Influence of the initial recurrence and metastatic site on outcomes in postoperative gastric cancer patients

Purpose To evaluate the impact of initial recurrence and metastatic sites on survival outcomes in gastric cancer patients after radical gastrectomy. Methods A retrospective study was conducted on 1,140 patients with pathologically confirmed gastric cancer who underwent radical gastrectomy between September 2016 and December 2023. Patients were categorized into six groups according to initial recurrence and recurrence site: no recurrence, local recurrence, liver metastasis, peritoneal/retroperitoneal metastasis, multiple-site peritoneal metastasis, and other sites. Overall survival (OS) and Post-Recurrence Survival (PRS) were analyzed using Kaplan–Meier and log-ran k tests. Independent prognostic factors were identified by Cox regression. Results Age ≥60 years, CEA ≥5U/ml, pathological type (other), T4 Stage, perineural invasion and lymphovascular invasion were independent risk factors for poor OS, while middle-third and lower-third gastric location were protective factors. As for DFS, tumor differentiation, lymph node metastasis, perineural invasion, and site of recurrence/metastasis were significantly associated with the PRS. In the multivariable analysis, only perineural invasion remained an independent risk factor for PRS ( HR = 1.588, P = 0.005). Among 222 patients with recurrence, patients with Multiple-site peritoneal metastasis had the numerically shortest median OS (20 months, 95% CI 17.040–22.960) among the five recurrence subgroups. No statistically significant between-group differences were identified after Bonferroni correction for 10 pairwise log-rank comparisons. PRS was longest in liver metastasis patients (10 months), significantly longer than multiple-site peritoneal metastasis (3 months, P < 0.001), and patients with peritoneal/retroperitoneal metastasis had longer PRS than those with multiple-site peritoneal metastases (5 vs. 3 months, P = 0.028). Comprehensive therapy was associated with significantly longer PRS compared with supportive care (9 vs. 3 months, P < 0.001) and chemotherapy alone (9 vs. 4 months, P = 0.002). Conclusion PRS differed among patients with different initial recurrence/metastatic patterns. However, the initial recurrence/metastatic site was not independently associated with PRS. Perineural invasion was independently associated with shorter PRS and OS and may serve as a reference factor for post-recurrence prognostic assessment. Patients receiving comprehensive therapy after recurrence showed longer PRS.