基于SEER数据库构建肺肉瘤样癌的生存预测模型

目的 分析影响肺肉瘤样癌（PSC）患者预后的因素，构建PSC患者预后列线图预测模型。方法 基于SEER数据库收集1988—2015年间诊断为PSC患者1671例，按照7:3的比例分为建模组和验模组。对建模组患者进行单因素和多因素Cox回归分析影响PSC患者预后的独立因素并构建列线图预测模型，通过一致性指数和校准曲线分别在建模组和验模组进行验证。结果 单因素和多因素分析年龄、性别、组织学类型、TNM分期、肿瘤直径>50 mm、手术、放疗和化疗都是影响PSC患者预后的独立因素。基于独立因素构建列线图预测模型并进行验证。建模组和验模组一致性指数分别为0.790（95%CI: 0.776~0.804）和0.781（95%CI: 0.759~0.803）。建模组和验模组的校准曲线提示预测生存率与实际生存率基本一致。结论 基于多因素分析结果构建的列线图预测模型可预测PSC患者的预后，并且具有较高的准确性和一致性。

A Survival Prediction Model of Pulmonary Sarcomatoid Carcinoma Based on SEER Database

Objective To analyze the factors affecting the prognosis of patients with pulmonary sarcomatoid carcinoma (PSC) and construct a nomogram prediction model for the prognosis of PSC patients. Methods Based on the SEER database, 1671 patients diagnosed as PSC from 1988 to 2015 were collected and divided into modeling group and validation group according to the ratio of 7:3. Univariate and multivariate Cox regression analysis were performed in the modeling group to explore independent risk factors affecting the prognosis and construct a nomogram survival prediction model. The consistency index and calibration curve were used for verification in the modeling group and the test module respectively. Results Age, gender, histological type, TNM stage, tumor diameter >50mm, surgery, radiotherapy and chemotherapy were independent factors that affected the prognosis of PSC patients. The nomogram prediction model was constructed and verified based on independent factors. The C indexes of the modeling group and the test model were 0.790 (95%CI: 0.776-0.804) and 0.781 (95%CI: 0.759-0.803), respectively. The calibration curves of the modeling group and the test model indicated that the predicted survival rate was basically the same as the actual survival rate. Conclusion The nomogram prediction model constructed based on the results of multivariate analysis can predict the prognosis of PSC patients, and has high accuracy and consistency.