基于常规指标的肺癌及胃癌早期诊断模型

Anearly diagnosis model of lung cancer and gastric cancer based on conventional indicators

目的:通过化学计量学模型评估分析肺癌、胃癌患者与健康体检者血常规、生化指标的区别,为肿瘤早期诊断提供新方法。方法:采集2014年1月到2015年12月无锡市第四人民医院肺癌(72例)、胃癌患者(57例)和健康体检者(50例)空腹血。应用SIMCA-P 13.0对患者血常规及生化指标进行主成分分析(Principal component analysis,PCA)、正交-偏最小二乘回归(Orthogonal partial least-squares regression-discriminate analysis,OPLS-DA)分析,并利用OPLS-DA模型的变量投影重要性(Variable importance in projection,VIP)参数筛选组间差异指标。结果:对于肺癌和胃癌患者,有监督的PLS-DA模型较稳定(Q20.395),根据分组信息构建模型,组间差异较明显。对于肺癌组和对照组以及胃癌组和对照组,有监督的OPLS-DA模型较稳定(Q20.406和Q20.501),组间差异明显。结论:肿瘤伴有血常规、生化指标不同程度的变化,肺癌和胃癌患者与对照组可通过有监督的OPLS-DA模型进行辅助诊断以及鉴别诊断,为肿瘤早期诊断提新方法。

Objective:To analyze the levels and diagnostic value of detecting indicators of blood of cancer patients to develop new diagnostic models for lung cancer and gastric cancer.Methods:Plasma of cancer patients and healthy volunteers were collected under fasting from the affiliated hospital of Jiangnan university during 2014 and 2015.Data about blood were analyzed by SIMCA-P 13.0 with PCA and OPLS-DA.Inter-group differences were further analyzed by Q2 test which was calculated by OPLS-DA.Results:The stability between lung cancer group and gastric cancer group was much better by supervised OPLS-DA model,of which the Q2 was 0.395.According to the group information,models wereconstructed with obvious differences.Similarresults were drawn separately between lung cancer group,gastric cancer group and the control group,of which the Q2 were 0.406 and 0.501,respectively.Conclusion:The supervised OPLS-DA model may help to diagnose lung cancer and gastric cancer,which may provid theoretical basis for early diagnosis of tumor.