基于LASSO回归的CT引导下经皮肺穿刺术中并发症预测模型以及护理干预

Intraoperative Complication Prediction Model for CT-Guided Percutaneous Lung Biopsy Based on LASSO Regression and Nursing Intervention

摘要: 肺部病变的早期精准诊断是改善患者预后的关键，CT引导下经皮肺穿刺活检术是肺部占位性病变定性诊断的核心微创技术，但肺出血、术后气胸等并发症无法全部避免。本研究旨在构建术后并发症风险预测模型，并制定针对性护理干预措施。收集2023年1月~2024年12月于我院接受CT引导下经皮肺穿刺活检术的231例患者临床资料，根据术后并发症发生情况分为并发症组(106例)与无并发症组(125例)。通过单因素χ²检验筛选差异有统计学意义的指标，采用LASSO回归进行特征降维，结合多元logistic回归构建并发症预测模型，利用ROC曲线验证模型效能，并提出个体化护理干预方案。结果显示，231例患者穿刺活检成功率为100%，并发症发生率为45.89%，其中气胸占比最高(55.66%)。单因素分析提示年龄、合并慢性阻塞性肺疾病、穿刺深度、病灶距胸壁距离、穿刺针停留时间、病灶位置与并发症发生相关(P < 0.05)；LASSO回归进一步筛选出5个核心变量；多元logistic回归证实合并慢性阻塞性肺疾病、穿刺深度、病灶距胸壁距离、穿刺针停留时间是并发症发生的独立危险因素。模型预测并发症的AUC为0.848，优于单一指标预测效能。基于危险因素制定的个体化护理干预措施，可通过术前评估、术中配合及术后监测降低并发症风险。结论：合并慢性阻塞性肺疾病、穿刺深度过大、病灶距胸壁过远、穿刺针停留时间过长是术后并发症的高危因素；基于LASSO回归构建的预测模型可有效识别高风险患者，针对性护理干预有助于提升术中安全性。

Abstract: Early accurate diagnosis of pulmonary lesions is crucial for improving patient prognosis. CT-guided percutaneous lung biopsy is a core minimally invasive technique for the qualitative diagnosis of space-occupying pulmonary lesions, but postoperative complications such as pneumothorax and pulmonary hemorrhage remain unavoidable. This study aimed to construct a risk prediction model for postoperative complications and develop targeted nursing intervention measures. Clinical data of 231 patients who underwent CT-guided percutaneous lung biopsy in our hospital from January 2023 to December 2024 were collected. Patients were divided into the complication group (106 cases) and non-complication group (125 cases) based on the occurrence of postoperative complications. Univariate chi-square test was used to screen indicators with statistically significant differences, LASSO regression was applied for feature dimensionality reduction, and multivariate logistic regression was combined to construct a complication prediction model. The model’s performance was verified using ROC curves, and an individualized nursing intervention plan was proposed. Results showed that the success rate of percutaneous biopsy was 100% among the 231 patients, with a complication rate of 45.89%, of which pneumothorax accounted for the highest proportion (55.66%). Univariate analysis indicated that age, comorbidity with chronic obstructive pulmonary disease (COPD), puncture depth, distance from the lesion to the chest wall, puncture needle retention time, and lesion location were associated with the occurrence of complications (P < 0.05). LASSO regression further screened 5 core variables, and multivariate logistic regression confirmed that comorbidity with COPD, puncture depth, distance from the lesion to the chest wall, and puncture needle retention time were independent risk factors for complications. The AUC of the model for predicting complications was 0.848, which was superior to the predictive performance of a single indicator. Individualized nursing intervention measures developed based on risk factors can reduce the risk of complications through preoperative assessment, intraoperative coordination, and postoperative monitoring. Conclusion: Comorbidity with COPD, excessive puncture depth, excessive distance from the lesion to the chest wall, and prolonged puncture needle retention time are high-risk factors for postoperative complications. The prediction model constructed based on LASSO regression can effectively identify high-risk patients, and targeted nursing interventions are helpful for improving surgical safety.