分析肺癌患者发生上肢PICC相关性静脉血栓( PICC-catheter related thrombosis, PICC-CRT)危险因素，构建风险预测模型。

选取2019年5月至2022年5月北京市两家三级甲等医院肿瘤内科行PICC静脉化疗肺癌患者453例为对象，根据是否发生PICC相关性静脉血栓分为血栓组48例，非血栓组405例。收集两组临床资料，进行单因素分析筛选血栓危险因素，建立预测模型列线图，采用Bootstrap法进行模型内部验证，ROC曲线评价预测模型区分度，校准曲线评价模型校准度。

Logistic回归分析显示癌症分期、性别、置管前VTE风险评估、置管前活动能力评估、脑梗史、癌症转移史、D-二聚体水平是肺癌患者发生PICC相关性静脉血栓危险因素。ROC曲线分析显示曲线下面积为0.823〔95%CI(0.767，0.879)〕，最佳截断值为0.087，灵敏度为0.792，特异度为0.699，模型具有较好的区分能力和校准能力。

癌症晚期(Ⅲ~Ⅳ期)、性别男、VTE评估风险偏高、活动能力评估偏低、脑梗史、癌症转移史、D-二聚体水平偏高是肺癌患者PICC相关性静脉血栓危险因素，列线图模型具有较好的区分能力和校准能力。

To analyze the risk factors of PICC catheter-associated venous thrombosis in the upper limb of lung cancer patients after PICC catheterization, and to construct a preliminary risk prediction model presented in a visual graph.

A total of 453 patients with lung cancer who received PICC intravenous chemotherapy in the Department of Oncology of two third-class A hospitals in Beijing were selected from May 2019 to May 2022. According to whether the patients had PICC-associated venous thrombosis in half a year, they were divided into thrombe group(48 cases) and non-thrombe group(405 cases). The clinical data of the two groups of patients were collected, and R. 4.2.0 software was used for univariate analysis to screen the risk factors of thrombosis. Variables with P<0.05 were included in multivariate Logistic regression analysis to determine the independent risk factors of thrombosis. A line diagram of the prediction model was established, and the Bootstrap method was used to verify the model internally. The ROC curve was used to evaluate the differentiation of the prediction model, and the calibration curve was used to evaluate the calibration degree of the model.

The independent risk factors of PICC-associated venous thrombosis in patients with lung cancer were determined by Logistic regression analysis, including cancer stage, gender, pre-catheterization VTE risk assessment, pre-catheterization mobility assessment, history of cerebral infarction, history of cancer metastasis, and D-dimer level. The above risk factors were constructed by R software. The ROC curve analysis results showed that the area under the curve was 0.823 (95%CI: 0.767, 0.879), the optimal cut-off value was 0.087, the sensitivity was 0.792, and the specificity was 0.699. The model had good distinguishing ability and calibration ability.

Advanced cancer, male, high risk of VTE assessment, low mobility assessment, history of cerebral infarction, history of cancer metastasis, and high D-dimer level are independent risk factors for PICC-associated venous thrombosis in lung cancer patients. The line graph model built based on the above risk factors has a good ability to distinguish and calibrate.