中文摘要：

卵巢癌是目前死亡率最高的妇科疾病之一,利用信息学手段挑选特征肿瘤标志物已被广泛用于包括卵巢癌在内的肿瘤分类、诊断研究。但是研究中单纯以提高分类率为指标而忽视敏感性和特异性的均衡,且模型为多变量或者复杂模型,成本过高,不太适合临床应用。为此,提出一种基于＂极少＂特征标志物的两步预测模型,利用先期提取的多个特征作敏感性和特异性测试,然后构建特征变量的两步预测模型。先用单个变量预测,在一个变量不能得到可靠结果时,才增加另一变量参与模型。实验显示,筛选出的PPE8＋LPE4和PPE8＋LPC0两对变量组合的敏感性和特异性显著、均衡,变量之间的相关性较小,且分类结果和4个变量的分类结果相当,与9个变量的分类率只差4%～5%。所提出的基于极少特征标志物的两步预测模型结构简单,在保持相同分类效果的前提下大大减少了用于预测的变量,为实际应用提供方便,同时在一定程度上节约了经济成本。

英文摘要：

Ovarian Carcinoma（OvCa） is the most lethal type of gynecological cancer.Now biomarkers selection methods are extensively used in OvCa classification and diagnosis.However many studies only focused on improving classification accuracies and ignored the balance of the sensitivity and specificity.Moreover,most of the built models include too many variables,which make the models clinically hard to apply.To this end,this paper presented a novel simple two-stage model based on minimal number of biomarkers.Firstly,the candidate biomarkers were examined in terms of the sensitivity and specificity.Then we built a two-stage prediction model which included only two variables,and instead of apply the two variables simultaneously,the second variable was applied only when necessary.Experimental study shows that the selected biomarkers such as PPE8 ＋ LPE4 and PPE8 ＋ LPC0 have balanced sensitivity and specificity.Statistical permutation test shows that the results based on our simple model are comparable to the complex four or more variables models.It maintains a high classification accuracy and uses less variables as well therefore is more cost effective.