职业性肺癌高危人群筛检的聚类与判别分析

Cluster and discriminant analysis of screening for high risk individuals of the occupational pulmonary carcinoma

目的 :通过职业性肺癌高发人群多因素聚类分析并建立判别函数 ,判定职业人群中个体在已知类中的归类。方法 :用系统聚类法和逐步判别法对 2 7名肺癌高发人群的细胞水平变化、分子水平变化及其它好发因素进行分类并建立判别函数。结果 :共聚为 2类 ,3人被聚为肺癌高危个体类 ,2 4人被聚为肺癌发生危险度较低的另一类 ;经逐步判别 ,姐妹染色单体互换频率 (SCE)、P2 1进入了判别函数 ,分别建立了 2类的判别函数 :Y1=3.2 2 1×10 -3 ×P2 1+1.6 2 8×SCE - 14.70 9和Y2 =1.11× 10 -2 ×P2 1+0 .41×SCE - 44 .46。以此判别函数进行回代分析 ,表明回代率为 10 0 %。结论 :可以按多因素对肺癌高发人群进行聚类、判别分析 ,以数学模型对职业人群作出未来是否患癌的预报。

Aim:Multi factors were detected and analyzed in high risk workers of occupational pulmonary cancer in order to determine which of the known the individuals belong to.Methods:Multi factors including changes of cellular level and molecular level in 27 cases of hazard workers were detected, then what kind of the individuals was determined and established 2 kinds of discriminant function with the statistical methods of systematic cluster analysis and stepward discriminant analysis,respectively.Results: 2 kinds of the 27 individuals were clustered. 3 individuals were clustered as high risk ones to develop to lung cancer in future.24 cases were determined without risk of lung cancer. SCE, P21 of multi factors were chosen to step into the discriminant function after stepward analysis. 2 kinds discriminant functions were attained as below: Y 1=3.221×10 -3 ×P21+1.628×SCE-14.709 and Y 2=1.11×10 -2 ×P21+0.41×SCE-44.46.Revert test was done by using these functions and the correction rate was 100%. Conclusion: Cluster and discriminant analysis can be used to determine whether the high risk individuals develop to cancer in future, so mathematical model can be widely used in occupational workers.