探讨乳腺癌中肿瘤相关巨噬细胞（TAM）的异质性及不同TAM亚群的功能差异，分析TAM相关基因对三阴性乳腺癌（TNBC）患者预后的影响。

收集GSE161529数据集中30例乳腺癌患者与13例正常人乳腺组织样本，采用Seurat流程和共识非负矩阵分解算法鉴定TAM亚群，通过GO和免疫反应富集分析不同TAM亚群的免疫功能。收集GSE148673数据集中5例TNBC乳腺组织样本，采用单因素Cox回归分析与Pearson相关性筛选显著关联基因对，构建TAM预后基因的共表达网络。采用K-means无监督聚类方法，根据共表达网络中的基因对TCGA与GEO数据库中的247例TNBC患者进行分子分型，比较不同亚型患者间的生存差异与临床特征。采用Kaplan-Meier法和Log-rank检验进行生存分析。

GSE161529数据集中注释了8种细胞（上皮细胞、T细胞、成纤维细胞、巨噬细胞、内皮细胞、组织干细胞、B细胞和共同髓系祖细胞），共鉴定出6个功能不同的TAM亚群，分别在免疫炎症、能量代谢以及细胞黏附等方面发挥作用。GSE148673数据集中，TAM相关基因CLEC4E、CTSC、CTSH和CTSS的高表达与TNBC良好预后相关（P<0.01），而STAB1、RNASE1、SDS、SPP1、TREM2的高表达则提示不良预后（P<0.01）。通过单因素Cox回归分析和Pearson相关性分析，构建了26个与患者预后相关的TAM基因共表达网络。根据26个TAM基因的表达将TNBC患者分为保护型（A组，96例）和风险型（B组，99例）。二分类聚类热图显示组内具有高度一致；基因表达热图显示保护型TAM基因在A组高表达、在B组低表达。生存分析显示A组患者总生存率显著优于B组（χ²=6.63，P=0.010）；临床热图进一步显示两组患者在年龄、临床分期上存在差异。

乳腺癌及其肿瘤微环境中的TAM存在异质性，6个TAM亚群在免疫炎症、能量代谢以及细胞黏附等方面具有不同作用。TAM相关基因在TNBC患者中具有临床分型和预后预测价值。

To explore the heterogeneity of tumor-associated macrophage （TAM） in breast cancer and the functional differences among different TAM subgroups， and analyze the impact of TAM-related genes on the prognosis of triple negative breast cancer （TNBC） patients.

Thirty breast cancer tissue samples and 13 normal breast tissue samples from the GSE161529 dataset were collected. TAM subgroups were identified using the Seurat pipeline and consensus non-negative matrix factorization algorithm. The immune functions of different TAM subgroups were analyzed through GO and immune response enrichment analysis. The tissue samples of 5 TNBC patients from the GSE148673 dataset were collected. Univariate Cox regression and Pearson correlation analysis were used to screen for significant gene pairs related to TAM， and a co-expression network of TAM prognostic genes was constructed. K-means unsupervised clustering was used to get molecular subtypes of 247 TNBC patients in the TCGA and GEO databases based on the co-expression network genes, and survival differences and clinical characteristics were compared among different subtypes. Survival analysis was performed using the Kaplan-Meier method and log-rank test.

Eight cell types （epithelial cells， T cells， fibroblasts， macrophages， endothelial cells， tissue stem cells， B cells and common myeloid progenitor cells） were annotated in the GSE161529 dataset. Among them， six functionally distinct TAM subgroups were identified， which played roles in immune inflammation, energy metabolism and cell adhesion. In the GSE148673 dataset， high expression of TAM-related genes CLEC4E， CTSC， CTSH， and CTSS was associated with a favorable prognosis in TNBC patients （P<0.01）， while high expression of STAB1， RNASE1， SDS， SPP1 and TREM2 indicated a poor prognosis （P<0.01）. Through univariate Cox analysis and Pearson correlation analysis， a co-expression network of 26 TAM genes related to patient prognosis was constructed. TNBC patients were classified into protective type （Group A， 96 cases） and risky type （Group B， 99 cases） based on the expression of 26 TAM genes. The binary clustering heatmap showed high consistency within groups； the gene expression heatmap indicated that protective TAM genes were highly expressed in Group A and lowly expressed in Group B. Survival analysis showed that the overall survival of Group A patients was significantly better than that of Group B （χ²=6.63，P=0.010）； the clinical heatmap further revealed differences in age and clinical stage between two groups.

TAMs in breast cancer and its tumor microenvironment are heterogeneous， and six TAM subgroups have distinct roles in immune inflammation, energy metabolism and cell adhesion. TAM-related genes have clinical typing and prognostic prediction value in TNBC patients.