To address the substantial variability in immune checkpoint blockade (ICB) therapy effectiveness, we developed an innovative R package called integrated Machine Learning and Genetic Algorithm‐driven Multiomics analysis (iMLGAM), which establishes a comprehensive scoring system for predicting treatment outcomes through advanced multi‐omics data integration. Our research demonstrates that iMLGAM scores exhibit superior predictive performance across independent cohorts, with lower scores correlating significantly with enhanced therapeutic responses and outperforming existing clinical biomarkers. Detailed analysis revealed that tumors with low iMLGAM scores display distinctive immune microenvironment characteristics, including increased immune cell infiltration and amplified antitumor immune responses. Critically, through clustered regularly interspaced short palindromic repeats screening, we identified Centrosomal Protein 55 (CEP55) as a key molecule modulating tumor immune evasion, mechanistically confirming its role in regulating T cell‐mediated antitumor immune responses. These findings not only validate iMLGAM as a powerful prognostic tool but also propose CEP55 as a promising therapeutic target, offering novel strategies to enhance ICB treatment efficacy. The iMLGAM package is freely available on GitHub (https://github.com/Yelab1994/iMLGAM), providing researchers with an innovative approach to personalized cancer immunotherapy prediction.

Background Lung adenocarcinoma (LUAD) is a highly heterogeneous disease, posing significant challenges to accurate prognosis prediction. Mitochondria play a central role in the energy metabolism of eukaryotic cells and can influence programmed cell death (PCD) mechanisms, which are critical in tumorigenesis and cancer progression. However, the prognostic significance of the interplay between mitochondrial function and PCD in LUAD requires further investigation. Methods We analyzed data from 1231 LUAD patients across seven global cohorts to develop a mitochondrial-related PCD signature (MPCDS) using machine learning. Validation was done using six immunotherapy cohorts (LUAD, melanoma, clear cell renal cell carcinoma; n=935) and a pan-cancer cohort of 21 tumor types. An in-house LUAD tissue cohort (n=100) confirmed the prognostic significance of nucleoside diphosphate kinase 4 (NME4). In vivo and in vitro experiments explored NME4’s role in immune exclusion. Results The MPCDS demonstrated strong predictive performance for prognosis in LUAD patients, surpassing 114 previously published LUAD signatures. Additionally, MPCDS effectively predicted outcomes in immunotherapy patients (including those with LUAD, melanoma, and clear cell renal cell carcinoma). Biologically, MPCDS was significantly associated with immune features, with the high MPCDS group exhibiting reduced immune activity and a tendency towards cold tumors. NME4, a key gene within the MPCDS (correlation=0.55, p<0.05), was associated with poorer prognosis in LUAD patients with high expression, particularly in CD8 desert phenotypes, as validated by our in-house cohort. Multiplex immunofluorescence confirmed the spatial colocalization and exclusion relationship between NME4 and immune cells such as CD3+ T cells and CD20+ B cells. Further experiments revealed that NME4 regulated the proliferation and invasion of LUAD cells both in vitro and in vivo. Importantly, inhibiting NME4 increased the abundance and activity of CD8+ T cells and enhanced the antitumor immunity of anti-programmed cell death protein-1 therapy in vivo. Conclusion The MPCDS provides personalized risk assessment and immunotherapy interventions for individual LUAD patients. NME4, a key gene within the MPCDS, has been identified as a novel oncogene associated with immune exclusion and may serve as a new target for LUAD intervention and immunotherapy.

The human microbiome is now recognized as a central regulator of cancer biology, intricately shaping tumor development, immune dynamics, and therapeutic response. This comprehensive review delineates the multifaceted roles of bacteria, viruses, and fungi in modulating the tumor microenvironment and systemic immunity across diverse cancer types. We synthesize current evidence on how microbial dysbiosis promotes carcinogenesis via chronic inflammation, metabolic reprogramming, genotoxic stress, immune evasion, and epigenetic remodeling. This review emphasizes organ‐specific microbiome signatures and highlights their potential as non‐invasive biomarkers for early detection, treatment stratification, and prognosis. Furthermore, we explore the impact of intratumoral microbiota on cancer therapies, uncovering how microbial metabolites and host–microbe interactions shape therapeutic efficacy and resistance. Finally, advances in microbiome‐targeted strategies, such as probiotics, fecal microbiota transplantation, and engineered microbes offer new avenues for adjunctive cancer therapy. This review provides a roadmap for future investigation and underscores the transformative promise of microbiome modulation in cancer prevention and treatment.

Immunotherapy has brought significant advancements in the treatment of lung adenocarcinoma (LUAD), but identifying suitable candidates remains challenging. In this study, we investigated tumour cell heterogeneity using extensive single‐cell data and explored the impact of different tumour cell cluster abundances on immunotherapy in the POPLAR and OAK immunotherapy cohorts. Notably, we found a significant correlation between CKS1B+ tumour cell abundance and treatment response, as well as stemness potential. Leveraging marker genes from the CKS1B+ tumour cell cluster, we employed machine learning algorithms to establish a prognostic and immunotherapeutic signature (PIS) for LUAD. In multiple cohorts, PIS outperformed 144 previously published signatures in predicting LUAD prognosis. Importantly, PIS reliably predicted genomic alterations, chemotherapy sensitivity and immunotherapy responses. Immunohistochemistry validated lower expression of immune markers in the low‐PIS group, while in vitro experiments underscored the role of the key gene PSMB7 in LUAD progression. In conclusion, PIS represents a novel biomarker facilitating the selection of suitable LUAD patients for immunotherapy, ultimately improving prognosis and guiding clinical decisions.

Epidermal growth factor receptor (EGFR) mutation is a key oncogenic driver in lung adenocarcinoma (LUAD), but its impact on the tumor immune microenvironment (TIME) remains unclear. By integrating single-cell transcriptomes from 153 LUAD samples using machine learning, we generated an atlas of over one million cells that delineates immune heterogeneity. EGFR-mutant tumors exhibited enrichment of TIGIT⁺regulatory T cells, neutrophils, and macrophages, whereas wild-type tumors contained abundant ZNF683⁺CD8⁺tissue-resident memory T cells, diverse memory B cells, and FGFBP2⁺CD16^high natural killer cells, reflecting an immune-active TIME. Non-negative matrix factorization defined five TIME subtypes, with EGFR-mutant patients clustering into immunosuppressive profiles linked to poor prognosis. Flow cytometry and mouse models confirmed the cytotoxic and PD-1 blockade-enhancing functions of FGFBP2⁺NK cells. These findings reveal distinct TIME landscapes in EGFR-mutant LUAD and illustrate the potential of machine learning-based immunogenomic analysis to inform precision immunotherapy.