Gastric cancer, one of the most common malignancies globally, was previously the most prevalent cancer in Korea [1,2]. Since the introduction of the National Gastric Cancer Screening Program in 1999, the proportion of patients diagnosed at earlier stages has markedly increased, leading to a 35% increase in the incidence of surgically treatable cancers over the past two decades. Consequently, nearly half of the patients are now diagnosed at surgically resectable stages, and the 5-year survival rate has improved to 77.5% by 2019 [1,2].

Surgical resection of the affected portions of the stomach, with lymphadenectomy, has long been considered the only curative treatment. However, advances in endoscopic techniques have introduced endoscopic submucosal dissection (ESD) as a minimally invasive curative option for selected patients without deep submucosal invasion [1,3-5]. Historically, Schlatter [6] first introduced total gastrectomy for gastric cancer, and Macguire [7] later reported subtotal gastrectomy as an alternative procedure for preserving part of the stomach. Subsequent randomized controlled trials comparing total versus subtotal gastrectomy for distal third gastric cancers demonstrated comparable survival outcomes, and subtotal gastrectomy is now generally preferred for tumors located in the distal and middle thirds of the stomach [8,9].

Following the initial curative treatment, long-term surveillance is essential to detect remnant gastric cancer (RGC), which refers to any newly diagnosed gastric cancer arising in the residual stomach after partial gastrectomy, regardless of the original disease [10]. Although there is a strong consensus regarding the need for surveillance, optimal strategies, including examination modality, timing, and intervals, remain undefined. Among the available modalities, computed tomography is useful for detecting hematogenous or nodal recurrence, whereas esophagogastroduodenoscopy is considered the gold standard for detecting early RGC lesions owing to its high sensitivity [1].

However, endoscopic examination after gastrectomy remains technically challenging owing to several factors, including the limited space of the remnant stomach, deformity of the anastomotic sites, mucosal alterations associated with chronic bile reflux, retained food material that obscures visualization, and a tendency toward bleeding that may hinder adequate biopsy sampling [11]. This review provides a comprehensive overview of the endoscopic features of RGC after gastrectomy, including characteristic morphological patterns, common anatomical locations, and differences according to reconstruction types.

Subtotal gastrectomy is a well-known risk factor for RGC development, regardless of whether the initial disease was benign or malignant. Studies have reported that patients who underwent distal gastrectomies for gastric ulcers were more likely to develop RGC than those with duodenal ulcers, and the majority of RGC occurred in patients with Billroth-II (B-II) reconstruction as initial surgery [12-17]. Follow-up of more than 3470 patients who underwent gastric surgery for benign peptic ulcer disease revealed that 87 (2.5%) developed RGC at a median interval of 28.4 years since the initial operation, and the relative risk increased up to 7.3-fold in patients followed up for 40–45 years compared to that of those followed up for less than 15 years postoperatively [18]. Another large-scale observational study reported following up 4131 patients with surgery for benign peptic ulcer disease revealed RGC in 1.1%; the relative risk increased by 2.1-fold 25 years after the operation compared to less than 17 years after the operation [19].

For malignant cases, Sano et al. [20] reported that among 1475 patients who underwent distal gastrectomies for gastric cancer, RGC occurred in 1.4%, which was consistent with a literature review of 20 studies involving 12785 patients who underwent curative resections, showing an overall RGC rate of 1.9%. The mean interval from resection to recurrence was 42 months, ranging from 4 to 105 months [20]. Western studies reported similar findings, with a cumulative incidence of 2.9% at 5 years, rising to 4% at 20 years post-surgery, with the majority of recurrences located at the cardia of the remnant stomach [21]. A recent United States population-based study using the SEER database demonstrated that among 17782 patients who underwent partial gastrectomies for gastric cancer, RGC was found in 1.3% at a median interval of 5.6 years since the initial operation and had a 7.7-fold increased risk of developing RGC compared to the general population, with a 20-year cumulative incidence of 1.88% [22]. These findings highlight the importance of long-term surveillance, even decades after the initial surgery [23].

Notably, Hosokawa et al. [24] further corroborated these findings, reporting a cumulative 5-year RGC prevalence of 2.4% and a 10-year prevalence of 6.1%, with the most frequent recurrence occurring along the lesser curvature of the remnant stomach rather than at the anastomosis site. Data from Korean patients have shown comparable results. In a study of 220 patients who underwent distal gastrectomies for gastric cancer, RGC occurred in 3.6% of the cases, with a mean interval of 32 months; among these, 41% occurred in the remnant stomach and 59% occurred at the anastomosis site [25].

Various nomenclatures and definitions have been proposed for RGC based on factors such as location, time interval since the initial surgery, and type of reconstruction; however, no consensus has been reached. Depending on its anatomical location, RGC may develop at the gastroduodenal or gastrojejunal anastomosis sites, along the gastric stump line, or in nonstump areas remote from the anastomosis or stump line (Fig. 1). Kaminishi [10] classified RGC into three categories: residual cancer, defined as cancer developing in the remnant stomach excluding the anastomosis site within 10 years postoperatively;recurrent cancer, defined as cancer developing at the anastomosis site within 10 years; and newly developed cancer, defined as any RGC occurring more than 10 years after surgery. An observational study by Oh et al. [26] analyzed 391 patients with RGC according to Kaminishi’s classification: 33% were residual cancers, 19.9% were recurrent cancers, and 47.1% were newly developed cancers. Among patients with residual cancer, 99.2% underwent initial surgery for malignant disease, whereas 100% of those with recurrent cancer and 56.5% of those with newly developed cancer had malignant primary lesions. Notably, all residual cancers occurred at non-anastomotic sites, all recurrent cancers at the anastomosis site, and newly developed cancers were slightly more common at the anastomosis site (56.5%) [26]. Another observational study reviewed 55 patients with RGC using the Kaminishi classification [27]. It reported that all patients with residual and recurrent cancer had undergone surgery for malignant disease as the initial condition, whereas 14 of 23 patients with newly developed cancer had undergone surgery for benign disease. The mean time interval from initial surgery was 20 years for newly developed cancer, which was significantly longer than that for residual or recurrent cancer. However, early gastric cancer (EGC) was less frequently detected in newly developed cancers (8.7%) than in residual and recurrent cancers (63.6% and 33.3%, respectively). This study also described the endoscopic appearance of RGC. Among the 16 cases of EGC, 71% of the residual cancers appeared as elevated-type lesions, whereas 86% of the recurrent cancers presented as depressed-type lesions. Among the 39 patients with advanced RGC, 66.7% of the newly developed cancers were ulcerative Bormann type II lesions, and 78.6% of the recurrent cancers were ulceroinfiltrative Bormann type III lesions.

Other studies have emphasized that fibrotic remodeling and granulomatous reactions are often accompanied by elevated lesions after gastrectomies, further underscoring the importance of histopathological confirmation [28,29]. Representative endoscopic images of RGC at various locations are shown in Figs. 2 and 3.

A multicenter Japanese study reported that in patients with B-I reconstructions for benign disease, RGC occurred most commonly in the residual remnant stomach, excluding the anastomosis site (46%), followed by the anastomosis site (26%) and the stump line (22%) [30]. Similar patterns were observed for B-I reconstructions in malignant disease, with the highest RGC frequency at non-stump sites (51%), followed by the stump line (23%) and the anastomosis site (20%). In contrast, for B-II reconstructions in benign disease, the highest RGC rate was observed at the anastomosis site (60%), followed by the non-stump areas (20%) and stump lines (12%). The corresponding rates for B-II reconstruction in malignant diseases were 38%, 31%, and 17%, respectively. The mean interval for primary benign disease was 32 years, with a relatively longer interval for B-II reconstructions (33.9 years), while the interval for primary malignant disease was 18 years, with a relatively longer interval for B-II reconstruction (21.3 years).

These trends are summarized in Table 1, which highlights key Western studies on the incidence, interval, and anatomical distribution of RGC after partial gastrectomy [15-19,21-23]. To complement these findings, Table 2 presents major observational studies from Asia, including data from Japan and Korea, providing additional insights into the regional differences in RGC incidence patterns, reconstruction types, and recurrence locations [24,26,27,29-33].

Total gastrectomies are generally performed in patients with proximal gastric cancer. In these patients, recurrence can occur at the esophagojejunal anastomosis site because of either tumor implantation or residual microscopic intramural cancer cells. The reported incidence of anastomotic site recurrence varies widely among studies; therefore, meticulous endoscopic evaluation of the anastomotic site is recommended, even in patients who undergo total gastrectomies [34,35].

In a large-scale study by Lee et al. [36] involving 848 patients who underwent total gastrectomies with R0 resections, recurrence was observed in 167 patients (19.7%) during a median follow-up of 58 months; however, only five patients (0.6%) developed loco-regional recurrence at the peri-anastomotic area. Based on these results, the authors concluded that in patients who underwent total gastrectomies, radiologic examinations, particularly abdominal computed tomography, play a more critical role than endoscopy in detecting recurrence, as the majority of recurrent lesions could not be adequately evaluated with endoscopy only.

In another study, Na et al. [37] analyzed 17 patients with recurrence after total gastrectomies with negative resection margins and found that five patients (29.4%) had recurrence localized to the anastomosis site, while the remaining 12 patients (70.6%) presented with combined distant metastases. The mean time to recurrence was 20 months, suggesting that early recurrence is not uncommon.

Furthermore, a prospective analysis by Lee et al. [38] that included 622 patients with advanced gastric cancer who underwent total gastrectomies revealed an overall recurrence rate of 37.5% (233 patients), with 17 patients (4.3%) developing local recurrence specifically at the anastomosis site. Among these, anastomotic stenosis was the most common endoscopic finding, followed by mass formation, ulceration, mucosal discoloration, and nodularity. These findings indicate that endoscopic recognition of subtle mucosal and structural changes at the anastomosis site is essential because recurrent lesions may present with a nonspecific appearance.

Collectively, these studies suggest that while the overall rate of anastomosis site recurrence after total gastrectomies is lower than that after distal gastrectomies, clinicians should maintain a high index of suspicion. Routine endoscopic surveillance, particularly in combination with cross-sectional imaging, may complement early detection. Nevertheless, given the limited diagnostic yield of endoscopy alone in detecting deep-seated perianastomotic recurrences, a multimodal surveillance strategy that integrates both endoscopic and radiological assessments is recommended.

Endoscopic resection, particularly ESD, has emerged as a viable option for selected cases of early stage RGC, offering organ preservation and lower morbidity rates than that via surgery. Indications for ESD in RGC are generally extrapolated from those for primary EGC. These include differentiated-type mucosal cancers without ulceration, traditionally defined as lesions ≤2 cm (classical absolute indication), and expandedcriteria lesions such as differentiated-type mucosal cancers with ulceration ≤3 cm, differentiated-type cancers with minute submucosal invasion (SM1, ≤500 μm) up to 3 cm, and undifferentiated-type mucosal cancers without ulceration ≤2 cm, provided there is no lymphovascular invasion [1,39,40].

Recent multicenter studies from Japan have demonstrated that ESD is a feasible and effective treatment for early stage RGC, despite technical challenges due to altered anatomy and fibrosis.

In a large multicenter retrospective cohort study involving 256 lesions from 12 institutions, Tsuda et al. [41] reported a 5-year overall survival rate of 81.3% and a gastric cancer-specific survival rate of 98.1%, supporting the long-term efficacy of ESD in appropriately selected patients. Similarly, Nonaka et al. [42] reported en bloc and curative resection rates of 94% and 78%, respectively, with an acceptable perforation rate of 1.4%, underscoring the procedural safety in expert hands. These findings suggest that ESD should be considered a primary therapeutic option for early RGC lesions that fulfill curative criteria, such as differentiated-type histology, size ≤20 mm, absence of ulceration, and no lymphovascular invasion. However, when post-ESD histopathology reveals non-curative factors (e.g., submucosal invasion >500 μm, lymphovascular invasion, positive margins), conversion to completion total gastrectomy with lymphadenectomy is recommended, given the risk of lymph node metastases [39,43]. In such cases, a multidisciplinary team approach is essential to determine surgical fitness, especially in older adult patients or those with prior major abdominal surgery. Recent advances, including laparoscopic or robotic gastrectomy, may offer less invasive surgical options for selected patients requiring completion total gastrectomies [44].

Owing to the improved efficacy of acid-suppressive agents, such as proton pump inhibitors, the incidence of gastrectomies for benign conditions, including gastric and duodenal ulcers, has markedly decreased over the past two decades. In contrast, advancements in endoscopic technology and the implementation of national cancer screening programs have facilitated the earlier detection of gastric cancer, leading to an increased proportion of surgically resectable cases.

Several studies have reported differences in the interval to RGC, depending on the underlying indication for initial gastrectomy. In Korean studies, An et al. [28] reported that the mean interval of RGC after malignant disease was 6.8 years, with 49% of cases detected within 5 years, while the mean interval for benign disease was 32.4 years. Han et al. [45] reported a mean interval of 13.4 years for malignant disease, and an observational study by Lee et al. [46] reported a mean interval of 7 years. The interval to RGC after benign disease ranged from 22 to 40 years [31,45,46]. A large Japanese multicenter analysis reported that approximately 55% of RGC cases were detected within three years of the initial operation, and a pooled review of 20 studies demonstrated a mean interval of 40 months, with 23% of cases detected more than five years after surgery [20].

In a Korean study, An et al. [31] reported that among 38 patients with RGC, those who underwent annual surveillance endoscopy were diagnosed at an earlier stage and had a significantly better prognosis than did patients who underwent less frequent examination. In comparison, in Europe, post-gastrectomy surveillance is often symptom-driven rather than routine endoscopy-based because there is a lack of randomized trial data to support intensive surveillance [47]. Therefore, despite this evidence, there is still no universal consensus regarding the optimal surveillance interval or duration. Currently, recommendations are largely based on expert opinions and clinical experience, as evidence linking surveillance intervals to overall survival outcomes remains limited.

As the majority of RGCs after gastrectomy for malignant disease occur within the first five years, endoscopic surveillance during this period is critical. However, long-term monitoring is also advisable, particularly for patients who underwent gastrectomies for benign diseases because late-onset RGC remains a concern. Early detection of RGC significantly improves prognosis and increases the likelihood of eligibility for minimally invasive therapies, such as ESDs, especially for lesions confined to the mucosa or superficial submucosa.

Considering that early stage RGC is often asymptomatic, endoscopy remains the only reliable modality for visual assessment and histological confirmation. Based on current data, annual endoscopic surveillance for at least 10 years is recommended for asymptomatic patients [25,27,45]. For patients with persistent symptoms, endoscopy should be performed as soon as possible, regardless of the routine interval. Additionally, for patients who undergo distal gastrectomy for benign diseases, surveillance should begin 15–20 years postoperatively, as the risk of late RGC remains significant even decades after surgery.

Detecting RGC in a postoperative stomach can be challenging because of altered anatomy, surgical staples, bile reflux, and food residue that impair mucosal visibility. To overcome these limitations and improve the diagnostic yield, several advanced endoscopic imaging techniques have been developed and are increasingly being applied in the clinical surveillance of patients after gastrectomies.

Image-enhanced endoscopy technologies such as narrow band tmaging, blue laser imaging, and i-scan provide real-time mucosal and vascular contrast enhancement without the need for dyes. These technologies have demonstrated improved detection of early gastric neoplasia, especially when combined with magnifying endoscopy, which allows visualization of the microsurface and microvascular patterns of suspicious lesions [48]. Similarly, chromoendoscopy, particularly with indigo carmine or acetic acid, is still valuable in the remnant stomach setting to enhance subtle lesion margins, especially along the suture line or anastomosis site [5].

Recent advancements in artificial intelligence (AI)-assisted systems for endoscopic image analysis have shown promise for improving detection rates and reducing inter-observer variability. Convolutional neural network-based algorithms have been trained to detect gastric neoplasia, and pilot studies have demonstrated high sensitivity even in challenging post-gastrectomy anatomies [49,50]. Although AI tools have not yet been widely adopted for RGC surveillance, their integration into routine practice is expected to further enhance early detection, especially in high-risk populations.

Taken together, these advanced technologies offer synergistic benefits when integrated into RGC surveillance strategies. Endoscopists should be encouraged to use these modalities based on their availability, patient-specific risks, and anatomical considerations.

Endoscopic evaluation of postgastrectomy patients is technically challenging because of anatomical deformities, mucosal changes, and limited visualization of the remnant stomach. Clinicians must be aware of the different potential sites of recurrence and pay particular attention to the anastomosis, where distinguishing minimal changes associated with RGC from non-specific inflammatory changes can be difficult.

Meticulous and systematic inspection of the anastomosis site and remnant gastric mucosa is essential and should be performed with the same level of diligence as routine endoscopic surveillance of gastric cancer in a native stomach. Given that early RGC lesions may present with only subtle endoscopic findings, obtaining targeted biopsies from the anastomosis site is strongly recommended, even when there is no clear endoscopic suspicion. Early histologic confirmation can improve diagnostic accuracy and facilitate timely intervention, ultimately improving patient outcomes through the early detection and treatment of RGC. In addition, clinicians should remain vigilant during long-term surveillance because timely diagnosis and intervention are critical for improving patient outcomes. Future research is warranted to establish standardized surveillance protocols tailored to individual RGC risk profiles, with a particular emphasis on defining the optimal interval, duration, and modality of endoscopic follow-up after gastrectomy.