Proximal gastric cancer (PGC), which refers to tumors located in the upper third of the stomach, has been increasing in incidence globally. This trend is especially notable in East Asia and Western countries, possibly due to changes in diet, rising obesity rates, and widespread eradication of Helicobacter pylori (1–3). Advancements in early detection and the growing use of neoadjuvant therapy have led to an increasing number of patients being diagnosed with locally advanced PGC (LAPGC). For these patients, surgery remains the primary curative treatment (4, 5). Traditionally, total gastrectomy (TG) has been the preferred approach for LAPGC due to concerns about achieving complete oncologic resection, performing extensive lymphadenectomy, and avoiding proximal margin involvement (6, 7). However, TG has several well-known postoperative drawbacks, such as severe nutritional impairment, loss of the gastric reservoir and digestive hormones, and long-term reliance on nutritional supplements. These complications can significantly reduce quality of life and survival, particularly in elderly or nutritionally vulnerable patients (3, 8–10).

Proximal gastrectomy (PG), initially used only for early-stage PGC, has recently regained attention as a function-preserving alternative to total gastrectomy (TG). This renewed interest is largely due to the development of advanced reconstruction techniques, including double-tract reconstruction, jejunal interposition, and side overlap anastomosis (9, 11, 12). These techniques are designed to reduce reflux esophagitis and anastomotic complications, which were major concerns in earlier surgical practices. The main advantage of PG is the preservation of the distal stomach and pyloric function. This allows for improved food intake, better absorption of micronutrients (especially iron and vitamin B12), and a lower risk of sarcopenia and hypoalbuminemia (9). In patients with LAPGC who respond well to neoadjuvant chemotherapy, PG combined with adequate lymphadenectomy may provide oncologic outcomes comparable to TG while preserving better postoperative function (8, 13, 14).

Despite these theoretical advantages, PG remains underutilized for LAPGC due to persistent doubts regarding its oncologic safety (15). Concerns include potentially inadequate dissection of lymph node stations 5 and 6, increased risk of locoregional recurrence, and technical challenges related to anastomotic reconstruction (14). Furthermore, the lack of randomized controlled trials (RCTs) has limited the generalizability of available evidence. Most existing data come from retrospective observational studies, many of which are subject to confounding by indication and selection bias (16, 17). However, the recent proliferation of well-designed propensity score-matched (PSM) studies has helped mitigate these limitations by balancing baseline characteristics between groups, thereby enhancing the validity of comparative outcomes (1, 8, 13, 18, 19).

Given the emergence of PSM studies comparing PG and TG for LAPGC, there is now an opportunity to synthesize the available evidence in a comprehensive and statistically robust manner. Unlike previous meta-analyses that combined unmatched or early-stage gastric cancer populations, our study uniquely focuses on PSM cohorts specifically in the setting of proximal advanced gastric cancer. This approach helps reduce confounding and provides a more targeted and reliable comparison of surgical strategies in this clinically challenging subgroup. In this study, we conducted a meta-analysis of PSM studies comparing PG and TG in patients with LAPGC, aiming to evaluate whether PG can achieve equivalent oncologic outcomes while preserving postoperative function and reducing complications. Our primary endpoints were overall survival (OS) and recurrence-free survival (RFS), while secondary outcomes included surgical outcomes and postoperative complications. The findings aim to support surgical decision-making in this increasingly relevant clinical context.

This systematic review and meta-analysis was performed and reported following the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) Statement (20) and AMSTAR-2 (Assessing the methodological quality of systematic reviews) (21) guidelines (Supplemental materials).

The comparison between PG and TG for treating LAPGC has long attracted attention in surgical oncology. This systematic review and meta-analysis offers a comprehensive evaluation of PG versus TG, focusing on propensity score-matched studies and comparing surgical, perioperative, and oncological outcomes. The results suggest that PG and TG yield comparable long-term survival, despite differences in surgical techniques, postoperative complications, and recovery times.

One key finding of this meta-analysis was that PG resulted in significantly shorter operative times than TG, with an average reduction of approximately 25.7 minutes. This finding is consistent with previous studies suggesting that PG, which involves resection of the proximal stomach while preserving the pylorus, is technically less demanding and therefore requires less operative time. However, PG was associated with significantly greater intraoperative blood loss than TG (P = 0.02), possibly due to more complex dissection around the cardia and greater curvature, which are areas rich in blood vessels. Despite the increased blood loss, the overall surgical complexity and risk profile of PG remain acceptable. The reduced operative time may be particularly beneficial for elderly or frail patients.

Postoperative complications, which often lead to prolonged hospital stays and increased healthcare costs, are critical indicators when assessing the safety and technical feasibility of surgical procedures (9, 24, 25). Although previous studies have reported a higher incidence of anastomotic stenosis and reflux esophagitis after PG, these complications are largely influenced by the type of reconstruction technique used (3, 26–28). This study did not evaluate reflux esophagitis, as its occurrence is well known to depend on the specific reconstruction technique used in PG. For other postoperative complications, available data in the literature remain relatively limited. In this study, the overall rate of postoperative complications was slightly lower in the PG group than in the TG group; however, the difference was not statistically significant. This suggests that although PG may help reduce the risk of certain complications, the benefit may be insufficient to achieve a statistically significant difference in a propensity score-matched cohort. Moreover, the incidence of severe complications (Clavien–Dindo grade≥II) was also lower in the PG group, possibly reflecting the less invasive nature of the procedure. Another notable finding was a significant reduction in hospital stay among PG patients, with an average decrease of 0.81 days. This aligns with the general view that PG, as a less extensive procedure, promotes faster recovery and shorter hospital stays (9, 28, 29). Shorter hospital stays benefit patients by lowering the risk of hospital-acquired infections and reducing overall healthcare costs (30). Faster recovery associated with PG may be especially beneficial for elderly patients or those with multiple comorbidities, who are more prone to delayed recovery after major surgeries such as TG.

Oncological safety should remain the primary consideration when choosing between PG and TG (17). TG is a more extensive procedure that ensures complete resection of potential residual tumor at the gastric margin, often resulting in a lower recurrence rate than PG (1). However, previous studies have reported no significant difference in survival outcomes between PG and TG for early-stage proximal gastric cancer (1, 9, 28), which aligns with our findings in LAPGC. Khalayleh et al. further recommended PG for differentiated cT1–T3N0/N1 or poorly differentiated cT1N0/N1 gastric cancers smaller than 4 cm (31). Beyond tumor size, the extent of lymph node metastasis is widely recognized as a key prognostic factor for survival outcomes (32–34). Subgroup analyses of OS by ypN stage have shown that higher ypN stages are associated with worse survival, although multivariate analysis revealed no significant differences (13). Evaluating lymph node metastasis rates at stations #4d, #5, #6, and #12a—which are typically dissected in TG but not in PG—can help clarify this concern (35). Yura et al. (36) reported minimal metastasis rates at these stations in patients with T2/T3 proximal gastric cancer: 0.99%, 0%, 0%, and 0.006%, respectively. Similarly, Sasako et al. (37) found that stations #5 and #6 had the lowest therapeutic indices in upper-third tumors compared to those in the middle or lower third. Ooki et al. (38) reported nodal metastasis rates of 3.7%, 2.4%, and 0% at stations #4d, #5, and #6 in T3 proximal cancers. Similarly, Haruta et al. (39) found rates of 3.3%, 0.5%, 1.6%, and 0% at the same stations in proximal cancers. According to Niihara et al. (40), in early proximal gastric cancer, lymph nodes along the left gastric artery (#1, #3, and #7) serve as primary drainage sites, while those along the right gastric artery (#5, #8, and #12) and right gastroepiploic arteries (#4d and #6) are more distant and rarely involved in early metastasis. These findings suggest that prophylactic distal lymphadenectomy may have limited clinical value. Moreover, a recent multivariate analysis by Yang et al. (41) revealed that tumors ≥4 cm and metastases to stations #4, #7, #8, or #9 are independent risk factors for involvement of stations #5 and/or #6. Since these distal stations are typically excluded in PG, such findings underscore the potential risk of residual nodal disease when PG is applied to patients with high-risk profiles. Thus, in cases of advanced proximal tumors with suspected extended nodal spread, TG with standard D2 lymphadenectomy may remain the more appropriate oncologic option. Lymph node dissection was another important aspect evaluated in this meta-analysis. PG was associated with a significantly lower number of harvested lymph nodes compared to TG. This is expected, as TG generally includes more extensive lymphadenectomy, particularly in the perigastric and para-aortic regions. Nevertheless, despite fewer lymph nodes being dissected in PG, the metastatic rates were comparable between the two groups, suggesting that PG does not compromise oncologic outcomes related to lymph node involvement. These findings are consistent with previous studies showing that limited lymphadenectomy in PG may still provide sufficient oncologic control, especially when combined with adjuvant therapy (42). In this context, it is important to acknowledge the controversial role of extended lymphadenectomy, particularly para-aortic lymph node dissection (PAND), in the treatment of LAPGC following neoadjuvant chemotherapy. Although PAND is not routinely recommended, selective studies have reported that it may offer survival benefit in patients with radiologically suspected para-aortic node metastases who respond well to preoperative chemotherapy. However, given the technical demands, increased perioperative morbidity, and limited evidence supporting routine use, PAND is rarely performed, especially in conjunction with proximal gastrectomy. Its feasibility is mostly confined to highly selected patients in high-volume centers with experienced surgical teams (43). Adjuvant chemotherapy has been proven to improve survival, particularly in advanced gastric cancer, by targeting micrometastases and lowering recurrence risk. This highlights the importance of integrating surgical resection with chemotherapy to optimize long-term outcomes, regardless of the surgical approach.

Despite the differences in surgical techniques and perioperative outcomes, both PG and TG demonstrated comparable long-term oncologic outcomes, as reflected in 5-year OS and RFS. Although PG entails a less extensive resection, it can achieve oncologic results equivalent to those of TG in patients with LAPGC. Several factors may explain the lack of significant differences in survival outcomes. First, the high rate of adjuvant chemotherapy administration—without significant difference between the PG and TG groups—indicates that postoperative treatment protocols were largely comparable, potentially offsetting any differences in surgical extent. Second, although PG involves more limited lymphadenectomy, the similar rates of metastatic lymph nodes in both groups suggest that oncologic control is not substantially compromised. Furthermore, the established role of chemotherapy in managing advanced gastric cancer likely mitigates the potential drawbacks of less extensive surgery. Sensitivity analyses confirmed the robustness of these findings, showing no material change in results upon exclusion of any single study. This enhances the reliability of the conclusion that PG is not inferior to TG in terms of survival outcomes. Additionally, the absence of significant publication bias, as demonstrated by funnel and L’Abbé plots, further reinforces the validity of the results and suggests minimal influence from selective reporting or unpublished data.

Quality of life (QoL) has been an important factor in comparing PG and TG, particularly in early-stage proximal gastric cancer. Most studies suggest that PG is more beneficial in maintaining nutritional status, as it is a function-preserving procedure that leads to less postoperative weight loss and better overall nutrition, primarily due to the preservation of the gastric fundic gland region (9, 11, 44). However, there is a lack of research on QoL in patients with locally advanced gastric cancer. Therefore, future studies should focus on examining QoL outcomes in this patient group, especially in the context of comparing PG and TG in LAPGC.

Nevertheless, several limitations of this study should be acknowledged. First, although PSM improves group comparability, it cannot eliminate all biases inherent in retrospective studies. Thus, our findings should be interpreted as hypothesis-generating rather than definitive evidence. Second, we used the GRADE framework to assess key outcomes such as OS and RFS, which were rated as low-certainty due to study design and imprecision. Third, very few published reports provided outcome data stratified by minimally invasive surgical approach (laparoscopic vs. robotic) or by receipt of neoadjuvant therapy. Hence, the prespecified subgroup analyses could not be performed, leaving the influence of these factors on perioperative and oncologic end points uncertain. These findings underscore the importance of more detailed and standardized data reporting in future research, particularly in multi-institutional studies and real-world datasets.

In conclusion, PG appears to be a feasible alternative to TG, potentially providing better short-term outcomes without compromising long-term survival in patients with LAPGC. However, large-scale, multicenter, prospective randomized controlled trials are needed to confirm these findings and provide more definitive guidance for clinical decision-making.