探讨预后营养指数(PNI)在食管癌放射治疗(简称放疗)患者预后中的意义,并构建相关列线图风险模型。 方法 选取 2018 年 3 月至 2019 年 10 月于阜阳市肿瘤医院行放疗的食管癌患者作为研究对象,根据首次放疗前 1 周内实验室 检查数据计算 PNI。 采用受试者操作特征曲线计算 PNI 的最佳截断值,并据此将患者分为低 PNI 组与高 PNI 组。 比较两组临 床特征。 采用单因素和 Cox 回归多因素分析食管癌放疗患者预后的影响因素,并据此建立列线图预测模型。 结果 PNI 最佳 截断值为 48. 03,低 PNI 组与高 PNI 组的年龄、TNM 分期和肿瘤直径资料间差异有统计学意义(P<0. 05)。 TNM 分期高、肿瘤 直径≥3 cm、处方剂量<60 Gy、和 PNI<48. 03 是食管癌放疗患者死亡的独立危险因素(P<0. 05)。 据此建立预测食管癌放疗患 者死亡的列线图风险模型,模型验证结果显示 C-index 为 0. 846,校正曲线趋近于理想曲线,ROC 曲线的 AUC 为 0. 881 (95%CI = 0. 847~ 0. 912),表明模型具有良好的预测能力。 结论 PNI 对食管癌放疗患者的预后具有较高的预测价值,基于 PNI 构建的列线图模型能够有效预测食管癌放疗患者死亡的风险。

To explore the prognostic nutritional index PNI in patients with esophageal cancer undergoing radiotherapy and to establish an associated nomogram risk model. Method Patients with esophageal cancer who received radiotherapy in our hospital from March 2018 to October 2019 were selected as subjects. PNI was calculated according to the laboratory examination data within one week before the first radiotherapy. The best cut-off value of PNI was calculated by ROC curve and patients were divided into low PNI group and high PNI group. The clinical characteristics of the two groups were compared. Univariate and Cox regression were used to analyze the factors influencing the prognosis of patients with esophageal cancer radiotherapy. Result The optimal PNI cut-off value was 48. 03. There were statistically significant differences in age TNM stage and tumor diameter between the low PNI group and the high PNI group P<0. 05 . High TNM stage tumor diameter ≥3 cm prescription dose < 60 Gy and PNI < 48. 03 were independent risk factors for death in patients with esophageal cancer radiotherapy P < 0. 05 . The results of model verification showed that the C-index was 0. 846 the calibration curve was close to the ideal curve and the AUC of ROC curve was 0. 881 95%CI = 0. 847-0. 912 indicating that the model had good predictive ability. Conclusion PNI has a high predictive value for the prognosis of patients with esophageal cancer radiotherapy and the nomogram model constructed based on PNI can effectively predict the risk of death in patients with esophageal cancer radiotherapy.