Published online Sep 28, 2026. doi: 10.3748/wjg.119724
Revised: March 10, 2026
Accepted: May 6, 2026
Published online: September 28, 2026
Processing time: 202 Days and 19.2 Hours
Subepithelial lesions (SELs) of the upper gastrointestinal tract are frequently detected during routine endoscopy. To date, their natural history remains unclear, and evidence-based guidelines for their management are lacking.
To investigate the clinical course of asymptomatic SELs through a large-scale multicenter cohort.
This multicenter study was conducted at seven university hospitals in South Korea. We retrospectively analyzed the clinical, endoscopic, endoscopic ultrasonography, computed tomography, and histological data of patients who were newly diagnosed with asymptomatic SELs between January and December 2011.
The study followed 2546 patients with SELs for ≥ 12 months (median follow-up duration: 74 months; range: 12-205 months). A ≥ 25% increase in lesion size was observed in 287 patients (11.3%), including increases of ≥ 5 mm in 145 (5.7%) patients and ≥ 10 mm in 44 (1.7%) patients. The 5-year cumulative risks of intervention and a ≥ 25% increase in size were 1% and 4%, respectively. The risk factors for lesion growth included younger age, larger baseline size, absence of rolling signs, mucosal erosion, and heterogeneous echogenicity. Notably, the presence of a rolling sign was protective.
The risk of significant lesion growth and intervention in small, asymptomatic SELs was low. These findings support conservative management and less intensive surveillance of SELs < 3 cm without high-risk features.
Core Tip: In this large multicenter cohort of asymptomatic upper gastrointestinal subepithelial lesions (SELs), long-term follow-up demonstrated a low cumulative risk of progression, with a 5-year risk of only 1% for intervention and 4% for significant (≥ 25%) size increase. Younger age, larger baseline size, mucosal erosion, and heterogeneous echogenicity were associated with lesion growth, whereas the rolling sign was protective. These findings support conservative management with less intensive surveillance for SELs < 3 cm without high-risk features.
- Citation: Choi Y, Park JY, Kim SE, Park JM, Kim JS, Chung H, Kim SH, Park CH, Kim SH, Park SY, Lee BE, Lee SW, Lee A, Lee JH, Jung DH, Jo HH, Yoo IH, Min BH, Jung HY. Natural history of asymptomatic subepithelial lesions of the upper gastrointestinal tract: A multicenter study in Korea. World J Gastroenterol 2026; 32(36): 119724
- URL: https://www.wjgnet.com/1007-9327/full/v32/i36/119724.htm
- DOI: https://dx.doi.org/10.3748/wjg.119724
Asymptomatic subepithelial lesions (SELs) of the upper gastrointestinal (GI) tract are relatively common. In South Korea and Japan, the detection of asymptomatic upper GI SELs is steadily increasing, likely because of nationwide gastric cancer screening programs, which include endoscopic examinations. Recent single-center[1] and multicenter[2] studies have reported an incidence of upper GI SELs of approximately 1.6%-1.7% in South Korea.
The diagnosis and management of upper GI SELs vary among countries. For example, guidelines in the United States[3,4] and Europe[5] generally recommend endoscopic ultrasonography (EUS) and tissue sampling to distinguish potentially malignant lesions, such as GI stromal tumors (GISTs). In Japan, additional tests, such as computed tomography (CT) or EUS, are recommended only in the presence of risk factors, such as mucosal necrosis, bleeding, or irregular parenchymal features or borders[6]. Symptomatic SELs are treated in clinical practice, but the diagnosis and management of incidentally found asymptomatic SELs are controversial[7].
Several reports have described the natural course of asymptomatic upper GI SELs, including significant lesion growth in 3.6%[8] and 7.4%[9] of patients. Additionally, a large single-center study and a multicenter study reported that interventions were performed in 3.9%[9] and 5.2%[2] of cases, respectively. The evidence primarily came from single-center retrospective studies with relatively small sample sizes, often lacking information on the final histological diagnosis, which limits the generalizability of the study. Additionally, although many studies have used a criterion of “≥ 25% increase in size relative to the initial size” to define significant growth, the appropriateness of this threshold in clinical practice remains uncertain. For example, an interim report from an ongoing prospective study in Japan used absolute size increases of 5 mm or 10 mm, regardless of the initial size, as criteria[10].
This study analyzed the natural course of upper GI SELs in a large patient cohort across multiple centers in South Korea using EUS and CT findings, as well as data on interventions and histological diagnosis. This study also applied various criteria for defining “size increase” to identify clinically meaningful outcomes relevant to real-world practice.
This retrospective cohort study was conducted at seven university hospitals in South Korea. All patients diagnosed with SELs between January 1 and December 31, 2011, at the participating institutions were retrospectively enrolled through a review of their medical records. To minimize the limitations of retrospective studies, such as selection bias, this study included all patients with SELs who had undergone at least one esophagogastroduodenoscopy (EGD) examination at the participating centers during the study period. Among the overall population, cases with a follow-up duration of ≥ 12 months and at least one follow-up EGD were classified as the follow-up cohort and included in the analysis of the natural course. Patients without available EGD results (e.g., patients who underwent EGD at an outside institution and were seen only for consultation at the participating institutions) were excluded from the study.
The study protocol was reviewed and approved by the Institutional Review Board of Seoul National University Bundang Hospital (No. B-2311-864-104), which waived the requirement for written informed consent owing to the retrospective design.
Data collected through retrospective medical record review included basic demographic information (such as sex and age), EGD findings at diagnosis and follow-up [including lesion size (longest diameter), location, surface mucosal changes such as erosion or ulceration, gross appearance such as roundness, and presumed diagnosis], EUS findings (size, rolling sign, cushion sign, originating layer, echogenicity, homogeneity, internal changes, and presumed diagnosis), CT findings (size and presumed diagnosis), and histological diagnosis. Biopsy forceps were used as a reference for size comparison when measuring the gross size of the lesions during EGD. When multiple examinations were performed at the time of diagnosis, lesion size was determined based on EUS measurements when available. If EUS data were not available, CT measurements were used when the lesion was identifiable on CT; otherwise, endoscopic measurements were used.
Intervention was defined as therapeutic resection of the lesion during follow-up, and in cases in which an intervention was performed, the method, timing, and histopathological findings were recorded.
Student’s t-test and χ2 test were used to compare continuous and categorical variables between groups, respectively. Cox univariate and multivariate analyses were performed to identify risk factors, and variables with P < 0.2 in the univariate analysis were included in the multivariate model. Cumulative risk was analyzed using Kaplan-Meier curves and the log-rank test. The analyses were performed using IBM SPSS Statistics for Windows, version 25.0 (IBM Corp., Armonk, NY, United States). In all analyses, a P < 0.05 was considered statistically significant.
A total of 4052 patients received medical care for upper GI SELs at least once during the study period across all participating institutions. Of these, 3440 patients were included in the primary analysis after excluding those with missing medical records or test results. A follow-up of ≥ 12 months was achieved in 2546 patients. The median follow-up duration was 74 months (interquartile range: 37-125 months; range: 12-205 months), who were classified according to initial tumor size: < 1 cm (n = 1600), 1-2 cm (n = 757), 2-3 cm (n = 147), and ≥ 3 cm (n = 42). The proportion of patients in each group who experienced an increase in size of ≥ 25% during follow-up was 12.8%, 9.2%, 5.4%, and 9.5%, respectively. When classified by absolute increase, the proportions of patients with an increase of ≥ 5 mm was 4.3%, 7.3%, 5.4%, and 9.5%, respectively, while those with an increase of ≥ 10 mm was 1.2%, 2.5%, 1.4%, and 9.5%, respectively. Among the 1600 patients with tumors < 1 cm, 13 (0.8%) experienced lesion growth to ≥ 2 cm. Among the 2357 patients with tumors < 2 cm, 11 (0.5%) experienced lesion growth to ≥ 3 cm (Figure 1).
Among the 2546 patients in the study, EUS was performed at the time of diagnosis in 881 patients (34.6%). The comparison of baseline characteristics according to whether EUS was performed at diagnosis is presented in Supplementary Table 1. Briefly, patients in the EUS group were younger (mean age: 63.92 ± 13.62 years vs 67.19 ± 13.01 years, P < 0.001), and the lesions were larger (lesion size: 11.11 ± 6.05 mm vs 7.27 ± 5.34 mm, P < 0.001). Furthermore, the proportion of lesions ≥ 2 cm (15.1% vs 3.4%, P < 0.001) and those with an irregular shape (12.8% vs 5.7%, P < 0.001) was higher in the EUS group. The proportion of lesions with a rolling sign was also lower (49.4% vs 60.3%, P < 0.001) in the EUS group. CT imaging was performed in 283 patients (11.1%). In contrast, 124 patients (4.9%) underwent concurrent EGD, EUS, and CT.
The EGD and EUS characteristics were compared between patients with an SEL size increase of ≥ 25% and those without such an increase. Endoscopically, the size-increased group demonstrated a smaller initial tumor size (7.45 ± 5.50 vs 8.75 ± 5.92, P < 0.001) and a lower proportion with a positive rolling sign (47.9% vs 58.0%, P = 0.005). On EUS, a higher proportion of tumors in this group were located in the proper muscle layer (69.9% vs 56.9%, P = 0.015). Additionally, a higher proportion also showed heterogeneous echogenicity (33.0% vs 24.8%, P = 0.063), although this difference was not statistically significant (Table 1).
| Increased (n = 287) | Not increased (n = 2259) | P value | ||
| Sex | Male | 119 (41.5) | 1043 (46.2) | 0.132 |
| Female | 168 (58.5) | 1216 (53.8) | ||
| Age at diagnosis (mean ± SD) | 60.52 ± 13.29 | 66.76 ± 13.15 | < 0.001 | |
| Duration of F/U (mean ± SD, month) | 91.75 ± 46.89 | 77.79 ± 46.17 | < 0.001 | |
| Size (mean ± SD) | 7.45 ± 5.50 | 8.75 ± 5.92 | < 0.001 | |
| Size group | < 1 cm | 205 (71.4) | 1395 (61.8) | 0.007 |
| 1-2 cm | 70 (24.4) | 687 (30.4) | ||
| 2-3 cm | 8 (2.8) | 139 (6.2) | ||
| ≥ 3 cm | 4 (1.4) | 38 (1.7) | ||
| Shape | Round | 198 (89.6) | 1814 (91.7) | 0.297 |
| Irregular | 23 (10.4) | 165 (8.3) | ||
| Ulceration | Absent | 287 (100.0) | 2252 (99.8) | 0.425 |
| Present | 0 (0.0) | 5 (0.2) | ||
| Erosion | Absent | 287 (100.0) | 2237 (99.1) | 0.109 |
| Present | 0 (0.0) | 20 (0.9) | ||
| Rolling sign | Absent | 111 (52.1) | 657 (42.0) | 0.005 |
| Present | 102 (47.9) | 908 (58.0) | ||
| Cushion sign | Absent | 176 (84.2) | 1437 (85.9) | 0.499 |
| Present | 33 (15.8) | 235 (14.1) | ||
| Originating layer | Mucosa | 9 (8.0) | 128 (16.9) | 0.015 |
| Submucosa | 25 (22.1) | 198 (26.2) | ||
| Proper muscle | 79 (69.9) | 431 (56.9) | ||
| Echogenicity | Hypoechoic | 101 (87.8) | 702 (92.4) | 0.250 |
| Isoechoic | 4 (3.5) | 18 (2.4) | ||
| Hyperechoic | 10 (8.7) | 40 (5.3) | ||
| Homogeneity | Homogenous | 75 (67.0) | 570 (75.2) | 0.063 |
| Heterogenous | 37 (33.0) | 188 (24.8) | ||
| Internal change | None | 92 (80.7) | 536 (73.2) | 0.421 |
| Anechoic space | 7 (6.1) | 64 (8.7) | ||
| Calcification | 5 (4.4) | 37 (5.1) | ||
| Echogenic foci | 10 (8.8) | 85 (11.6) | ||
| Septation | 0 (0.0) | 10 (1.4) | ||
The results differed when using absolute values to determine the increase in lesion size. Compared with patients who had stable lesions without a size increase, the group with a size increase of ≥ 5 mm had a lower proportion of cases with a positive rolling sign on endoscopy (33.3% vs 58.3%, P < 0.001) and a higher proportion with heterogeneous echogenicity (39.8% vs 24.4%, P = 0.002). In the group with a ≥ 5 mm increase, the initial size was also larger (10.48 ± 7.29 vs 8.49 ± 5.78, P < 0.001), and the proportion of tumors with an irregular shape was higher (15.3% vs 8.2%, P = 0.007) (Table 2). Similar results were observed when using a threshold of ≥ 10 mm increase in size (Table 3). Taken together, these findings indicated that when the size increase was defined by relative proportion, the initial lesion size was smaller in the size-increased group. In contrast, when the increase in size was defined by absolute value, the initial size was larger in the size-increased group.
| Increased (n = 145) | Not increased (n = 2401) | P value | ||
| Sex | Male | 61 (42.1) | 1101 (45.9) | 0.374 |
| Female | 84 (57.9) | 1300 (54.1) | ||
| Age at diagnosis (mean ± SD) | 62.29 ± 12.45 | 66.28 ± 13.33 | < 0.001 | |
| Duration of F/U (mean ± SD, month) | 96.05 ± 45.75 | 78.36 ± 46.31 | 0.001 | |
| Size (mean ± SD) | 10.48 ± 7.29 | 8.49 ± 5.78 | < 0.001 | |
| Size group | < 1 cm | 68 (46.9) | 1532 (63.8) | < 0.001 |
| 1-2 cm | 55 (37.9) | 702 (29.2) | ||
| 2-3 cm | 18 (12.4) | 129 (5.4) | ||
| ≥ 3 cm | 4 (2.8) | 38 (1.6) | ||
| Shape | Round | 100 (84.7) | 1912 (91.8) | 0.007 |
| Irregular | 18 (15.3) | 170 (8.2) | ||
| Ulceration | Absent | 145 (100.0) | 2394 (99.8) | 0.582 |
| Present | 0 (0.0) | 5 (0.2) | ||
| Erosion | Absent | 145 (100.0) | 2379 (99.2) | 0.270 |
| Present | 0 (0.0) | 20 (0.8) | ||
| Rolling sign | Absent | 70 (66.7) | 698 (41.7) | < 0.001 |
| Present | 35 (33.3) | 975 (58.3) | ||
| Cushion sign | Absent | 88 (80.7) | 1525 (86.1) | 0.123 |
| Present | 21 (19.3) | 247 (13.9) | ||
| Originating layer | Mucosa | 7 (8.4) | 130 (16.5) | 0.153 |
| Submucosa | 24 (28.9) | 199 (25.3) | ||
| Proper muscle | 52 (62.7) | 458 (58.2) | ||
| Echogenicity | Hypoechoic | 74 (89.2) | 729 (92.0) | 0.205 |
| Isoechoic | 1 (1.2) | 21 (2.7) | ||
| Hyperechoic | 8 (9.6) | 42 (5.3) | ||
| Homogeneity | Homogenous | 50 (60.2) | 595 (75.6) | 0.002 |
| Heterogenous | 33 (39.8) | 192 (24.4) | ||
| Internal change | None | 59 (70.2) | 569 (74.7) | 0.439 |
| Anechoic space | 9 (10.7) | 62 (8.1) | ||
| Calcification | 3 (3.6) | 39 (5.1) | ||
| Echogenic foci | 13 (15.5) | 82 (10.8) | ||
| Septation | 0 (0.0) | 10 (1.3) | ||
| Increased (n = 44) | Not increased (n = 2502) | P value | ||
| Sex | Male | 17 (38.6) | 1145 (45.8) | 0.347 |
| Female | 27 (61.4) | 1357 (54.2) | ||
| Age at diagnosis (mean ± SD) | 61.86 ± 12.54 | 66.13 ± 13.32 | 0.031 | |
| Duration of F/U (mean ± SD, month) | 89.79 ± 45.92 | 79.18 ± 46.45 | 0.136 | |
| Size (mean ± SD) | 11.32 ± 9.42 | 8.55 ± 5.80 | 0.059 | |
| Size group | < 1 cm | 19 (43.2) | 1581 (63.2) | < 0.001 |
| 1-2 cm | 19 (43.2) | 738 (29.5) | ||
| 2-3 cm | 2 (4.5) | 145 (5.8) | ||
| ≥ 3 cm | 4 (9.1) | 38 (1.5) | ||
| Shape | Round | 26 (72.2) | 1986 (91.8) | < 0.001 |
| Irregular | 10 (27.8) | 178 (8.2) | ||
| Ulceration | Absent | 44 (100.0) | 2495 (99.8) | 0.767 |
| Present | 0 (0.0) | 5 (0.2) | ||
| Erosion | Absent | 44 (100.0) | 2480 (99.2) | 0.551 |
| Present | 0 (0.0) | 20 (0.8) | ||
| Rolling sign | Absent | 22 (71.0) | 746 (42.7) | 0.002 |
| Present | 9 (29.0) | 1001 (57.3) | ||
| Cushion sign | Absent | 27 (87.1) | 1586 (85.7) | 0.829 |
| Present | 4 (12.9) | 264 (14.3) | ||
| Originating layer | Mucosa | 2 (7.4) | 135 (16.0) | 0.474 |
| Submucosa | 8 (29.6) | 215 (25.5) | ||
| Proper muscle | 17 (63.0) | 493 (58.5) | ||
| Echogenicity | Hypoechoic | 27 (100.0) | 776 (91.5) | 0.287 |
| Isoechoic | 0 (0.0) | 22 (2.6) | ||
| Hyperechoic | 0 (0.0) | 50 (5.9) | ||
| Homogeneity | Homogenous | 12 (44.4) | 633 (75.1) | < 0.001 |
| Heterogenous | 15 (55.6) | 210 (24.9) | ||
| Internal change | None | 20 (74.1) | 608 (74.2) | 0.650 |
| Anechoic space | 1 (3.7) | 70 (8.5) | ||
| Calcification | 1 (3.7) | 41 (5.0) | ||
| Echogenic foci | 5 (18.5) | 90 (11.0) | ||
| Septation | 0 (0.0) | 10 (1.2) | ||
Multivariate analysis identified younger age [odds ratio (OR), 0.98; 95%CI: 0.96-0.99; P = 0.006], smaller initial size (OR, 0.95 per mm; 95%CI: 0.91-0.99; P = 0.007), absence of a rolling sign (OR, 0.52; 95%CI: 0.34-0.80; P = 0.003), and heterogeneous echogenicity (OR, 2.35; 95%CI: 1.44-3.84; P = 0.001) as significant risk factors for a ≥ 25% size increase (Supplementary Table 2). When absolute growth was considered, younger age, larger initial size, and absence of a rolling sign remained significant, while heterogeneous echogenicity showed marginal significance (Supplementary Tables 3 and 4). Kaplan-Meier analysis demonstrated a 5-year cumulative risk of 0.04 for ≥ 25% growth, with significantly faster progression in lesions ≥ 3 cm. The 5-year risks were 0.02 for ≥ 5 mm and 0.006 for ≥ 10 mm increases, and 0.004 and 0.001, respectively, for growth from < 1 cm to ≥ 2 cm and from < 2 cm to ≥ 3 cm (Figure 2).
Among the 2546 patients with a follow-up of at least 12 months, the annual growth rate was 1.17 ± 2.02 mm per year. According to the initial size of the SELs, the annual growth rate was 0.98 ± 1.46 mm, 1.16 ± 1.19 mm, 1.57 ± 0.71 mm, and 10.29 ± 9.89 mm per year for lesions measuring < 1 cm, 1-2 cm, 2-3 cm, and ≥ 3 cm, respectively. In other words, larger tumors at the time of diagnosis tended to grow more rapidly (P < 0.001) (Table 4).
| Total population | < 1 cm | 1-2 cm | 2-3 cm | ≥ 3 cm | P for trend | |
| Annual growth rate | 1.17 ± 2.02 | 0.98 ± 1.46 | 1.16 ± 1.19 | 1.57 ± 0.71 | 10.29 ± 9.89 | < 0.001 |
Comparing the length of time taken to reach a significant (≥ 25%) increase in size between the groups showed a mean follow-up duration of 536 days in the < 1 cm group, 1051 days in the 1-2 cm group, 1550 days in the 2-3 cm group, and 234 days in the ≥ 3 cm group. When size increase was defined based on absolute values, the mean follow-up duration for a lesion to increase in size by ≥ 5 mm was 1487 days in the < 1 cm group, 1733 days in the 1-2 cm group, 1809 days in the 2-3 cm group, and 511 days in the ≥ 3 cm group. The mean follow-up duration until a ≥ 10 mm increase in size was 1466 days in the < 1 cm group, 1837 days in the 1-2 cm group, 2102 days in the 2-3 cm group, and 695 days in the ≥ 3 cm group. Additionally, the mean duration for a size increase from < 1 cm to ≥ 2 cm was 260 days, while that from < 2 cm to ≥ 3 cm was 1494 days (Supplementary Table 5).
Among the entire cohort, histologic confirmation was attempted at the time of diagnosis in 210 patients (8.2%). Of these, conventional endoscopic biopsy was performed in 70 patients, bite-on-bite biopsy in 126 patients, and EUS-guided fine-needle biopsy in 14 patients. As a result, four cases of GIST, eight leiomyomas, five heterotopic pancreas lesions, and six lipomas were histologically diagnosed.
In this study, 159 patients underwent interventions within 12 months, and 53 patients underwent interventions after 12 months of follow-up. The histological diagnoses of the 53 lesions that were therapeutically resected included 23 GISTs, 11 leiomyomas, 2 neuroendocrine tumors, 2 heterotopic pancreas lesions, 6 schwannomas, 2 duplication cysts, and 7 other benign lesions (e.g., simple cysts, lipomas, and gastritis cystica profunda).
Regression analyses were performed to identify risk factors for intervention. According to the multivariate analysis, significant risk factors included age (OR, 0.98 per year; 95%CI: 0.96-1.00; P = 0.014), initial size (OR, 1.08 per mm; 95%CI: 1.06-1.09; P < 0.001), mucosal erosion (OR, 7.17; 95%CI: 2.17-23.74; P = 0.002), an absence of a rolling sign (OR, 0.41; 95%CI: 0.24-0.72; P = 0.002), and heterogeneous echogenicity (OR, 2.24; 95%CI: 1.41-3.55; P = 0.001) (Supplementary Ta
Assessment of the cumulative risk of intervention using Kaplan-Meier curves showed a 5-year cumulative risk of 1%. When classified by initial size, a significantly higher risk was observed in patients with an initial SEL size of ≥ 2 cm (Figure 3).
The SELs were analyzed according to location (esophagus, stomach, and duodenum). Esophageal and duodenal lesions were more common in males, whereas gastric lesions were more common in females (P < 0.001). At the time of diagnosis, duodenal lesions had a slightly larger mean size than esophageal and gastric lesions (8.68 ± 5.99 mm vs 8.47 ± 5.98 mm vs 9.35 ± 4.92 mm, P < 0.001). However, there was no statistically significant difference in the rate of intervention according to lesion location. Notably, the annual growth rate was significantly higher in gastric lesions (0.81 ± 3.37 mm/year vs 1.91 ± 9.10 mm/year vs 0.70 ± 2.49 mm/year, P = 0.004). Furthermore, EUS was performed significantly less frequently in duodenal lesions (40.6% vs 34.8% vs 21.8%, P < 0.001), which may reflect anatomical limitations in adequately visualizing duodenal lesions using an EUS scope (Supplementary Table 7).
This retrospective multicenter study analyzed the natural course of upper GI SELs in a large cohort of patients. Lesions that were initially larger in size tended to have a higher annual growth rate. In particular, lesions measuring ≥ 3 cm at diagnosis showed a higher proportion of size increase and at a faster rate. In contrast, lesions measuring < 1 cm or < 2 cm at diagnosis had a very low risk of growing to ≥ 2 cm or ≥ 3 cm, respectively.
Previous studies on the natural course of upper GI SELs reported variable rates of size increase[11-15]. Song et al[8] analyzed 964 incidentally detected SELs and found that 3.6% showed a ≥ 25% size increase during a mean follow-up of 47 months. Kim et al[9] examined 1859 cases at a tertiary center and reported a ≥ 25% size increase in 7.4% and interventions in 3.9% over 5 years. A prospective Japanese study of 567 patients reported a 5-year cumulative risk of ≥ 5 mm growth of 4.5%[10]. These discrepancies may be related to differences in baseline lesion size, follow-up duration, and measurement methods, with EUS generally providing more reliable assessments than EGD[15]. A recent systematic review found that large-scale studies (> 410 patients) reported relatively low growth rates (2.0%-8.5%), whereas smaller studies (< 145 patients) showed higher and more variable rates (5.4%-28.4%)[14]. This highlights the need for standardized criteria for measuring lesion growth. Importantly, individualized SEL surveillance and management is essential.
In the present study, 11.3% of lesions showed a size increase of ≥ 25% during follow-up, a proportion higher than those reported previously. However, when absolute growth was used as the criterion rather than a percentage increase, size increases of ≥ 5 mm and ≥ 10 mm were observed in 5.7% and 1.7% of all cases, respectively. Specifically, size increases of ≥ 5 mm were observed in 4.3%, 7.3%, 5.4%, and 9.5% of each group, while increases of ≥ 10 mm were seen in 1.2%, 2.5%, 1.4%, and 9.5%, respectively, indicating a significant difference in the ≥ 3 cm group. These findings indicate a low risk of clinically significant enlargement of upper GI SELs. Importantly, percentage-based criteria may overestimate growth risk, especially in cohorts with a high proportion of small lesions.
The optimal endoscopic follow-up interval according to the initial size of SELs has yet to be established. A higher detection rate of asymptomatic upper GI SELs is observed in Korea. This may be attributed to upper endoscopy, including follow-up endoscopic examinations, being widely accessible because of a nationwide health screening program[16]. In Japan, where circumstances are comparable to those in Korea, guidelines consistently recommend endoscopic follow-up at 6-12 months for asymptomatic SELs < 2 cm without high-risk features and for SELs measuring 2-5 cm in the absence of malignant features[6,17,18]. In the present study, analysis of the annual growth rate according to initial size revealed a marked difference at the 3 cm threshold, with lesions measuring ≥ 3 cm at diagnosis showing a significantly higher growth rate compared with that of those < 3 cm (10.29 ± 9.89, P for trend < 0.001). In addition, the interval from diagnosis to size increase was also significantly shorter for lesions ≥ 3 cm compared with that of those < 3 cm, regardless of whether the criterion for growth was based on percentage or absolute measures. The 5-year cumulative risk was 4%, 2%, and 0.6% for size increases of ≥ 25%, ≥ 5 mm, and ≥ 10 mm, respectively. The 5-year cumulative risk for increases from < 1 cm to ≥ 2 cm and from < 2 cm to ≥ 3 cm was 0.4% and 0.1%, respectively, both of which were extremely low. In this study, lesions measuring ≥ 3 cm were relatively underrepresented in the surveillance cohort because such lesions are typically considered for intervention at the time of diagnosis. Therefore, 3 cm cannot be proposed as an absolute cutoff value for intervention based on our data. Nevertheless, our findings suggest that lesions < 3 cm generally showed a favorable clinical course during follow-up. Even though lesions ≥ 3 cm included in this cohort were considered relatively indolent at diagnosis and were therefore not initially resected, they demonstrated a higher size increase and at a faster rate during follow-up. Considering these, resection or more intensive surveillance should be considered for upper GI SELs measuring ≥ 3 cm.
The risk factors for tumor size increase and intervention were similar to those reported previously[6,8,9,14,19], including larger initial tumor size, surface ulcers or erosions, and irregular shape (lobulated shape). In the present study, younger age at diagnosis and the absence of the rolling sign were also identified as risk factors for size increase, in addition to a smaller initial tumor size when using a ≥ 25% increase as the criterion. However, when size increase was assessed using an absolute threshold of 5 mm or 10 mm, a larger initial size was identified as a risk factor for growth. When a substantial number of small lesions are included, percentage-based criteria may lead to misleading interpretations, highlighting the need for caution when interpreting such results.
This study has several limitations. First, this study is limited by its retrospective design, which potentially increases the risk of bias and inadvertently influences the evaluation of the natural course, such as decisions regarding whether to perform follow-up examinations and the appropriate follow-up intervals. To minimize selection bias, we included all patients who underwent at least one follow-up examination at the participating centers during the enrollment period. Furthermore, not all patients underwent EUS examinations at baseline. Patients who underwent EUS at baseline were likely those with a relatively large SEL or those considered at high risk of malignancy. Additionally, the method used to measure lesion size during endoscopy was not fully standardized, which may have introduced inter-observer variability. The timing and modality of follow-up examinations were also not standardized and may have been influenced by lesion size or morphology. Consequently, additional biases may have arisen owing to incomplete EUS data, inter-institutional differences, and variations among individual clinicians, potentially limiting the generalizability of our findings. Another limitation relates to potential measurement variability in lesion size assessment. Although we defined clinically significant growth as an increase of ≥ 5 mm, a measurement error of approximately 2-3 mm is generally unavoidable even when using EUS, which provides relatively objective size estimation. Therefore, differences in measurement may have affected the classification of growth, particularly in small lesions. Third, only patients with a follow-up period of at least 12 months were included to observe the natural course of UGI SELs during follow-up. Consequently, cases in which the intervention was performed within 12 months of diagnosis were excluded. Therefore, caution is warranted when interpreting the results related to the risk factors for intervention in this study. This point should be particularly considered when interpreting the findings for lesions ≥ 3 cm. As lesion size is an important factor when determining the need for intervention at the time of diagnosis, lesions ≥ 3 cm included in this study were more likely to represent relatively indolent cases with a lower malignant risk, as lesions suspected to have a higher risk may have undergone intervention at diagnosis. Therefore, the findings of this study should not be interpreted as representing the overall characteristics of all lesions ≥ 3 cm. Finally, although this study was designed as a nationwide, multicenter study in Korea and initially enrolled > 4000 patients, the final number of patients included in the analysis was comparable to that of previous single-center studies in Korea, owing to the exclusion of cases with missing medical records or short follow-up durations.
This study also has several strengths. First, in addition to using the percentage-based criteria for size increase commonly employed in previous studies, absolute size changes were also analyzed. Moreover, size changes and the interval to growth based on the initial size at diagnosis were also examined, and the 5-year cumulative risks of size increase and intervention were determined. The results showed that among lesions < 3 cm, the proportion exhibiting clinically significant growth during follow-up was very low, and the time to size increase was notably long. However, the findings for lesions ≥ 3 cm should be interpreted with caution. These features distinguish our study from previous reports investigating the natural course of SELs.
This study demonstrated a low rate of clinically significant size increase in upper GI SELs during long-term follow-up. Consequently, for lesions < 3 cm without accompanying high-risk features, excessive concern regarding lesion growth or intensive surveillance may not be necessary. Additionally, considering the very low cumulative risk of meaningful growth among lesions < 1 cm, an initial follow-up of approximately 2 years or 3 years is a reasonable conservative approach in selected patients.
| 1. | Lee JH, Lee HL, Ahn YW, Lee KN, Jun DW, Lee OY, Han DS, Yoon BC, Choi HS. Prevalence of Gastric Subepithelial Tumors in Korea: A Single Center Experience. Korean J Gastroenterol. 2015;66:274-276. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 23] [Cited by in RCA: 20] [Article Influence: 1.8] [Reference Citation Analysis (0)] |
| 2. | Choe Y, Cho YK, Kim GH, Choi JH, Kim ES, Kim JH, Choi EK, Kim TH, Kim SH, Kim DH; Research Group for Endoscopic Ultrasound in Korean Society of Gastrointestinal Endoscopy. Prevalence, natural progression, and clinical practices of upper gastrointestinal subepithelial lesions in Korea: a multicenter study. Clin Endosc. 2023;56:744-753. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 10] [Cited by in RCA: 25] [Article Influence: 8.3] [Reference Citation Analysis (0)] |
| 3. | Jacobson BC, Bhatt A, Greer KB, Lee LS, Park WG, Sauer BG, Shami VM. ACG Clinical Guideline: Diagnosis and Management of Gastrointestinal Subepithelial Lesions. Am J Gastroenterol. 2023;118:46-58. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 139] [Cited by in RCA: 128] [Article Influence: 42.7] [Reference Citation Analysis (5)] |
| 4. | Sharzehi K, Sethi A, Savides T. AGA Clinical Practice Update on Management of Subepithelial Lesions Encountered During Routine Endoscopy: Expert Review. Clin Gastroenterol Hepatol. 2022;20:2435-2443.e4. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 127] [Cited by in RCA: 108] [Article Influence: 27.0] [Reference Citation Analysis (0)] |
| 5. | Deprez PH, Moons LMG, OʼToole D, Gincul R, Seicean A, Pimentel-Nunes P, Fernández-Esparrach G, Polkowski M, Vieth M, Borbath I, Moreels TG, Nieveen van Dijkum E, Blay JY, van Hooft JE. Endoscopic management of subepithelial lesions including neuroendocrine neoplasms: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy. 2022;54:412-429. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 325] [Cited by in RCA: 282] [Article Influence: 70.5] [Reference Citation Analysis (5)] |
| 6. | Nishida T, Hirota S, Yanagisawa A, Sugino Y, Minami M, Yamamura Y, Otani Y, Shimada Y, Takahashi F, Kubota T; GIST Guideline Subcommittee. Clinical practice guidelines for gastrointestinal stromal tumor (GIST) in Japan: English version. Int J Clin Oncol. 2008;13:416-430. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 288] [Cited by in RCA: 282] [Article Influence: 15.7] [Reference Citation Analysis (2)] |
| 7. | Kim SH, Park JY, Lee A, Lee BE, Min BH, Park CH, Jung DH, Jo HH, Chung H, Yoo IH, Park SY, Lee SW, Choi Y, Lee JH, Kim SH, Park JM, Kim JS, Kim SE, Jung HY. Current Practices in Histological Diagnosis and Management of Asymptomatic Gastric Subepithelial Lesions: A Multicenter Survey in Korea. Korean J Helicobacter Up Gastrointest Res. 2025;25:159-166. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 1] [Cited by in RCA: 3] [Article Influence: 3.0] [Reference Citation Analysis (0)] |
| 8. | Song JH, Kim SG, Chung SJ, Kang HY, Yang SY, Kim YS. Risk of progression for incidental small subepithelial tumors in the upper gastrointestinal tract. Endoscopy. 2015;47:675-679. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 50] [Cited by in RCA: 46] [Article Influence: 4.2] [Reference Citation Analysis (0)] |
| 9. | Kim B, Kang S, Lee E, Choi J, Chung H, Cho SJ, Kim SG. Gastric subepithelial tumor: long-term natural history and risk factors for progression. Surg Endosc. 2022;36:5232-5242. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 8] [Reference Citation Analysis (0)] |
| 10. | Iwamuro M, Mitsuhashi T, Inaba T, Matsueda K, Nagahara T, Takeuchi Y, Doyama H, Mizuno M, Yada T, Kawai Y, Nakamura J, Matsubara M, Nebiki H, Niimi K, Toyokawa T, Takenaka R, Takeda S, Tanaka S, Nishimura M, Tsuzuki T, Akahoshi K, Furuta T, Haruma K, Okada H. Results of the interim analysis of a prospective, multicenter, observational study of small subepithelial lesions in the stomach. Dig Endosc. 2024;36:323-331. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 4] [Cited by in RCA: 9] [Article Influence: 4.5] [Reference Citation Analysis (1)] |
| 11. | Gill KR, Camellini L, Conigliaro R, Sassatelli R, Azzolini F, Messerotti A, Woodward TA, Wallace MB, Jamil LH, Raimondo M. The natural history of upper gastrointestinal subepithelial tumors: a multicenter endoscopic ultrasound survey. J Clin Gastroenterol. 2009;43:723-726. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 49] [Cited by in RCA: 47] [Article Influence: 2.8] [Reference Citation Analysis (0)] |
| 12. | Kim MY, Jung HY, Choi KD, Song HJ, Lee JH, Kim DH, Choi KS, Lee GH, Kim JH. Natural history of asymptomatic small gastric subepithelial tumors. J Clin Gastroenterol. 2011;45:330-336. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 60] [Cited by in RCA: 53] [Article Influence: 3.5] [Reference Citation Analysis (0)] |
| 13. | Hu ML, Wu KL, Changchien CS, Chuah SK, Chiu YC. Endosonographic surveillance of 1-3 cm gastric submucosal tumors originating from muscularis propria. World J Gastroenterol. 2017;23:2194-2200. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in CrossRef: 5] [Cited by in RCA: 13] [Article Influence: 1.4] [Reference Citation Analysis (0)] |
| 14. | Iwamuro M, Okada H, Otsuka M. Natural Course and Long-Term Outcomes of Gastric Subepithelial Lesions: A Systematic Review. J Clin Med. 2025;14:1055. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in RCA: 5] [Reference Citation Analysis (0)] |
| 15. | Park H, Ahn JY, Kim GH, Na HK, Jung KW, Lee JH, Kim DH, Choi KD, Song HJ, Lee GH, Jung HY. Reliability of endoscopic ultrasonography and endoscopy in measurement of gastric subepithelial tumor size. Surg Endosc. 2023;37:2604-2610. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 2] [Reference Citation Analysis (0)] |
| 16. | Suh YS, Lee J, Woo H, Shin D, Kong SH, Lee HJ, Shin A, Yang HK. National cancer screening program for gastric cancer in Korea: Nationwide treatment benefit and cost. Cancer. 2020;126:1929-1939. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 98] [Cited by in RCA: 81] [Article Influence: 13.5] [Reference Citation Analysis (3)] |
| 17. | Nishida T, Blay JY, Hirota S, Kitagawa Y, Kang YK. The standard diagnosis, treatment, and follow-up of gastrointestinal stromal tumors based on guidelines. Gastric Cancer. 2016;19:3-14. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 365] [Cited by in RCA: 360] [Article Influence: 36.0] [Reference Citation Analysis (5)] |
| 18. | Nishida T, Kawai N, Yamaguchi S, Nishida Y. Submucosal tumors: comprehensive guide for the diagnosis and therapy of gastrointestinal submucosal tumors. Dig Endosc. 2013;25:479-489. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 220] [Cited by in RCA: 204] [Article Influence: 15.7] [Reference Citation Analysis (0)] |
| 19. | Jeon SW, Park YD, Chung YJ, Cho CM, Tak WY, Kweon YO, Kim SK, Choi YH. Gastrointestinal stromal tumors of the stomach: endosonographic differentiation in relation to histological risk. J Gastroenterol Hepatol. 2007;22:2069-2075. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 33] [Cited by in RCA: 40] [Article Influence: 2.1] [Reference Citation Analysis (0)] |