Objective  To explore factors affecting the prognosis of poorly differentiated and undifferentiated gastric cancer and to construct a prognostic nomogram model on survival prediction of the cancer patients for providing a reference to the assessment on prognosis of gastric cancer patients.

  Methods  The data on National Cancer Institute (NCI)-registered 9 223 patients with poorly differentiated and undifferentiated gastric cancer were extracted from the 2000 – 2018 database of Surveillance, Epidemiology, and End Results (SEER) program with SEER*Stat 8.3.9.2. Multivariate Cox proportional hazards regression model was adopted to explore main factors influencing the prognosis of the patients and nomogram diagram was constructed to evaluate main factors′ significance in prognosis prediction combined with correction curve.

  Results  For all the patients, the average survival time was 31.05 ± 29.12 months; the overall survival rate was 29.3% and the 1-, 3-, and 5-year survival rate were 62.4%, 36.4%, and 18.6%, respectively. The results of Cox model analysis revealed that aged ≥ 60 years, black race, II/III tumor stage, first/second/third node stage, tumor size of 5 – 10 centimeters, primary tumor at antrum/pylorus and other sites than cardia/fundus or corpus of stomach, and preoperative radiotherapy were associated with a poor prognosis; while, female, Asian pacific yellow race, and tumor types other than adenocarcinoma or signet ring cell carcinoma were associated with a better prognosis. The constructed nomogram diagram based on 6 influencing factors (age, tumor stage, node stage, overall tumor staging, surgery, and chemotherapy) showed prognostic predictions well consistent with those from the correction curves for 1-, 3-, and 5-year survival rate.

  Conclusion  The prognosis of patients with poorly differentiated and undifferentiated gastric cancer is mainly influenced by gender, age, race, node stage, overall tumor staging, tumor size, primary site, histological type, surgery, radiotherapy, and chemotherapy. The constructed nomogram diagram is of good prognostic value for predicting the patients′ survival.