Spatial Analysis Reveals Targetable Macrophage-Mediated Mechanisms of Immune Evasion in Hepatocellular Carcinoma Minimal Residual Disease

We utilized CO-Detection by indEXing (CODEX) technology to image a total of 1,065,284 hepatocellular carcinoma (HCC) cells, analyzing them with a comprehensive 41-marker panel. In this procedure, DNA-barcoded antibodies targeted tissue antigens and were visualized through iterative hybridization with complementary fluorescent DNA oligonucleotides. Our algorithmic pipeline processed the raw imaging data to segment and identify single cells, accurately localize them within tissues, and quantify their marker expressions. Further, unsupervised clustering, enhanced by manual curation, utilized marker expression, and tissue localization to identify specific cell types. The imaging data were segmented into single cells based on nuclear staining, using DAPI as a reference. Spatial coordinates (X/Y) were assigned to each cell, allowing us to pinpoint each cell's precise location on the slide. The methodologies employed are detailed further in the accompanying manuscript. The "Celllevel_rawdata_8_12_24.csv" table has been compiled to include annotations of cell types, expression profiles, coordinates of all segmented objects, and the cohort to which each cell belongs.本研究采用索引共检测（CO-Detection by indEXing, CODEX）技术，对总计1,065,284个肝细胞癌（hepatocellular carcinoma, HCC）细胞完成成像，并通过包含41个标志物的全面检测组合对其进行分析。实验中，靶向组织抗原的DNA条形码标记抗体，通过与互补荧光DNA寡核苷酸进行迭代杂交实现信号可视化。我们搭建的算法流程对原始成像数据进行处理，实现单细胞分割与识别、精准定位细胞在组织中的位置，并量化其标志物表达水平。进一步地，经人工校正优化的无监督聚类方法，依托标志物表达特征与组织定位信息，识别出特定细胞类型。成像数据以核染色为依据分割为单细胞，以4',6-二脒基-2-苯基吲哚（DAPI）作为参照标准。本研究为每个细胞分配了空间坐标（X/Y轴），从而可精准确认其在玻片上的具体位置。本研究所采用的方法学细节已在随附的研究手稿中进一步详述。 我们已编译生成"Celllevel_rawdata_8_12_24.csv"表格，该表格涵盖了所有已分割细胞的细胞类型注释、表达谱、空间坐标，以及每个细胞所属的队列信息。