白蛋白球蛋白比值、放疗剂量及同期放化疗联合预测鼻咽癌患者口腔真菌医院感染的价值

目的 探究白蛋白球蛋白比值（AGR）、放疗剂量及同期放化疗联合预测鼻咽癌（NPC）患者口腔真菌医院感染的临床价值。方法 回顾性分析2014年1月 - 2016年12月439例就诊鼻咽癌患者的资料，根据住院期间是否发生口腔真菌医院感染分为感染组与对照组。用单因素分析及Cox 比例风险模型分析影响口腔真菌医院感染可能性的因素。同时，通过建立ROC曲线分析确定最佳临界值及其预测发生口腔真菌医院感染的敏感度和特异性。结果 Cox回归分析显示△AGR水平（HR = 0.221，95%CI:0.086～0.565）、原发灶放疗剂量（HR = 5.330，95%CI:1.307～21.733）及同期放化疗（HR = 6.879，95%CI:3.988～11.865）的差异均有统计学意义。ROC显示三者联合时，AUC 为0.843。3个危险标记患者的感染率明显高于2个、1个危险标记和无危险标记的患者（P＜0. 001）。结论 △AGR（＜0.080）、原发灶放疗剂量（≥68.425Gy）及同期放化疗联合预测的效能较高，是预测NPC患者口腔真菌医院感染的较好指标。

Value of combining albumin-to-globulin ratio,dose of radiotherapy and concurrent radio-chemotherapy for predicting oral fungal nosocomial infection in nasopharyngeal carcinoma patients

Objective To explore the prognostic ability of the albumin-to-globulin ratio (AGR), dose of radiotherapy and concurrent radio-chemotherapy for oral fungal nosocomial infection in nasopharyngeal carcinoma (NPC) patients. Methods 439 patients diagnosed as NPC were enrolled from January 2014 to December 2016 in our hospital. Measurement was made and recorded. The patients were divided into the infection group and the control group according to whether they were infected. Univariate analysis and Cox model were used to show that the risk factors influencing the oral fungal nosocomial infection. ROC curves were applied to establish optimal cutoff points and the statistical significance factors were calculated. Results Cox regression analysis showed that the difference of △AGR (HR=0.221, 95%CI: 0.086-0.565), dose of radiotherapy to primary lesion (HR=5.330, 95%CI: 1.307-21.733) and concurrent radio-chemotherapy (HR=6.879, 95%CI: 3.988-11.865) were statistically significant. The ROC analysis showed that the AUC of the combination was 0.843. NPC patients with 3 markers have a significantly higher infection rate than the patients with 2 markers or with 1 marker or with no markers (P<0.001). Conclusion The combinations of △AGR (<0.080), dose of radiotherapy to primary lesion (≥68.425Gy) and concurrent radio-chemotherapy are good indicators for predicting oral fungal nosocomial infection in NPC patients.