Abstract
OBJECTIVE: The aim of this study was to investigate the value of early postoperative ctDNA dynamic monitoring (TP53/KRAS/PIK3CA gene mutations) combined with clinicopathologic factors in risk stratification of postoperative peritoneal metastasis in advanced gastric cancer. METHODS: One hundred and twenty-eight patients with advanced gastric cancer (stage II/III) who underwent R0 resection were enrolled, of whom 32 developed peritoneal metastasis after surgery (peritoneal metastasis group) and 96 did not (non-peritoneal metastasis group). The study collected peripheral blood samples from patients at 7 days (T0) and 4 weeks (T1) after surgery, detected mutations in ctDNA at three loci, TP53 p.R248W, KRAS p.G12D, and PIK3CA p.H1047R, by ddPCR and simultaneously collected clinicopathological features such as Lauren classification and serosal invasion. The primary endpoint was peritoneal metastatic events confirmed by imaging or pathology. RESULTS: The proportion of Lauren diffuse-type peritoneal metastases (65.6% vs. 34.4%, P = 0.002), serosal invasion (87.5% vs. 65.6%, P = 0.018), and persistent ctDNA positivity (53.1% vs. 19.8%, P < 0.001) were significantly higher. The Boruta algorithm identified Lauren diffuse type, serosal invasion, and persistent ctDNA positivity as key variables. The combined model (Model 1) constructed based on these variables demonstrated optimal performance (C-index = 0.853) and significantly outperformed the traditional clinical model (C-index = 0.840). A nomogram developed from this model enables personalized risk prediction. Based on the number of risk features (≥ 2 features in the high-risk group and ≤ 1 feature in the low-risk group), Kaplan-Meier analysis showed that the cumulative incidence of peritoneal metastasis was significantly higher in the high-risk group (51 cases) than in the low-risk group (77 cases) (Log-rank P = 0.0012). CONCLUSION: Early postoperative persistent ctDNA positivity (especially TP53 mutation) combined with Lauren diffuse-type classification and serosal invasion can effectively identify patients at high risk of peritoneal metastasis. This integrated model provides a new strategy for precise postoperative surveillance and intervention in advanced gastric cancer.