结直肠癌个体危险度及人群筛检数量化评价的研究

Quantitative assessment of individual risk and population screening for colon and rectal cancer

目的 尝试按结肠癌、直肠癌分别建立可作为第一线初筛方法,包含环境暴露和遗传因素(遗传多态性)的大肠癌危险度数量化评估模型。方法 在以随访队列为基础的病例对照研究的基础上,分别选择在单因素统计分析中,对结肠癌、直肠癌有显著意义的若干因素(危险因素和保护因素),应用模糊数学和隶属函数等原理和方法分别对结肠癌、直肠癌危险因素和保护因素数量化,尝试建立包含内外环境因素的结肠癌、直肠癌危险度筛检数量化评价方法(数学模型),并进行数据回代和模型评价。结果 结肠癌、直肠癌的AD阈值分别设定为0.13和0.19时,结肠癌的灵敏度和特异度分别为82.2%和73.3%,直肠癌的灵敏度和特异度分别为73.1%和78.0%。用ROC曲线比较只包含环境暴露因素的环境模型和包含内外环境的全模型判别结肠癌、直肠癌的诊断价值,判别结肠癌的曲线下面积分别为0.851和0.845,无显著性差异;判别直肠癌的曲线下面积分别为0.828和0.824,也无统计学差异。结论 分别按结肠癌、直肠癌建立的危险度评估模型诊断价值都较高,包含内外环境的结直肠癌危险度评估的数量化模型的标准因素群得到了进一步优化。

Objective To set up and evaluate two quantitative assessment models including environmental factors and polymorphisms of metabolic enzymes, one for the primary screening of colon cancer and the other for rectal cancer. Methods Based on some principles and methods of fuzzy mathematical and membership function, risk and protective factors of colon cancer and of rectal cancer including environmental exposure and genetic polymorphisms from a population-based case-control study were quantified. Accordingly, two quantitative assessment models were set up. The models were applied to the general population for evaluateing the efficacies. Results The quantitative models for colon cancer and rectal cancer related to different factors of genetic polymorphisms and environmental exposure. When the AD value of the mathematical model increased, the sensitivities of coloncancer and of rectal cancer screening were increased and the specificities were decreased. Considering the satisfaction of both the sensitivities and specificities, the thresholds of AD values of population screening for colon cancer and rectal cancer were configured at 0.13 and 0.19, respectively. The sensitivity and specificity for colon cancer were 82.2% and 73.3%, respectively, while those for rectal cancer were 73.1% and 78%. The area under the ROC curve of the screening model for colon cancer related to both environmental exposure and genetic polymorphisms was 0.851, while that just related to environmental exposures was 0.845. The difference was no significant in statistics. The area under the ROC curve of rectal cancer related to both environmental exposures and genetic polymorphisms was 0.828, while that just related to environmental exposure was 0.824, also no significant difference in statistics. Conclusion When the colon and rectal cancers are considered separately, the validities of the actual identification of the quantitative assessment models are satisfactory. The groups of standard factors relate to the quantitative assessment models for colon and rectal cancer are improved.