直肠癌是一种常见的致死性疾病。如何治疗前有效预测局部晚期直肠癌新辅助放化疗疗效，以及分类筛选放化疗敏感性不同的患者并制定个体化的治疗方案是关键的临床问题。已有研究显示多参数磁共振放射组学量化分析可高准确性预测疗效，其内在原理可能为反映肿瘤微环境，但未被阐明。本研究将首次最大化的实现高质量多参数磁共振和肿瘤病理微环境在相同层面上对应匹配分析。通过数据深度学习，赋予放射组学参数具体肿瘤病理微环境内涵。通过放射组学与放化疗疗效相关性研究，把肿瘤微环境与放化疗结局进行串联，寻找影响治疗结局的微环境关键因子（如微血管、血管生成拟态、肿瘤相关成纤维细胞、细胞外基质纤维状态、肿瘤相关巨噬细胞等单一或组合因素）。同时进行组织RNA-seq检测，全转录组测序及细胞亚型分析，深入阐明肿瘤微环境影响放化疗结局可能机制 ，建立和优化预测模型，指导临床转化应用。

Rectal cancer is a common lethal disease. How to effectively predict the efficacy of neoadjuvant chemoradiation for locally advanced rectal cancer, as well as classify and screen the patients with different sensitivity of chemoradiation, and optimize individualized treatment options is the key issues in clinical practice. Current studies have shown that radiomics quantitative analysis of multi-parametric magnetic resonance imaging can predict efficacy with high accuracy. The intrinsic principle may be to reflect the tumor micro-environment, but it has not been elucidated. This study will maximize the matching correlation analysis of high-quality multi-parametric magnetic resonance imaging features and tumor pathological microenvironment at the same imaging anatomy level. The results would contribute for specificly understanding of the tumor microenvironment connotation reflected on the detail radiomics features through the deep learning of the data. Then through the correlation study between radiomics features and the tumor regression grading of chemoradiotherapy, we will get the result of the correlation between the tumor microenvironment and the chemoradiotherapy outcomes, so the key microenvironment factors (microvessel density, angiogenesis mimicry, tumor associated fibroblasts, tumor associated macrophages, extracellular matrix fibrosis，single or combined factors)will be found. In this study，the tissue RNA-seq detection will be made, full transcriptome sequencing and cell subtype analysis will be performed, and the possible mechanism of tumor microenvironment factors affecting the outcome of chemoradiation will be further clarified. Finally this is encouraging to accomplish the radiomics predictive modol optimized by tumor microenvironment factors may lead to better clinical translational applocations to help tailoring the treatment into the era of personalized medicine.