周围型小细胞肺癌影像诊断模型的建立

Establishment of an imaging diagnosis model for peripheral small cell lung cancer

目的:建立周围型直径≤3 cm小细胞肺癌(SCLC)的影像诊断模型,提高对周围型SCLC和非小细胞肺癌(NSCLC)的鉴别诊断能力。方法:收集武汉大学中南医院2004年3月到2018年1月直径≤3 cm的周围型肺癌患者的影像及临床资料,以周围型SCLC为实验组,以周围型NSCLC为对照组,单因素分析比较实验组、对照组一般资料(年龄、性别及吸烟史等)、CT征象的差异性,将P<0.1的因素纳入多因素Logistic回归分析,探讨SCLC的诊断影响因素。使用多因素Logistic的β系数建立一个预测积分,结合年龄、肿瘤阴影消失率(TDR)和相对强化等连续性变量建立影像征象评分预测模型,并进行诊断模型价值分析。结果:通过建立诊断的影像评分预测模型,可以有效区分周围型SCLC和NSCLC。结论:利用建立的影像评分预测模型对于区分周围型SCLC和NSCLC诊断效能较高。

Objective:To establish a diagnosis model for peripheral small cell lung cancer(SCLC),and to improve the differential diagnostic ability for peripheral SCLC from non-small cell lung cancer(NSCLC) in tumor diameter of 3 cm or less.Methods:We collected the clinical data which included images of peripheral lung cancer cases(SCLC group and NSCLC group) with tumor diameter ≤3 cm in Zhongnan Hospital of Wuhan University from March 2004 to January 2018.Single factor analysis was done to compare SCLC group and NSCLC group(control group) for general information(age,gender,and smoking history,etc.) and CT features.The differential factors with P<0.1 were included into multivariate logistic regression analysis to explore the diagnostic factors of SCLC.Multivariable logistics was used with a beta coefficient of predicted points,combined with continuous variables like age,tumor disappearance rate(TDR),and relative strengthening imaging signs,to establih an imaging score prediction model,and then the model was evaluated.Results:The imaging score prediction model was established to effectively distinguish peripheral SCLC from NSCLC.Conclusion:The imaging scoring prediction model is effective in differentiating peripheral SCLC from NSCLC.