An IFN-γ-related signature predicts prognosis and immunotherapy response in bladder cancer: Results from real-world cohorts

Improve clinical feature-based bladder cancer survival prediction models through integration with gene expression profiles and machine learning techniques

Background

Bladder cancer (BCa), one of the most common cancers worldwide, is characterized by high rates of recurrence, progression, and mortality. Machine learning algorithms offer promising advancements in enhancing predictive models. This study aims to develop robust machine learning models for predicting BCa survival using clinical and gene expression data.

Methods

Clinical data from BCa patients were obtained from the Surveillance, Epidemiology, and End Results database. Cox proportional hazards regression models assessed the association between clinical variables and overall survival. Machine learning algorithms, including logistic regression, random forest, XGBoost, decision tree, and LightGBM, were employed to predict survival at 1, 3, and 5 years. The TAGO database, combined with the data from The Cancer Genome Atlas and four databases from the Gene Expression Omnibus, which have available genomic data and clinical data, were selected. Gene expression data were transformed into gene sets data, and the performance of models based on clinical data and gene sets data and their combination were compared. Furthermore, the impact of model-derived scores on overall survival was evaluated.

Results

Among 138,741 BCa patients with available clinical data, key independent predictors of survival included age, race, marital status, surgery, chemotherapy, radiation, and TNM stages. Clinical data machine learning (CML) models used these clinical predictors to achieve AUC values of 0.860, 0.821, and 0.804 in the testing sets for predicting survival at 1, 3, and 5 years, respectively. In the TAGO database, which has 863 patients with clinical and genomic data, the integrated clinical and gene expression machine learning model (IML) outperformed the CML and gene expression machine learning (GML) models in survival prediction. Patients with higher IML and GML model scores exhibited poorer survival outcomes.

Conclusions

This study successfully identifies key clinical and genomic predictors, a significant step forward in BCa research. The development of predictive models for BCa survival underscores the potential of integrated data approaches in improving BCa management and treatment strategies.

Immuno-Transcriptomic Profiling of Blood and Tumor Tissue Identifies Gene Signatures Associated with Immunotherapy Response in Metastatic Bladder Cancer

Blood-based biomarkers represent ideal candidates for the development of non-invasive immuno-oncology-based assays. However, to date, no blood biomarker has been validated to predict clinical responses to immunotherapy. In this study, we used next-generation sequencing (RNAseq) on bulk RNA extracted from whole blood and tumor samples in a pre-clinical MIBC mouse model. We aimed to identify biomarkers associated with immunotherapy response and assess the potential application of simple non-invasive blood biomarkers as a therapeutic decision-making assay compared to tissue-based biomarkers. We established that circulating immune cells and the tumor microenvironment (TME) display highly organ-specific transcriptional responses to ICIs. Interestingly, in both, a common lymphocytic activation signature can be identified associated with the efficient response to immunotherapy, including a blood-specific CD8+ T cell activation/proliferation signature which predicts the immunotherapy response.

Construction of a Liver Cancer Prognostic Model Based on Interferon-Gamma-Related Genes for Revealing the Immune Landscape

Inferferon-gamma (LFN-γ) exerts anti-tumor effects, but there is currently no reliable and comprehensive study on prognostic function of IFN-γ-related genes in liver cancer. In this study, IFN-γ-related differentially expressed genes (DEGs) in liver cancer were identified through GO/KEGG databases and open-access literature. Based on these genes, individuals with liver cancer were clustered. A prognostic model was built based on the intersection genes between differential genes in clusters and in liver cancer. Then, model predictive performance was analyzed and validated in GEO dataset. Regression analysis was fulfilled on the model, and a nomogram was utilized to evaluate model ability as an independent prognostic factor and its clinical application value. An immune-related analysis was conducted on both the H- and L-groups, with an additional investigation into link of model genes to drug sensitivity. Significant differential expression of IFN-γ-related genes was observed between the liver cancer and control groups. Subsequently, individuals with liver cancer were classified into two subtypes based on these genes, which displayed a notable difference in survival between the two subtypes. A 10-gene liver cancer prognostic model was constructed, with good prognostic performance and was an independent prognosticator for patient analysis. L-group patients possessed higher immune infiltration levels, immune checkpoint expression levels, and immunophenoscore, as well as lower TIDE scores. Drugs that had high correlations with the feature genes included SPANXB1: PF-04217903, SGX-523, MMP1: PF-04217903, DUSP13: Imatinib, TFF1: KHK-Indazole, and Fulvestrant. We built a 10-gene liver cancer prognostic model. It was found that L-group patients were more suitable for immunotherapy. This study provided valuable information on the prognosis of liver cancer.