ALOX5 Expression and Pathomics Features Reveal New Insights Into Lung Adenocarcinoma Prognosis: Model Construction and Functional Validation.

[BACKGROUND] Arachidonic acid 5-lipoxygenase (ALOX5) is a biomarker of lung adenocarcinoma (LUAD). This research seeks to establish a prognostic model for LUAD by examining ALOX5 expression.

[METHODS] The Cancer Genome Atlas database provided the pathological images and transcriptome data. To determine the prognostic value of ALOX5, survival analysis and Cox regression (univariate and multivariate) were conducted, along with subgroup analysis and interaction tests. The OTSU algorithm and the PyRadiomics package were used to segment the pathological images of patients with LUAD and extract pathological features. The gradient-enhanced model algorithm was used to construct the pathological omics model. The prognostic value of pathomics mechanism analysis was confirmed using the pathomics score (PS) output from the model. Cell experiments were used to verify gene function.

[RESULTS] A total of 327 samples and seven best pathological features were included in the analysis. In LUAD, elevated levels of ALOX5 and PS were associated with improved overall survival. The gradient boosting machine (GBM) pathomics model demonstrated strong predictive performance and clinical applicability, achieving an area under the curve (AUC) of 0.786 in the training set ( = 230) and 0.741 in the validation set ( = 97). According to the model, samples with elevated ALOX5 expression exhibited higher PS values. Moreover, macrophage infiltration was significantly increased in groups with high PS expression. Gene set enrichment analysis (GSEA) indicated that genes differentially expressed between PS subgroups were involved in apoptosis and inflammatory-response pathways. Apoptosis-related genes were positively correlated with PS values ( < 0.001), and 63 hub genes associated with the inflammatory response were enriched in cytokine-mediated signaling pathways. In vitro experiments showed that ALOX5 knockdown in lung cancer cells enhanced tumor cell proliferation and migration.

[CONCLUSIONS] There was a strong link between ALOX5 expression levels and OS in LUAD patients. A pathomics-based model can effectively predict the expression level of ALOX5; as a result, LUAD patients' prognoses can be predicted.