Lung cancer prediction. Homo sapiens

We generated a blood-derived transcriptional signature that discriminates patients with lung cancer from non-affected smokers. When applied to blood samples from one of the largest prospective population-based cancer studies (the European Prospective Investigation into Cancer and Nutrition), this signature accurately predicted the occurrence of lung cancer in smokers within two years before the onset of clinical symptoms. Such a blood test could be used as a screening tool to enable early diagnosis of lung cancer at a curable stage. Overall design: All samples were taken from smokers. EPIC cases were taken from the EPIC trial (a large prospective epidemiological trial in Europe, and are incident cases), Cosmos and BC cases are prevalent cases from a Cologne, Germany based epidemiological trial.

我们构建了一种血液来源的转录特征（transcriptional signature），可用于区分肺癌患者与未患病吸烟者。将该特征应用于目前规模最大的基于人群的前瞻性癌症研究之一——欧洲癌症与营养前瞻性调查（European Prospective Investigation into Cancer and Nutrition，简称EPIC）的血液样本时，其可在临床症状出现前两年内，精准预测吸烟者的肺癌发生风险。此类血液检测可作为筛查工具，实现可治愈阶段肺癌的早期诊断。
整体实验设计：所有样本均取自吸烟者。EPIC队列样本取自EPIC研究（一项在欧洲开展的大型前瞻性流行病学研究），均为新发病例；Cosmos队列与乳腺癌（breast cancer，BC）病例则取自德国科隆的一项流行病学研究中的现患病例。