Background: Lung adenocarcinoma (LUAD), the most common subtype of non-small cell lung cancer (NSCLC), is associated with poor prognosis. However, current stage-based clinical methods are insufficient for survival prediction and decision-making. This study aimed to establish a novel model for evaluating the risk of LUAD based on hypoxia, immunity, and epithelial-mesenchymal transition (EMT) gene signatures.Methods: In this study, we used data from TCGA-LUAD for the training cohort and GSE68465 and GSE72094 for the validation cohorts. Immunotherapy datasets GSE135222, GSE126044, and IMvigor210 were obtained from a previous study. Using bioinformatic and machine algorithms, we established a risk model based on hypoxia, immune, and EMT gene signatures, which was then used to divide patients into the high and low risk groups. We analyzed differences in enriched pathways between the two groups, following which we investigated whether the risk score was correlated with stemness scores, genes related to m6A, m5C, m1A and m7G modification, the immune microenvironment, immunotherapy response, and multiple anti-cancer drug sensitivity.Results: Overall survival differed significantly between the high-risk and low-risk groups (HR = 4.26). The AUCs for predicting 1-, 3-, and 5-year survival were 0.763, 0.766, and 0.728, respectively. In the GSE68465 dataset, the HR was 2.03, while the AUCs for predicting 1-, 3-, and 5-year survival were 0.69, 0.651, and 0.618, respectively. The corresponding values in the GSE72094 dataset were an HR of 2.36 and AUCs of 0.653, 0.662, and 0.749, respectively. The risk score model could independently predict OS in patients with LUAD, and highly correlated with stemness scores and numerous m6A, m5C, m1A and m7G modification-related genes. Furthermore, the risk model was significantly correlated with multiple immune microenvironment characteristics. In the GSE135222 dataset, the HR was 4.26 and the AUC was 0.702. Evaluation of the GSE126044 and IMvigor210 cohorts indicated that PD-1/PD-LI inhibitor treatment may be indicated in patients with low risk scores, while anti-cancer therapy with various drugs may be indicated in patients with high risk scores.Conclusion: Our novel risk model developed based on hypoxia, immune, and EMT gene signatures can aid in predicting clinical prognosis and guiding treatment in patients with LUAD.

The widespread dissemination of forged images generated by Deepfake techniques has posed a serious threat to the trustworthiness of digital information. This demands effective approaches that can detect perceptually convincing Deepfakes generated by advanced manipulation techniques. Most existing approaches combat Deepfakes with deep neural networks by mapping the input image to a binary prediction without capturing the consistency among different pixels. In this paper, we aim to capture the subtle manipulation artifacts at different scales for Deepfake detection. We achieve this with transformer models, which have recently demonstrated superior performance in modeling dependencies between pixels for a variety of recognition tasks in computer vision. In particular, we introduce a Multi-modal Multi-scale TRansformer (M2TR), which uses a multi-scale transformer that operates on patches of different sizes to detect the local inconsistency at different spatial levels. To improve the detection results and enhance the robustness of our method to image compression, M2TR also takes frequency information, which is further combined with RGB features using a cross modality fusion module. Developing and evaluating Deepfake detection methods requires large-scale datasets. However, we observe that samples in existing benchmarks contain severe artifacts and lack diversity. This motivates us to introduce a high-quality Deepfake dataset, SR-DF, which consists of 4,000 DeepFake videos generated by state-of-the-art face swapping and facial reenactment methods. On three Deepfake datasets, we conduct extensive experiments to verify the effectiveness of the proposed method, which outperforms state-of-the-art Deepfake detection methods.

The Golgi apparatus (GA) is a cellular organelle that participates in the packaging, modification, and transport of proteins and lipids from the endoplasmic reticulum to be further fabricated before being presented to other cellular components. Recent studies have demonstrated that GA facilitates numerous cellular processes in cancer development. Therefore, this study aimed to establish a novel lung adenocarcinoma (LUAD) risk evaluation model based on GA gene signatures. In this study, we used TCGA-LUAD (n = 500) as the training cohort and GSE50081 (n = 127), GSE68465 (442), and GSE72094 (398) as the validation cohorts. Two immunotherapy datasets (GSE135222 and GSE126044) were also obtained from a previous study. Based on machine algorithms and bioinformatics methods, a GA gene-related risk score (GARS) was established. We found that the GARS independently predicted the prognosis of LUAD patients and remained effective across stages IA to IIIA. Then, we identified that the GARS was highly correlated with mutations in P53 and TTN. Further, this study identified that GARS is related to multiple immune microenvironmental characteristics. Furthermore, we investigated GSE135222 and GSE126044 and found that a lower GARS may be indicative of an improved therapeutic effect of PD-1/PD-L1 therapy. We also found that high GARS may lead to a better response to multiple anticancer drugs. Finally, we established a nomogram to better guide clinical application. To our knowledge, this is the first study to demonstrate a novel GA signature-based risk score formula to predict clinical prognosis and guide the treatment of LUAD patients.

BackgroundBreast cancer has become the malignancy with the highest mortality rate in female patients worldwide. The limited efficacy of immunotherapy as a breast cancer treatment has fueled the development of research on the tumor immune microenvironment.MethodsIn this study, data on breast cancer patients were collected from The Cancer Genome Atlas Breast Invasive Carcinoma (TCGA-BRCA) and Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) cohorts. Differential gene expression analysis, univariate Cox regression analysis, and least absolute shrinkage and selection operator (LASSO) Cox regression analysis were performed to select overall survival (OS)-related, tumor tissue highly expressed, and immune- and inflammation-related genes. A tumor immune-inflammation signature (TIIS) consisting of 18 genes was finally screened out in the LASSO Cox regression model. Model performance was assessed by time-dependent receiver operating characteristic (ROC) curves. In addition, the CIBERSORT algorithm and abundant expression of immune checkpoints were utilized to clarify the correlation between the risk signature and immune landscape in breast cancer. Furthermore, the association of IL27 with the immune signature was analyzed in pan-cancer and the effect of IL27 on the migration of breast cancer cells was investigated since the regression coefficient of IL27 was the highest.ResultsA TIIS based on 18 genes was constructed via LASSO Cox regression analysis. In the TCGA-BRCA training cohort, 10-year AUC reached 0.89, and prediction performance of this signature was also validated in the METABRIC set. The high-risk group was significantly correlated with less infiltration of tumor-killing immune cells and the lower expression level of the immune checkpoint. Furthermore, we recommended some small-molecule drugs as novel targeted drugs for new breast cancer types. Finally, the relationship between IL27, a significant prognostic immune and inflammation cytokine, and immune status was analyzed in pan-cancer. Expression of IL27 was significantly correlated with immune regulatory gene expression and immune cell infiltration in pan-cancer. Furthermore, IL27 treatment improved breast cancer cell migration.ConclusionThe TIIS represents a promising prognostic tool for estimating OS in patients with breast cancer and is correlated with immune status.