Page 118 - 南京医科大学自然版

南京医科大学学报（自然科学版） 第46卷第7期
·1064 · Journal of Nanjing Medical University（Natural Sciences） 2026年7月

·临床研究·

小细胞肺癌患者化疗期间肺部感染风险评分模型的构建与验证

张 梦，汤 婷，潘志娟，朱金星，刘扣英 *
南京医科大学第一附属医院呼吸与危重症医学科，江苏 南京 210029

［摘 要］ 目的：探究小细胞肺癌（small cell lung cancer，SCLC）患者化疗期间肺部感染的危险因素，建立风险预测模型并验
证。方法：回顾性纳入2020年4月—2022年3月在南京医科大学第一附属医院接受化疗的SCLC患者作为训练集（251例），同
期前瞻性连续纳入SCLC患者作为验证集（112例）。根据随访结果分为肺部感染组和非肺部感染组，在训练集采用单因素及
多因素Logistic回归分析筛选独立危险因素并构建评分模型，采用受试者工作特征（receiver operating characteristic，ROC）曲线
评价模型区分度，校准曲线评估模型一致性，并与既往预测模型进行比较，在验证集中对模型进行外部验证。结果：多因
素 Logistic回归分析显示，吸烟史、胸腔积液、声音嘶哑、单药化疗方案及首次化疗后白蛋白<35 g/L是SCLC患者化疗期间肺部
感染的独立危险因素（P均< 0.05）。基于上述5个变量建立SCLC⁃PIR评分模型。在训练集中应用此模型的ROC曲线下面积
（area under the curve，AUC）为0.870（95% CI：0.818~0.922），最佳截断值为5分，对应的灵敏度为71.7%，特异度为89.4%。校准
曲线显示此模型预测的感染风险与实际风险一致性良好。在验证集中应用此模型的预测效能保持稳定，AUC 为 0.896（95%
CI：0.832~0.961），且此模型的预测效能高于既往列线图模型。结论：基于吸烟史、胸腔积液、声音嘶哑、单药化疗方案及化疗后
白蛋白<35 g/L构建的SCLC⁃PIR评分模型具有良好的区分度和一致性，能够有效预测SCLC患者化疗期间肺部感染风险，可为
早期识别高风险患者提供参考。
［关键词］ 小细胞肺癌；肺部感染；风险预测；评分模型
［中图分类号］ R734.2 ［文献标志码］ A ［文章编号］ 1007⁃4368（2026）07⁃1064⁃09
doi： 10.7655/NYDXBNSN260404

Development and validation of a risk scoring model for pulmonary infection during
chemotherapy in patients with small cell lung cancer

ZHANG Meng，TANG Ting，PAN Zhijuan，ZHU Jinxing，LIU Kouying *
Department of Respiratory and Critical Care Medicine，the First Affiliated Hospital of Nanjing Medical University，
Nanjing 210029，China

［Abstract］ Objective： To investigate the risk factors for pulmonary infection during chemotherapy in patients with small cell lung
cancer（SCLC）and to develop and validate a risk prediction model. Methods：Patients with SCLC who received chemotherapy at the
First Affiliated Hospital of Nanjing Medical University from April 2020 to March 2022 were retrospectively enrolled as the training
cohort（n=251），and a prospective cohort of SCLC patients was consecutively included as the validation cohort（n=112）. According to
follow⁃up outcomes，patients were divided into the pulmonary infection and non⁃infection groups. In the training cohort，univariate and
multivariate logistic regression analyses were performed to identify independent risk factors and to construct a scoring model. The
discriminative ability of the model was evaluated using the receiver operating characteristic（ROC）curve，and calibration was assessed
using calibration curves. The model was compared with previously reported nomogram models and further validated in the validation
cohort. Results：Multivariate logistic regression analysis showed that smoking history，pleural effusion，hoarseness，single ⁃ agent
chemotherapy，and albumin <35 g/L after the first cycle of chemotherapy were independent risk factorsfor pulmonary infection in
patients with SCLC（all P < 0.05）. Based on these five variables，the SCLC⁃PIR scoring model was established. In the training cohort，
the area under the curve（AUC）was 0.870（95% CI：0.818-0.922），with an optimal cutoff value of 5 points，yielding a sensitivity of

［基金项目］ 江苏省人民医院“临床能力提升工程”护理项目（JSPH⁃NC⁃2021⁃17）
通信作者（Corresponding author），E⁃mail： liuky188@126.com（ORCID：0000⁃0002⁃4135⁃314X）
∗