胃癌由于早期症状不典型且缺乏特异性，检出率相对较低，导致许多患者在初次就诊时就已处于进展期。而进展期胃癌患者的生存预后总体较差且差异较大，5年生存率为20%~60%。因此，需要制定更加个性化的治疗策略来提高生存率，缩小不同患者的预后差异。建立准确、可靠又方便的胃癌预后评估模型是当前胃癌诊治中亟待解决的重要问题之一。Cox回归分析模型能够分析多种因素对患者预后的影响；列线图模型能够通过计算不同预后因素的得分，直观地显示不同预后因素对患者预后的贡献程度；神经网络模型可以自动学习，深度学习模型可以自动提取特征，建立高效的预测模型，适用于大规模数据的胃癌预后预测。诸如以上预测模型可以区分胃癌患者预后的危险程度，有望为个性化治疗方案和后续诊治提供参考，在一定程度上改善患者的预后生存。

Due to the atypical symptoms and lack of specificity, the detection rate of early gastric cancer is relatively low, leading to many patients already in advanced stage since first diagnosis. However, the prognosis of patients with advanced gastric cancer is generally poor and varies greatly, with the 5-year survival rate hovering between 20% and 60%. Therefore, it is necessary to develop more personalized treatment strategies to improve the survival rate of patients and reduce the difference in prognosis of different patients. The establishment of reliable models for gastric cancer is one of the important problems to be solved urgently in the diagnosis and treatment of gastric cancer. Cox regression analysis model can analyze the impact of multiple factors on the prognosis of patients; nomogram model can directly display the contribution of different prognostic factors to the prognosis of patients by calculating the scores of different prognostic factors; the artificial neural network model can be automatically learned, and the deep learning model can automatically extract features to establish an efficient prediction model, which are applicable to the prognosis prediction of gastric cancer with large-scale data. The above predictive models may be used to differentiate the risk of gastric cancer, which is expected to provide personalized reference for follow-up treatment for patients, and improve the survival of patients to some extent.