Background: The tumor microenvironment (TME) consists of heterogeneous cell populations, including malignant cells and nonmalignant cells that support tumor proliferation, invasion, and metastasis through extensive cross talk. The intra-tumor immune landscape is a critical factor influencing patient survival and response to immunotherapy.

Methods: Gene expression data were downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus databases. Immune cell infiltration was determined by single-sample Gene Set Enrichment Analysis (ssGSEA) depending on the integrated immune gene sets from published studies. Univariate analysis was used to determine the prognostic value of the infiltrated immune cells. Least absolute shrinkage and selection operator (LASSO) regression was performed to screen for the most survival-relevant immune cells. An immune-cell characteristic score (ICCS) model was constructed by using multivariate Cox regression analysis.

Results: The immune cell infiltration patterns across 32 cancer types were identified, and patients in the high immune cell infiltration cluster had worse overall survival (OS) but better progression-free interval (PFI) compared to the low immune cell infiltration cluster. However, immune cell infiltration showed inconsistent prognostic value depending on the cancer type. High immune cell infiltration (High CI) indicated a worse prognosis in brain lower grade glioma (LGG), glioblastoma multiforme (GBM), and uveal melanoma (UVM), and favorable prognosis in adrenocortical carcinoma (ACC), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), cholangiocarcinoma (CHOL), head and neck squamous cell carcinoma (HNSC), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), sarcoma (SARC), and skin cutaneous melanoma (SKCM). LUAD prognosis was significantly influenced by the infiltration of 13 immune cell types, with high infiltration of all but Type 2 T helper (Th2) cells correlating with a favorable prognosis. The ICCS model based on six most survival-relevant immune cell populations was generated that classified patients into low- and high-ICCS groups with good and poor prognoses, respectively. The multivariate and stratified analyses further revealed that the ICCS was an independent prognostic factor for LUAD.

Conclusions: The infiltration of immune cells in 32 cancer types was quantified, and considerable heterogeneity was observed in the prognostic relevance of these cells in different cancer types. An ICCS model was constructed for LUAD with competent prognostic performance, which can further deepen our understanding of the TME of LUAD and can have implications for immunotherapy.

背景：肿瘤微环境（TME）由异质细胞群组成，包括通过广泛的串扰来支持肿瘤增殖，侵袭和转移的恶性细胞和非恶性细胞。肿瘤内免疫状况是影响患者存活率和对免疫疗法反应的关键因素。

方法：基因表达数据从癌症基因组图谱（TCGA）和基因表达综合数据库下载。通过单样本基因集富集分析（ssGSEA）确定免疫细胞浸润，具体取决于已发表研究的整合免疫基因集。单因素分析用于确定浸润的免疫细胞的预后价值。进行最小绝对收缩和选择算子（LASSO）回归，以筛选与生存最相关的免疫细胞。通过使用多元Cox回归分析构建免疫细胞特征评分（ICCS）模型。

结果：确定了32种癌症类型的免疫细胞浸润模式，与低免疫细胞浸润簇相比，高免疫细胞浸润簇中的患者的总生存期（OS）较差，但无进展间隔（PFI）更好。然而，免疫细胞浸润根据癌症类型显示出不一致的预后价值。高免疫细胞浸润（High CI）表示脑低级神经胶质瘤（LGG），多形胶质母细胞瘤（GBM）和葡萄膜黑色素瘤（UVM）的预后较差，而肾上腺皮质癌（ACC），宫颈鳞状细胞癌和宫颈内膜的预后较好腺癌（CESC），胆管癌（CHOL），头颈部鳞状细胞癌（HNSC），肝肝细胞癌（LIHC），肺腺癌（LUAD），肉瘤（SARC）和皮肤皮肤黑色素瘤（SKCM）。13种免疫细胞的浸润显着影响LUAD的预后，除2型T辅助细胞（Th2）以外的所有细胞的高浸润均与良好的预后相关。建立了基于六个与存活率最相关的免疫细胞群的ICCS模型，该模型将患者分为预后良好和不良的低和高ICCS组。多元和分层分析进一步表明，ICCS是LUAD的独立预后因素。分别。多元和分层分析进一步表明，ICCS是LUAD的独立预后因素。分别。多元和分层分析进一步表明，ICCS是LUAD的独立预后因素。

结论：定量分析了32种癌症类型中免疫细胞的浸润情况，并观察到这些细胞在不同癌症类型中的预后相关性存在相当大的异质性。为LUAD构建了具有良好预后性能的ICCS模型，该模型可以进一步加深我们对LUAD TME的理解，并可能对免疫治疗产生影响。