Gastric cancer (GC) remains a major global public health burden, ranking as the fifth most commonly diagnosed malignancy and the fourth leading cause of cancer-related mortality worldwide. Approximately one million new cases are diagnosed each year, resulting in more than 650,000 deaths (1). With the accelerating trend of population aging, there has been a continuous increase in the number of elderly gastric cancer patients and their demand for diagnostic and therapeutic services. Epidemiological studies indicate that the median age at diagnosis of gastric cancer exceeds 70 years, and the aging population is projected to be the primary driver of rising incidence and mortality (2–4). Compared with younger patients, elderly patients more frequently present with multimorbidity, frailty, and malnutrition, resulting in a more complex risk-benefit profile for both surgical and systemic therapies. This underscores the urgent need for precise stratification and individualized management.

Neoadjuvant chemotherapy (NAC) has become a cornerstone perioperative strategy for locally advanced gastric cancer (LAGC), demonstrating the potential to improve R0 resection rates, achieve tumor downstaging, and confer survival benefits (5). Recent studies also suggest that NAC can improve oncological outcomes and long-term survival in elderly LAGC patients (6, 7). However, significant heterogeneity exists among elderly patients in terms of tolerance to NAC and the actual survival benefit derived from it. Previous research has predominantly focused on overall survival (OS) or pathological response as primary endpoints, with less systematic characterization of the risk of postoperative distant metastasis (DM) as a key outcome. In reality, DM is a critical event in gastric cancer mortality. Once it occurs, patient OS is significantly shortened, treatment strategies often shift to palliative intent, and clinical outcomes deteriorate rapidly. Epidemiological data indicate that approximately one-third of gastric cancer patients present with distant metastasis at initial diagnosis, with a dismal 5-year relative survival rate of only about 7%–8% and a median overall survival typically under one year (8). Common metastatic sites in gastric cancer include the liver, peritoneum, lungs, bones, and distant lymph nodes. Among patients with metastatic disease, liver involvement is reported in approximately 43%–48% of cases, peritoneal involvement in about 32%, pulmonary involvement in about 15%, and bone involvement in about 12%. Prognosis varies considerably depending on the metastatic site. Several recent studies indicate that median overall survival for peritoneal metastasis is typically only a few months to within a year, while median OS for liver metastasis is often approximately 10–12 months, although outcomes are significantly influenced by treatment modalities and metastatic burden (9–12). Therefore, identifying high-risk subgroups for DM among elderly gastric cancer patients who have undergone NAC and subsequent radical resection, and implementing personalized follow-up and intervention strategies at an early stage, represents a clear clinical necessity.

The nomogram has been widely adopted in recent years to evaluate cancer prognosis due to its convenience and precision, making it a suitable tool for our purpose (13). Accordingly, we conducted a retrospective study enrolling a consecutive cohort of elderly gastric cancer patients who underwent radical resection following NAC. This study aimed to determine the incidence, risk factors, and prognosis of postoperative distant metastasis in this population, and to develop two nomograms: one to predict the risk of postoperative DM, and another to predict OS among those who developed DM after NAC and radical surgery.

Gastric cancer imposes a substantial global disease burden due to its aggressive nature and poor prognosis, a concern that is particularly pronounced among the elderly population. However, high-quality evidence to guide treatment strategies in patients aged over 70 remains scarce. Importantly, therapeutic decisions in older adults should not rely solely on chronological age (15). Although existing studies indicate that the relative survival benefit from standard treatment is comparable between elderly and younger patients, older individuals face a higher risk of perioperative complications and therapy-related toxicities (16). Moreover, age-related physiological decline, comorbidities, and tumor-related factors—such as impaired gastric emptying, malabsorption, malnutrition, and sarcopenia—increase their susceptibility to dose reductions and cumulative toxicity during systemic anticancer treatment (17). For LAGC, NAC has become a cornerstone of perioperative management. Meta-analyses in Asian populations reveal that NAC significantly improves both disease-free survival and overall survival in elderly patients compared to surgery alone (18). Nonetheless, widespread adoption does not imply absence of risk. Marked interindividual variability exists among older patients in terms of tolerance and achievable dose intensity of NAC (2). Furthermore, even after completing NAC and achieving R0 resection, distant metastasis remains a leading cause of treatment failure, profoundly affecting long-term outcomes, with peritoneal and hepatic metastases being especially frequent (19). This underscores the urgent need to identify high-risk elderly subgroups predisposed to postoperative metastasis following NAC, which would allow tailored adjuvant strategies and intensified surveillance. Once DM develops, the absolute benefit that elderly patients gain from secondary surgeries, chemotherapy, or novel immunotherapies is often limited, and their prognosis tends to be worse than that of younger patients.

While several studies have explored prognostic predictors in gastric cancer, the majority have focused on the general population, with a paucity of dedicated analyses for elderly patients (20). Addressing this gap, our study developed and validated two specific tools for elderly patients with LAGC who received NAC followed by radical gastrectomy, a diagnostic nomogram for predicting the risk of postoperative DM and a prognostic nomogram for stratifying OS in those who developed DM. By incorporating readily accessible clinical variables, these nomograms generate individualized risk scores to inform decisions regarding postoperative surveillance intensity, consideration of intensified adjuvant therapy, and eligibility for clinical trials. Encouragingly, both models demonstrated robust and stable performance, with high AUC values, well-fitted calibration curves, and DCA indicating positive net clinical benefit across a wide range of risk thresholds. To our knowledge, this represents one of the largest studies specifically focused on predicting both “postoperative metastasis risk” and “post-metastasis survival” in this vulnerable population. Our models incorporate multidimensional variables-including laboratory indices, pathological grading, perioperative details, and chemotherapy-related data—collectively enhancing their general applicability and clinical utility.

NAC is widely considered a standard strategy for tumor downstaging and improving R0 resection rates in resectable LAGC. In our real-world elderly cohort, the specific NAC regimen or number of cycles did not show a significant independent impact on OS or DM. Previous research suggests that triplet regimens like docetaxel, oxaliplatin, and capecitabine DOX may offer superior 3-year OS and higher pathological complete response rates compared to doublets like XELOX, indicating potential biological advantage (21). However, in the clinical reality of elderly patients, this intensity benefit is often counterbalanced by tolerability issues and reduced treatment completion (22). Our finding that NAC-related adverse events independently correlated with increased DM risk underscores this trade-off. Treatment modifications (dose reduction, delay, or discontinuation) due to severe toxicity can compromise systemic control of micrometastases, increasing the risk of subsequent relapse. These findings underscore the importance of individualized dosing and proactive supportive care, including early nutritional and anti-sarcopenia interventions, primary prophylaxis with G-CSF for high-risk individuals, standardized antiemetic regimens, and comprehensive geriatric assessment (23). Ultimately, balancing regimen efficacy with personalized management is paramount, and overcoming chemoresistance and improving toxicity management remain critical challenges for future research.

The NLR, a readily available hematological marker reflecting systemic inflammation and immune suppression, emerged as a significant factor in both our DM risk and post-DM OS models, suggesting its role throughout the disease continuum from recurrence to progression and death. This aligns with numerous studies linking elevated NLR to worse OS and PFS (24). Neutrophils and tumor-associated neutrophils can promote angiogenesis, migration, and epithelial-mesenchymal transition through the secretion of factors like VEGF, MMP-9, and IL-17 (25), with the IL-17A–JAK2/STAT3 axis directly implicated in gastric cancer cells (26). Additionally, neutrophil extracellular traps can induce DNA damage/genomic instability and foster tumor progression and angiogenesis. Gastric cancer-derived extracellular vesicles can further activate and polarize neutrophils via the HMGB1/TLR4/NF-κB pathway, amplifying these effects (27). Regarding serum tumor markers, CA19–9 and CA72–4 showed distinct associations in our models, with CA19–9 linked to “DM risk” and CA72–4 to “OS after DM,” suggesting a meaningful division of labor. Previous evidence indicates that CA19–9 often rises with increasing tumor burden/dissemination risk and is associated with a higher likelihood of liver/peritoneal metastases (28). CA72-4, in contrast, is a mucin-related marker with greater specificity for gastric tissue. Its levels are particularly elevated in advanced or diffuse-type diseases that exhibit serosal or peritoneal spread, which likely accounts for its stronger prognostic signal in the overall survival model during the metastatic phase (29). The TRG assesses tumor response following NAC. When combined with the ypN stage, it significantly improves prognostic discrimination, prompting the proposal of integrated staging systems that incorporate both factors to refine survival prediction in gastric cancer (30). Our study corroborates this, demonstrating a stepwise decline in survival with higher TRG and ypN stage. This is consistent with the biological rationale that chemotherapy-resistant clones are more likely to survive treatment, seed micrometastases, and ultimately lead to clinical recurrence. While NAC is designed to eradicate micrometastases preoperatively and reduce recurrence risk (31), literature suggests that non-responders derive little improvement in distant metastasis-free survival even from adjuvant chemotherapy, implying that resistant clones may have already “escaped” during the neoadjuvant phase (32). Furthermore, an R1 resection margin can be viewed as direct evidence of “local microscopic residual disease leading to reseeding,” explaining its adverse impact in both our DM risk and OS models. Previous data also consistently show shorter OS for R1 patients compared to R0 (33).

The protective effects observed for both intraoperative intraperitoneal chemotherapy and postoperative adjuvant chemotherapy in our models align with the clinical reality that peritoneal recurrence is a common failure pattern in gastric cancer (34). It is important to note that conclusions from prior randomized trials have been inconsistent, and significant variations exist in techniques, drug selection, and patient criteria across centers, contributing to heterogeneous overall evidence. However, the “protective signal” observed in our real-world data is noteworthy. For patients at high risk of peritoneal recurrence (e.g., T3/T4, poorly differentiated) but without established DM, intraoperative intraperitoneal chemotherapy might be considered a preventive strategy (35). Evidence from colorectal cancer suggests that prophylactic intraoperative intraperitoneal chemotherapy can reduce the risk of subsequent peritoneal metastases, providing a rationale for cross-cancer strategy translation, though high-quality prospective data specific to gastric cancer are still needed (36). Postoperative adjuvant chemotherapy remains a critical component, with standard regimens and full course completion being essential. While delays in initiating adjuvant therapy have been associated with poorer outcomes, subsequent full-course completion can partially mitigate this negative impact (37). Given that elderly patients are more prone to incomplete treatment due to declining performance status, comorbidities, or frailty, a pragmatic approach involving dose individualization and intensive supportive care is crucial to balance relative dose intensity against severe toxicity, thereby maximizing net clinical benefit.

In the overall survival model for the distant metastasis cohort, a prolonged interval between neoadjuvant chemotherapy completion and surgery was independently associated with worse survival. This suggests that the time interval itself may represent a modifiable risk factor—an often-overlooked form of “time toxicity.” Particularly for patients with a poor response to NAC, delaying surgery risks missing the optimal window for resection while the patient continues to endure the side effects of chemotherapy and potential functional decline. Therefore, the decision to extend this interval must be guided by individualized response assessment and multidisciplinary discussion, carefully weighing the potential for further tumor regression against the risks of disease progression and dissemination (38). It is critical to emphasize the current absence of high-quality prospective trials directly comparing standardized intervals (e.g., 4 versus 8 weeks). Current evidence is largely derived from retrospective studies or combined-modality trials, wherein it is notoriously difficult to isolate the effect of the interval from the underlying tumor biology and treatment response. Consequently, in clinical practice, the strategy of intentionally extending the surgery interval should be approached with caution and pursued only alongside careful, dynamic monitoring of tumor response and resectability.

This study has several limitations. First, the relatively limited cohort of elderly patients with distant metastases from locally advanced gastric adenocarcinoma (N = 307) may have contributed to potential errors in the model. Second, although our nomograms were constructed using a training set and validated internally, external validation in independent cohorts is necessary to confirm their generalizability and mitigate inherent biases. Third, while we included key clinical, pathological, and laboratory variables, several important prognostic factors were not systematically integrated into the model, namely molecular markers, radiomic features, and comprehensive geriatric assessments. Finally, as a retrospective analysis, using “regimen/cycles” as exposure variables may not fully capture the impact of actual relative dose intensity, treatment delays, and dose modifications, potentially leading to an underestimation of how treatment intensity and completion influence outcomes. Future research should focus on prospective validation and incorporate these additional biomarkers, radiomic features, and comprehensive geriatric assessments, along with more granular treatment data, to refine predictive models. We believe that such advancements will significantly contribute to personalized perioperative management strategies and improve long-term outcomes in elderly patients with locally advanced gastric cancer (LAGC).

This study addresses the critical clinical challenge of predicting postoperative distant metastasis risk and individualized prognosis after metastasis in elderly patients with locally advanced gastric cancer following neoadjuvant chemotherapy, for whom specialized predictive tools are currently lacking. Through a retrospective analysis of 896 patients over 70 years of age, the study developed and validated two novel nomograms specifically for this population: the first model successfully identified eight independent risk factors for postoperative distant metastasis, including N stage, NAC-related adverse events, and CA19–9 levels; the second model, focusing on the 307 patients who developed metastasis, identified six independent prognostic factors, such as CA72–4 levels, the interval from NAC to surgery, and tumor regression grade. Validation results demonstrated that both nomograms exhibit exceptional predictive accuracy and clinical applicability. These tools enable the early identification of high-risk patients for metastasis from the preoperative to postoperative phases, guiding intensified monitoring and adjuvant therapy, while also providing individualized prognosis assessment after metastasis confirmation. This research offers a practical and innovative solution for advancing precision clinical decision-making in the management of elderly gastric cancer patients.