, fine-needle aspiration/biopsy; ER, endoscopic resection; NET, neuroendocrine tumor.
Characteristics of GISTs
Of the 52 patients with histological confirmation, 24 (46%) were diagnosed with GISTs, whereas the remainder had benign lesions (Table 3). The mean size of confirmed GISTs was 15.7±2.9 mm; however, post-resection measurements indicat-ed they were, on average, 1 cm larger than preoperative estimates. The most common location for GISTs was the fundus, accounting for 10 cases (42%), and most final pathologies were classified as low risk. Three GISTs were stratified as moderateto-high risk, including one measuring 0.9 cm (Supplementary Table 1 in the online-only Data Supplement).
Histologically confirmed types of gastric subepithelial lesions
Among the confirmed cases, 12 were identified as GISTs via EUS, of which 10 were confirmed as actual GISTs. The accuracy of EUS in identifying true GISTs was 56% (10/18). Additionally, eight patients were predicted to have a high probability of leiomyoma based on EUS findings but were ultimately diagnosed with GISTs.
Distribution of histologically confirmed SELs
A total of 24 patients were confirmed to have GISTs. The locations of the GISTs were as follows: fundus (n=10), upper body (n=6), lower body (n=5), midbody (n=1), antrum (n=1), and cardia (n=1). Although GISTs were found in various locations, the fundus was the most common site (Fig. 3). Notably, 67% (16 of 24) were located in the upper body and fundus.
Distribution of GIST (A) and leiomyomas (B). GIST, gastrointestinal stromal tumor.
Additionally, leiomyomas were identified in 16 patients, distributed as follows: upper body (n=6), cardia (n=5), and fundus (n=2). Sixty-nine percent (11 of 16) of leiomyomas were located in the upper body and cardia.
Comparative analysis of endoscopic findings and additional examinations
Compared with the size observed on the initial EGD, the lesions appeared slightly smaller on EUS and larger on CT. The size of SELs measured on EGD for those who underwent EUS was 1.7±0.3 cm, while the size on EUS was 1.5±0.3 cm. In patients with an abdominal CT, SELs on the initial EGD were 1.6±0.3 cm, and on the CT scan, 1.7±0.4 cm.
Most lipomas were diagnosed exclusively via EGD. However, EGD alone could not reliably distinguish between GISTs and leiomyomas. Notably, we observed high concordance between initial endoscopic impressions and histological diagnosis for leiomyomas located in the cardia.
Follow-up on unconfirmed subjects
The median surveillance period for the 28 patients with unconfirmed lesions was 55 months (range, 43–115 months). Two of these patients underwent endoscopic or wedge resection due to tumor enlargement; both lesions were histologically confirmed as low-risk.
DISCUSSION
The prevalence of gastrointestinal SELs is increasing due to advancements in modalities such as EUS and increased endoscopist awareness [11]. Active intervention is usually recommended for SELs exceeding 2 cm, consensus is lacking for lesions smaller than this threshold. Consequently, the absence of clear guidelines often leads to anxiety among both patients and physicians [10,12,13]. This study focused exclusively on SELs measuring 1–2 cm, with histological confirmation obtained in the majority of cases. Contrary to expectations, we observed a high prevalence of GISTs; notably, some were classified as highrisk despite measuring approximately 1 cm. This finding challenges the reliance on a strict 2 cm threshold for clinical decision-making.
This study has several strengths. First, literature specifically addressing SELs in the 1–2 cm range is scarce. Although a large-scale multicenter study reported an average incidental SEL size of 1.6 cm [7], specific research focusing on the 1–2 cm range remains limited [14]. Second, the high biopsy rate in our cohort enabled the validation of EGD diagnostic accuracy. EUS alone is often insufficient to clearly determine malignancy [15-17]. Although recent advancements in artificial intelligence show promise in predictive modeling, histological analysis remains the diagnostic gold standard [18]. Third, the single-center design ensured data homogeneity and minimized the risk of error by limiting examinations to four gastroenterologists.
Despite these strengths, this study has limitations. First, the single-center design limited the sample size. However, this design facilitated consistent data recording and allowed for the retrospective review of endoscopic images. Furthermore, longitudinal surveillance of this cohort is ongoing. Second, this study was retrospective. However, we mitigated this by ensuring the rigor of histological examination and surveillance data. This distinguishes our work from prior studies, which lack sufficient histological confirmation. Third, reliance on endoscopic estimation of lesion size represents a limitation. Post-resection dimensions of GISTs exceeded pre-operative endoscopic estimates by an average of 1 cm. Although size estimation was standardized using open biopsy forceps, significant discrepancies are likely, especially for lesions 1 to 2 cm in size. This suggests that endoscopic measurement is inherently difficult to obtain and inaccurate for predicting actual tumor size.
Despite the fact that histological confirmation remains the gold standard, imaging modalities such as EUS and CT are crucial for initial evaluation. Previous studies have demonstrated the limited accuracy of EUS in identifying true GISTs [13]. This underscores the inability of EUS alone to reliably differentiate GISTs from leiomyomas, particularly in the 1–2 cm range. Although CT is less effective for mass characterization, it is valuable for assessing extraluminal extension, lymphadenopathy, and distant metastases. The difference in lesion size measured by the two modalities was not statistically significant; however, CT measurements were slightly larger, likely reflecting the lesion’s full extent. Therefore, a comprehensive evaluation integrating EUS and CT findings is necessary; in equivocal cases, tissue acquisition is essential. Various tissue acquisition methods, including bite-on-bite biopsy, incision biopsy, and unroofing, demonstrate high diagnostic yields. In this study, the difference in diagnostic yields between methods was not statistically significant. However, incision biopsy and unroofing demonstrated high success rates and align with European Society of Gastrointestinal Endoscopy guidelines, which recommend mucosal incision-assisted biopsy as a firstline procedure [13].
GISTs were most frequently located in the upper stomach (fundus and proximal corpus). This anatomical location presents technical challenges for EGD and EUS, supporting the need for a more aggressive surveillance protocol for upper gastric SELs. In conclusion, approximately half of the gastric SELs measuring 1–2 cm in this study were diagnosed as GISTs, a subset of which were stratified as moderate-to-high risk. For SELs located in the upper stomach, even those smaller than 2 cm, we recommend close monitoring with short-term surveillance and aggressive efforts to obtain histological confirmation.
Supplementary Materials
The online-only Data Supplement is available with this article at https://doi.org/10.7704/kjhugr.2025.0057.
Clinical and pathological characteristics of histologically confirmed GISTs
Notes
Availability of Data and Material
The datasets generated or analyzed during the study are not publicly available due to patient confidentiality and ethical restrictions but are available from the corresponding author on reasonable request.
Conflicts of Interest
Younghee Choe, a contributing editor of the Korean Journal of Helicobacter and Upper Gastrointestinal Research, was not involved in the editorial evaluation or decision to publish this article. All remaining authors have declared no conflicts of interest.
Funding Statement
None
Acknowledgements
None
Authors’ Contribution
Conceptualization: Younghee Choe, Byung-Wook Kim. Data curation: Younghee Choe. Formal analysis: Younghee Choe, Byung-Wook Kim. Investigation: Younghee Choe. Methodology: Younghee Choe, Byung-Wook Kim. Project administration: all authors. Resources: all authors. Software: Younghee Choe. Supervision: Byung-Wook Kim. Validation: Younghee Choe, Byung-Wook Kim. Visualization: Younghee Choe. Writing—original draft: Younghee Choe. Writing—review & editing: all authors. Approval of final manuscript: all authors.