基于PD-L1分子影像学特征筛选结直肠癌免疫治疗生物学标志物和构建诊断预测模型

以抗体为探针的分子影像学检查是结直肠癌无创动态检测的新方法。免疫检测点PD-1/PD-L1是结直肠癌肿瘤耐受的重要机制。在肿瘤组织中精准监测PD-L1有助于免疫治疗选择、疗效及预后评估。课题组前期已成功制备光学显像剂NIR-PD-L1-mAb，检测到不同结直肠癌细胞及动物模型中PD-L1的表达变化。在此基础上，将采用放射性核素89Zr及131I分别制备PD-L1抗体PET与切伦科夫光学分子探针，检测结直肠癌中PD-L1的表达变化。结合肿瘤组织18F-FDG的SUV值、免疫浸润程度、淋巴细胞分布、细胞因子水平以及免疫相关基因突变频率和表达水平，研究PD-L1分子影像学指标与免疫治疗反应的相关性。基于朴素贝叶斯算法构建综合的预后和治疗反应的预测模型，并提供基于分子影像学特征筛选结直肠癌PD-L1免疫治疗相关生物学标志物的医学影像辅助诊断平台，有利于患者免疫治疗的选择和临床医生精准制定治疗策略。

Molecular imaging with antibodies as a probe is a new method for noninvasive dynamic detection of colorectal cancer. Immune checkpoint PD-1/PD-L1 is an important mechanism of tumor tolerance in colorectal cancer. Precise monitoring of PD-L1 expression level in tumor tissues is helpful for the selection of targeted immunotherapy for patients and timely evaluation of efficacy and prognosis. Our group successfully prepared optical imaging agent NIR-PD-L1-mAb and detected the expression of PD-L1 in different colorectal cancer cells and animal models. On this basis, PD-L1 antibody PET and Cerenkov optical molecular probes were prepared using radionuclide 89Zr and 131I, respectively, to detect the dynamic changes of PD-L1 expression in colorectal cancer. The correlation between PD-L1 molecular imaging and immunotherapy response was studied based on the SUV value, immune infiltration, lymphocyte distribution, cytokine levels and mutation frequencies and expression levels of immune-related genes in 18F-FDG of tumor tissues. Based on Naive Bayesian Algorithm, a comprehensive prediction model of prognosis and treatment response was built, and a medical imaging auxiliary diagnosis platform for screening biomarkers relative to colorectal cancer PD-L1 immunotherapy based on molecular imaging features, which is helpful for seletion of patients immunotherapy and clinicians to develop precise treatment strategies.