This study, utilizing a large-scale multicenter Eastern database, has established a Deep Learning-based predictive model for recurrence following gastric cancer surgery, which demonstrates robust discriminatory power for early recurrence. Furthermore, the individualized recurrence probability generated by this model can predict long-term postoperative prognosis and effectively stratify patients based on risk, thereby guiding personalized treatment choices. This individualized risk probability is also applicable to both adjuvant chemotherapy and neoadjuvant chemotherapy populations, offering valuable support for precision treatment in gastric cancer.

Inclusion Criteria:

Pathologically confirmed gastric adenocarcinoma; No distant metastases confirmed by preoperative examinations such as chest X-ray, abdominal ultrasonography, and upper abdominal computed tomography; Achievement of R0 resection.

Exclusion Criteria:

Presence of distant metastases detected preoperatively or intraoperatively; Prior neoadjuvant chemotherapy or radiotherapy; Incomplete general clinical data.

A retrospective analysis was conducted on the clinicopathological data of patients who underwent radical gastrectomy for gastric cancer between 2001 and 2022 at 13 tertiary hospitals in China.