Analysis of risk factors for bleeding after CT-guided percutaneous lung puncture biopsy and construction of a line graph model
10.3760/cma.j.cn115455-20220811-00718
- VernacularTitle:CT引导下经皮肺穿刺活检术后出血的危险因素分析及列线图模型构建
- Author:
Di FENG
1
;
Chengwei ZHOU
;
Haitao WANG
;
Jinfeng WEN
;
Wuliang YU
Author Information
1. 宁波大学医学院附属医院影像科，宁波 315020
- Keywords:
Hemorrhage;
CT-guided percutaneous lung biopsy;
Risk factor analysis;
Nomogram model
- From:
Chinese Journal of Postgraduates of Medicine
2023;46(10):876-880
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To analyze the risk factors of postoperative bleeding after CT-guided percutaneous fly biopsy, and to construct a nomogram model for predicting postoperative bleeding.Methods:A total of 328 patients with CT-guided percutaneous fly biopsy who were admitted to the Affiliated Hospital of Ningbo University School of Medicine from June 2019 to March 2021 were selected, and the general clinical data of the patients were retrospectively analyzed. The patients were divided into observation group and control group according to postoperative bleeding. Statistical analysis was performed on clinical data of patients with statistical significance, receiver operating characteristic (ROC) curve analysis on continuous variables with statistical significance, and Logistic multiple regression analysis on all variables with statistical significance. The risk factors of bleeding complications in CT-guided percutaneous lung biopsy were obtained, and a Nomogram model for predicting bleeding after percutaneous lung biopsy was constructed using the R language software 4.0 "rms" package.Results:The results of the study showed that the size of the mass , the depth of the mass, the number of punctures, the number of cases with inflammation around the lesion , and the number of cases with chronic lung lesions in the observation group were greater than those in the control group: (5.20 ± 1.20) cm vs. (4.30 ± 0.90) cm, (4.30 ± 0.60) cm vs. (2.90 ± 0.30) cm, (2.70 ± 0.60) times vs. (2.30 ± 0.50) times, 38(14, 70) cases vs. 17(24, 60) cases, 102(39, 40) cases vs. 41(59, 40) cases; while the number of normal preoperative prothrombin time (PT) in the observation group was less than that in the control group: 4(5, 80) cases vs. 151(58, 30) cases ( P<0.05). ROC curve analysis was performed on the continuous variables with statistical significance in the control table of patients′general clinical data. The results showed that the area under the curve for swelling size, swelling depth, number of punctures, and number of tissue blocks cut were 0.563, 0.714, 0.680, and 0.559, respectively; the optimal cut-off values were 53.00 cm, 5.56 cm, 2.00 times, and 1.00 blocks ( P<0.05). The univariate indicators were included in the Logistic multiple regression model, and the results showed that tumor depth, puncture times, inflammation around the lesion, and abnormal preoperative PT were the risk factors for complications of percutaneous lung biopsy under CT ( P<0.05). The internal validation results showed that the Nomogram model predicted the risk of bleeding complicated by percutaneous lung biopsy under CT, with a C index of 0.687 (95% CI 0.241 - 1.988). The calibration curve showed good agreement between the observed and predicted values. The Nomogram model predicted percutaneous lung biopsy under CT with a bleeding risk threshold of >0.16, and the Nomogram model provided a clinical net benefit; in addition, the Nomogram model had a higher clinical net benefit than independent indicators. Conclusions:In conclusion, patients with poor coagulation function, inflammation around tumor lesions, deeper lesions, and more puncture times are more prone to bleeding. The Nomogram model constructed in this study has a high clinical application value for predicting the bleeding complications of CT-guided percutaneous lung biopsy.
