Background: Lung adenocarcinoma (LUAD) is the most common subtype of non-small cell lung cancer. Fatty acid metabolism takes part in malignancy progression. However, the roles fatty acid metabolism plays in LUAD are still unclear.

Methods: The transcriptomic and clinical data of LUAD patients from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases were extracted. ssGSEA, WGCNA, univariable Cox regression, and LASSO Cox regression analyses were performed to identify the fatty acid metabolism-related genes which influenced the overall survival (OS) and build a fatty acid-related risk score (FARS) model. A nomogram was established based on the FARS and other clinicopathological features, and ROC and calibration plots were used to validate the prediction accuracy. The tumor microenvironment (TME) of patients with high and low FARS was compared.

Results: A total of 38 genes were identified to be independently related to the survival outcome and put into a FARS model. High FARS patients exhibited significantly worse OS. The nomogram included the FARS and pathological stage, and the AUC of the nomogram predicting 1-, 2-, 3-, 4-, and 5-year OS was 0.789, 0.807, 0.798, 0.809, and 0.753, respectively. Calibration plots also indicated good accuracy. Moreover, the samples of the high FARS had higher expression of PDL1.

Conclusion: We constructed a FARS model which could accurately predict the survival outcome of the LUAD patients. The genes of the FARS are related to the tumor microenvironment and patients with high FARS can potentially benefit more from anti-PD1/PDL1 immunotherapy. In addition, the mechanisms of the genes in the FARS affecting prognosis are worthy of further research to develop new gene-targeted drugs.

背景：肺腺癌（LUAD）是非小细胞肺癌最常见的亚型。脂肪酸代谢参与恶性肿瘤进展。然而，脂肪酸代谢在 LUAD 中的作用仍不清楚。

方法：从癌症基因组图谱 (TCGA) 和基因表达综合 (GEO) 数据库中提取 LUAD 患者的转录组和临床数据。进行ssGSEA、WGCNA、单变量Cox回归和LASSO Cox回归分析，以确定影响总生存期（OS）的脂肪酸代谢相关基因，并建立脂肪酸相关风险评分（FARS）模型。基于 FARS 和其他临床病理特征建立列线图，并使用 ROC 和校准图来验证预测准确性。比较了高和低 FARS 患者的肿瘤微环境 (TME)。

结果：共有 38 个基因被鉴定为与生存结果独立相关，并将其放入 FARS 模型中。高 FARS 患者表现出显着更差的 OS。列线图包括 FARS 和病理分期，预测 1 年、2 年、3 年、4 年和 5 年 OS 的列线图的 AUC 分别为 0.789、0.807、0.798、0.809 和 0.753。校准曲线也表明了良好的准确性。此外，高 FARS 的样本具有更高的 PDL1 表达。

结论：我们构建了一个可以准确预测 LUAD 患者生存结果的 FARS 模型。FARS 的基因与肿瘤微环境有关，FARS 高的患者可能会从抗 PD1/PDL1 免疫治疗中获益更多。此外，FARS中基因影响预后的机制值得进一步研究，以开发新的基因靶向药物。