基于外周血circRNAs表达谱筛查早期食管鳞癌及发病风险预测模型的建立

Esophageal squamous cell carcinoma (ESCC) is a complex and malignant tumor, and is one of the major cancers threatening human health. The early detection is critical for the cancer treatment and the patient survival. Circulating circRNAs is very stable and shows relatively specific expression profile in different diseases, it has potential to serve as a new kind of screening biomarker for ESCC. In the previous study, a high-throughput circRNA microarray was used to identify dysregulated circRNAs in the plasma of five ESCC patients and five frequency-matched healthy controls. However, the specificity and sensitivity of these dysregulated circRNAs in the screening of ESCC and their precancerous lesions are unclear. In this project, the candidate circRNAs were validated blindly in the plasma of ESCC, dysplasia and healthy population by quantitative reverse transcription polymerase chain reaction (qRT-PCR) at a population-based case-control study in ESCC high-risk area of China. Then, a 2-step approach combined with logistic regression and random forest analysis was performed to determine the panel of predictors included in the final classification model. The performance of model was assessed by the area under receiver operating characteristics curve (AUC) with the cross-validation and external validation. At last, we will calculate age- and sex-specific 5-year ESCC absolute risk estimates based on the espression level of circRNAs and the other identifiable risk factors for ESCC. The ratio of the expected to the observed number (E/O) of incident esophageal cancer cases was calculated to assess overall calibration. We intend to provide a model based on the blood circRNAs, have the potential for future use in identifying individuals who may be considered for endoscopic screening and targeted prevention of ESCC.

食管鳞癌早期预警是其防治研究中的瓶颈问题。分子标志能反映疾病的早期生物学效应。外周血中circRNAs的差异表达与肿瘤发生风险相关，有望成为一种新的生物标志。项目组前期研究发现食管鳞癌血浆中有33个显著差异表达的circRNAs，其在食管鳞癌早期筛查的价值，目前尚不清楚。本研究基于上消化道癌筛查的前瞻性评价研究队列，采用人群为基础的病例对照研究设计，通过qRT-PCR技术盲法验证候选circRNAs在食管鳞癌、异型增生和健康对照人群血浆中的表达水平，探讨其与食管鳞癌及其癌前病变的关联；采用logistic回归和随机森林算法结合的两步法筛选变量，结合环境因素，构建筛查模型；用交叉验证和Bootstrap法联合ROC曲线评价筛查模型的准确性；用人群性别和年龄别食管癌发病率，考虑竞争风险评估个体发生食管癌的5年绝对风险，计算E/O比值评估模型的校准度。为食管癌早期筛查的生物标志研究提供科学依据。

食管鳞癌（ESCC）早期预警是其防治研究中的瓶颈问题。外周血中circRNAs的差异表达与肿瘤发生风险相关，有望成为新的生物标志物。本项目利用高通量circRNAs芯片对circRNAs差异表达谱进行初筛。采用人群为基础的病例对照研究设计，通过qRT-PCR技术验证3个候选circRNAs在120例ESCC患者、75例异型增生患者及120例健康对照者血浆中的表达水平。结果显示，与健康对照组相比，hsa_circ_0020929(P<0.001)，hsa_circ_0044922(P<0.001)在ESCC血浆中表达上调，hsa_circ_0001714(P<0.001)在ESCC血浆中表达下调。基于危险因素+circRNAs表达谱构建的ESCC筛查模型ROC曲线下面积为0.946(95%CI:0.910-0.971)；食管异型增生（ESD）筛查模型ROC曲线下面积为0.731(95%CI:0.663-0.792)。采用以人群为基础的病例对照研究设计，以ESCC患者267例作为病例组，健康人群1251例为对照组。通过logistic回归分析建立ESCC的相对风险模型，校正竞争风险计算个体发生ESCC的5年绝对风险。结果显示建立的ESCC相对风险模型的AUC为0.731(95%CI:0.708-0.753)。随着年龄的增长和危险因素个数的增加，男性5年内发生ESCC的绝对风险从0.006%升至31.738%，女性从0.007%升至15.108%。ESCC相对风险模型在队列人群中的AUC为0.727(95%CI:0.721-0.734)。根据队列人群的基线信息计算5年内发生ESCC的绝对风险，男性人群的绝对风险为0.006%-8.012%，女性为0.003%-5.114%，模型在验证人群中的O/E值为0.97(95%CI:0.78-1.22)。因此，基于流行病学影响因素的风险预测模型对于食管鳞癌具有一定的预测价值。血浆中存在circRNAs差异表达谱，能有效提高流行病学危险因素对于食管鳞癌的筛查价值，有望作为食管鳞癌早诊早治的分子标志物。