A Novel Machine Learning 13-Gene Signature: Improving Risk Analysis and Survival Prediction for Clear Cell Renal Cell Carcinoma Patients

Transcriptome analysis revealed a novel nine-gene prognostic risk score of clear cell renal cell carcinoma

Clear cell renal cell carcinoma (ccRCC) continues to pose a significant global health concern, with rising incidence and high mortality rate. Accordingly, identifying molecular alternations associated with ccRCC is crucial to boost our understanding of its onset, persistence, and progression as well as developing prognostic biomarkers and novel therapies. Bulk RNA sequencing data and its associated clinicopathological variables of ccRCC were obtained from The Cancer Genome Atlas Program. Atypical differential gene expression analysis of advanced disease states using the extreme categories of staging and grading components was performed. Upregulated differentially expressed genes shared across the aforementioned components were selected. The risk-score construction pipeline started with univariate Cox logistic regression analysis, least absolute shrinkage and selection operator, and multivariate Cox logistic regression analysis in sequence. The generated risk score classified patients into low- vs high-risk groups. The predictive power of the constructed risk score was assessed using Kaplan-Meier curves analysis, multivariate Cox logistic regression analysis, and receiver operator curve of the overall survival. External validation of the risk score was performed using the E-MTAB-1980 cohort. The analysis work scheme established a novel nine-gene prognostic risk score composed of the following genes: ZIC2, TNNT1, SAA1, OTX1, C20orf141, CDHR4, HOXB13, IGFL2, and IGFN1. The high-risk group was associated with shortened overall survival and possessed an independent predictive power (hazard ratio: 1.942, 95% CI: 1.367-2.758, P < .0001, area under the curve = 0.719). In addition, the high-risk score was associated with advance clinicopathological parameters. The same pattern was observed within the external validation dataset (E-MTAB-1980 cohort), in which the high-risk score held a poor prognostic signature as well as independent predictive potential (hazard ratio: 5.121, 95% CI: 1.412-18.568, P = .013, area under the curve = 0.787). In the present work, a novel nine-gene prognostic risk score was constructed and validated. The risk score correlated with tumor immune microenvironment, somatic mutation patterns, and altered molecular pathways involved in tumorigenesis. Further experimental data are warranted to expand the work.

Survival prediction landscape: an in-depth systematic literature review on activities, methods, tools, diseases, and databases

Survival prediction integrates patient-specific molecular information and clinical signatures to forecast the anticipated time of an event, such as recurrence, death, or disease progression. Survival prediction proves valuable in guiding treatment decisions, optimizing resource allocation, and interventions of precision medicine. The wide range of diseases, the existence of various variants within the same disease, and the reliance on available data necessitate disease-specific computational survival predictors. The widespread adoption of artificial intelligence (AI) methods in crafting survival predictors has undoubtedly revolutionized this field. However, the ever-increasing demand for more sophisticated and effective prediction models necessitates the continued creation of innovative advancements. To catalyze these advancements, it is crucial to bring existing survival predictors knowledge and insights into a centralized platform. The paper in hand thoroughly examines 23 existing review studies and provides a concise overview of their scope and limitations. Focusing on a comprehensive set of 90 most recent survival predictors across 44 diverse diseases, it delves into insights of diverse types of methods that are used in the development of disease-specific predictors. This exhaustive analysis encompasses the utilized data modalities along with a detailed analysis of subsets of clinical features, feature engineering methods, and the specific statistical, machine or deep learning approaches that have been employed. It also provides insights about survival prediction data sources, open-source predictors, and survival prediction frameworks.

The Importance of HHLA2 in Solid Tumors-A Review of the Literature

Cancer immunotherapy is a rapidly developing field of medicine that aims to use the host's immune mechanisms to inhibit and eliminate cancer cells. Antibodies targeting CTLA-4, PD-1, and its ligand PD-L1 are used in various cancer therapies. However, the most thoroughly researched pathway targeting PD-1/PD-L1 has many limitations, and multiple malignancies resist its effects. Human endogenous retrovirus-H Long repeat-associating 2 (HHLA2, known as B7H5/B7H7/B7y) is the youngest known molecule from the B7 family. HHLA2/TMIGD2/KIRD3DL3 is one of the critical pathways in modulating the immune response. Recent studies have demonstrated that HHLA2 has a double effect in modulating the immune system. The connection of HHLA2 with TMIGD2 induces T cell growth and cytokine production via an AKT-dependent signaling cascade. On the other hand, the binding of HHLA2 and KIR3DL3 leads to the inhibition of T cells and mediates tumor resistance against NK cells. This review aimed to summarize novel information about HHLA2, focusing on immunological mechanisms and clinical features of the HHLA2/KIR3DL3/TMIGD2 pathway in the context of potential strategies for malignancy treatment.

Machine Learning Gene Signature to Metastatic ccRCC Based on ceRNA Network

Clear-cell renal-cell carcinoma (ccRCC) is a silent-development pathology with a high rate of metastasis in patients. The activity of coding genes in metastatic progression is well known. New studies evaluate the association with non-coding genes, such as competitive endogenous RNA (ceRNA). This study aims to build a ceRNA network and a gene signature for ccRCC associated with metastatic development and analyze their biological functions. Using data from The Cancer Genome Atlas (TCGA), we constructed the ceRNA network with differentially expressed genes, assembled nine preliminary gene signatures from eight feature selection techniques, and evaluated the classification metrics to choose a final signature. After that, we performed a genomic analysis, a risk analysis, and a functional annotation analysis. We present an 11-gene signature: SNHG15, AF117829.1, hsa-miR-130a-3p, hsa-mir-381-3p, BTBD11, INSR, HECW2, RFLNB, PTTG1, HMMR, and RASD1. It was possible to assess the generalization of the signature using an external dataset from the International Cancer Genome Consortium (ICGC-RECA), which showed an Area Under the Curve of 81.5%. The genomic analysis identified the signature participants on chromosomes with highly mutated regions. The hsa-miR-130a-3p, AF117829.1, hsa-miR-381-3p, and PTTG1 were significantly related to the patient's survival and metastatic development. Additionally, functional annotation resulted in relevant pathways for tumor development and cell cycle control, such as RNA polymerase II transcription regulation and cell control. The gene signature analysis within the ceRNA network, with literature evidence, suggests that the lncRNAs act as "sponges" upon the microRNAs (miRNAs). Therefore, this gene signature presents coding and non-coding genes and could act as potential biomarkers for a better understanding of ccRCC.

Artificial Intelligence in Pathomics and Genomics of Renal Cell Carcinoma

The integration of artificial intelligence (AI) with histopathology images and gene expression patterns has led to the emergence of the dynamic fields of pathomics and genomics. These fields have revolutionized renal cell carcinoma (RCC) diagnosis and subtyping and improved survival prediction models. Machine learning has identified unique gene patterns across RCC subtypes and grades, providing insights into RCC origins and potential treatments, as targeted therapies. The combination of pathomics and genomics using AI opens new avenues in RCC research, promising future breakthroughs and innovations that patients and physicians can anticipate.

A novel prognostic predictor of immune microenvironment and therapeutic response in clear cell renal cell carcinoma based on angiogenesis-immune-related gene signature

Background

Clear cell renal cell carcinoma (ccRCC), the most common type of RCC, typically produces no symptoms initially. Patients with ccRCC are at increased risk of developing advanced metastatic disease due to the absence of dependable and effective prognostic biomarkers. Therefore, it is particularly urgent to find optimal stratification of patients with ccRCC to distinguish the clinical benefits of different malignant degrees. Angiogenesis has a profound impact on the malignant behavior of renal cancer cells, and anti-angiogenic drugs have been applied to metastatic renal cancer patients. Moreover, immune function dysregulation is also a significant factor in tumorigenesis. We aim to construct a predictive model that combines angiogenesis and immune-related genes (AIRGs) to aid clinicians in predicting ccRCC prognosis.

Methods

We gathered transcriptome and clinicopathology data from two datasets, the E-MTAB-1980 dataset and the Cancer Genome Atlas (TCGA). We utilized consensus clustering to find new molecular subgroups. A predictive model for the prognosis of angiogenesis-immune-associated genes (AIRGs) was conducted by the lasso and multivariate Cox regression analysis. The signature's predictive ability was then tested in different datasets. Meticulous scrutiny and comprehensive assessment were undertaken, both internally and externally, to establish the prognostic model. Analyses of immunogenomics were carried out to examine the relationship between risk scores and clinical/immune features, including immune cell infiltration, genomic alterations, and response to targeted and immunotherapy therapy.

Results

Our prognostic signature, comprising 4 AIRGs, stood as an independent prognostic factor for ccRCC, while risk scores emerged as a novel indicator for forecasting overall survival. Risk scores exhibited significant associations with various immunophenotypic factors, such as oncogenic pathways, antitumor response, different immune cell infiltration, antitumor immunity, and response to targeted and immunotherapy therapy.

Conclusions

AIRGs-based prognostic prediction model could effectively predict immunotherapy responses and survival outcomes of ccRCC.

LINC01234 promoted malignant behaviors of breast cancer cells via hsa-miR-30c-2-3p/CCT4/mTOR signaling pathway

OBJECTIVE

Despite continuous progress in treatment, recurrence and metastasis limit further improvement in the prognosis of breast cancer (BC) patients. Our aim was to search for a crucial prognostic biomarker of BC.

MATERIALS AND METHODS

Patient data were selected from The Cancer Genome Atlas (TCGA) and GTEx databases. Several online public databases, including Gene Expression Profiling Interactive Analysis (GEPIA), miRWalk, miRDB, and LncBase Predicted v.2, were used to identify potential upstream miRNAs and lncRNAs. These findings were validated through in vitro experiments.

RESULTS

A total of 1, 097 invasive BC samples and 572 normal breast tissues (including 113 samples from TCGA and 459 samples from GTEx) were collected for the study. CCT4 was not only significantly overexpressed in BC compared with normal breast tissues but also had important prognostic significance (P < 0.001). By intersecting miRWalk and miRDB and conducting correlation analysis, hsa-miR-30c-2-3p was identified as the most probable upstream miRNA of CCT4. Following an extensive assessment that included survival analysis, correlation analysis, and common binding-site prediction, LINC01234 was chosen as the most likely upstream lncRNA. In vitro experiments showed that LINC01234-siRNA inhibited the proliferation, invasion, and migration abilities of BC cells. Western blot analysis further confirmed that LINC01234 promoted malignant behaviors of BC cells via the CCT4/mTOR signaling pathway.

CONCLUSION

The LINC01234/hsa-miR-30c-2-3p/CCT4/mTOR axis was identified as a potential ceRNA regulatory mechanism in BC. These findings established the foundation for systematically unveiling the pathological mechanisms of BC and provided new insights for targeted therapy of BC patients.

Mathematical and Machine Learning Models of Renal Cell Carcinoma: A Review

This review explores the multifaceted landscape of renal cell carcinoma (RCC) by delving into both mechanistic and machine learning models. While machine learning models leverage patients' gene expression and clinical data through a variety of techniques to predict patients' outcomes, mechanistic models focus on investigating cells' and molecules' interactions within RCC tumors. These interactions are notably centered around immune cells, cytokines, tumor cells, and the development of lung metastases. The insights gained from both machine learning and mechanistic models encompass critical aspects such as signature gene identification, sensitive interactions in the tumors' microenvironments, metastasis development in other organs, and the assessment of survival probabilities. By reviewing the models of RCC, this study aims to shed light on opportunities for the integration of machine learning and mechanistic modeling approaches for treatment optimization and the identification of specific targets, all of which are essential for enhancing patient outcomes.

B7-H7: A potential target for cancer immunotherapy

Cancer immunotherapy enhances the body's immunity against tumors by mitigating immune escape. Compared with traditional chemotherapy, immunotherapy has the advantages of fewer drugs, a wider range of action and fewer side effects. B7-H7 (also known as HHLA2, B7y) is a member of the B7 family of costimulatory molecules that was discovered more than 20 years ago. B7-H7 is mostly expressed in organs such as the breast, intestine, gallbladder and placenta and is detected predominantly in monocytes/macrophages in the immune system. Its expression is upregulated after stimulation by inflammatory factors such as lipopolysaccharide and interferon-γ. B7-H7/transmembrane and immunoglobulin domain containing 2 (TMIGD2) and killer cell immunoglobulin-like receptor, three Ig domains and long cytoplasmic tail 3 (KIR3DL3)-B7-H7 are the two currently confirmed signaling pathways for B7-H7. An increasing number of studies have demonstrated that B7-H7 is widely present in a variety of human tumor tissues, especially in programmed cell death-1 (PD-L1)-negative human tumors. B7-H7 promotes tumor progression, disrupts T-cell-mediated antitumor immunity, and inhibits immune surveillance. B7-H7 also triggers tumor immune escape and is associated with clinical stage, depth of tumor infiltration, metastasis, prognosis, and survival related to different tumor types. Multiple studies have shown that B7-H7 is a promising immunotherapeutic target. Herein, review the current literature on the expression, regulation, receptors and function of B7-H7 and its regulation/function in tumors.

HHLA2 immune-regulatory roles in cancer

Human endogenous retrovirus H long terminal repeat-associating protein 2 (HHLA2 or B7-H7) is a newly discovered B7 family member. HHLA2 is aberrantly expressed in solid tumors and exerts co-stimulatory or co-inhibitory activities dependent on interaction with counter receptors. HHLA2 represents co-stimulatory effects upon interaction with transmembrane and immunoglobulin domain containing 2 (TMIGD2, also called CD28H), but its interaction with killer cell Ig-like receptor, three Ig domains and long cytoplasmic tail 3 (KIR3DL3) renders co-inhibitory effects. TMIGD2 is mainly expressed on resting or naïve T cells, whereas expression of KIR3DL3 occurs on activated T cells. HHLA2/KIR3DL3 attenuates responses from both innate and adaptive anti-tumor immunity, and the activity within this axis is regarded as a biomarker of weak prognosis in cancer patients. HHLA2/KIR3DL3 promotes CD8⁺ T cell exhaustion and induces macrophage polarity toward pro-tumor M2 phenotype. HHLA2 represents diverse expression profile and activity in tumor and stroma. Tumoral expression of HHLA2 is presumably higher compared with programmed death-ligand 1 (PD-L1), and HHLA2 co-expression with PD-L1 is indicative of more severe outcomes. A suggested strategy in patients with HHLA2^high cancer is to use monoclonal antibodies for specifically suppressing the HHLA2 inhibitory receptor KIR3DL3, not the HHLA2 ligand. TMIGD2 can be a target for development of agonistic bispecific antibodies for hampering tumor resistance to the programmed death-1 (PD-1)/PD-L1 blockade therapy.

Incremental value of radiomics with machine learning to the existing prognostic models for predicting outcome in renal cell carcinoma

Purpose

To systematically evaluate the potential of radiomics coupled with machine-learning algorithms to improve the predictive power for overall survival (OS) of renal cell carcinoma (RCC).

Methods

A total of 689 RCC patients (281 in the training cohort, 225 in the validation cohort 1 and 183 in the validation cohort 2) who underwent preoperative contrast-enhanced CT and surgical treatment were recruited from three independent databases and one institution. 851 radiomics features were screened using machine-learning algorithm, including Random Forest and Lasso-COX Regression, to establish radiomics signature. The clinical and radiomics nomogram were built by multivariate COX regression. The models were further assessed by Time-dependent receiver operator characteristic, concordance index, calibration curve, clinical impact curve and decision curve analysis.

Result

The radiomics signature comprised 11 prognosis-related features and was significantly correlated with OS in the training and two validation cohorts (Hazard Ratios: 2.718 (2.246,3.291)). Based on radiomics signature, WHOISUP, SSIGN, TNM Stage and clinical score, the radiomics nomogram has been developed. Compared with the existing prognostic models, the AUCs of 5 years OS prediction of the radiomics nomogram were superior to the TNM, WHOISUP and SSIGN model in the training cohort (0.841 vs 0.734, 0.707, 0.644) and validation cohort2 (0.917 vs 0.707, 0.773, 0.771). Stratification analysis suggested that the sensitivity of some drugs and pathways in cancer were observed different for RCC patients with high-and low-radiomics scores.

Conclusion

This study showed the application of contrast-enhanced CT-based radiomics in RCC patients, creating novel radiomics nomogram that could be used to predict OS. Radiomics provided incremental prognostic value to the existing models and significantly improved the predictive power. The radiomics nomogram might be helpful for clinicians to evaluate the benefit of surgery or adjuvant therapy and make individualized therapeutic regimens for patients with renal cell carcinoma.

High VSX1 expression promotes the aggressiveness of clear cell renal cell carcinoma by transcriptionally regulating FKBP10

BACKGROUND

Clear cell renal cell carcinoma (ccRCC), the most common urological malignancy, has an unfavorable prognosis and an unknown mechanism of progression. Through survival analyses screening of The Cancer Genome Atlas (TCGA) dataset, we identified Visual system homeobox1 (VSX1) as a novel potential prognostic biomarker in ccRCC and subsequently investigated the oncogenic role of VSX1 in ccRCC.

METHODS

The differential expression of VSX1 in human tumors and the clinical prognoses were analyzed in the TCGA dataset and Gene Expression Omnibus. Spearman's correlation coefficient was determined for the correlation analysis of VSX1 expression and other genes of interest. The roles of VSX1 in cell proliferation, invasion, and migration of ccRCC cells were evaluated via the CCK-8 assay, colony formation assay, and Transwell assay, respectively. Further results were demonstrated by western blotting, immunohistochemistry, qRT-PCR, tumor sphere formation, flow cytometry, and the dual‑luciferase reporter assay.

RESULTS

VSX1 mRNA upregulation was generally observed in multiple human malignancies from the TCGA database and was confirmed in ccRCC clinical specimens from our department. High VSX1 expression usually indicated that overall and disease-free survival were unfavorable for patients with ccRCC. In terms of mechanism, knockdown or overexpression of VSX1 affected ccRCC aggressiveness in vitro. The dual-luciferase reporter gene assay implied that VSX1 overexpression significantly increased the luciferase activity of TMEM44, FKBP10, and TRIB3, which indicated that VSX1 promoted ccRCC invasiveness via transcriptional regulation of these genes. The significantly enhanced growth in vitro that was induced by stable VSX1 overexpression was almost restored to normal by the knockdown of FKBP10.

CONCLUSIONS

This study demonstrated that VSX1 was a novel prognostic biomarker in ccRCC and that high VSX1 expression promoted cell proliferation, invasion, and migration in ccRCC via transcriptional activation of downstream target genes.

A predictive signature based on enhancer RNA associates with immune infiltration and aids treatment decision in clear cell renal cell carcinoma

Clear cell renal cell carcinoma (ccRCC) is a prevalent urinary malignancy. Despite the recent development of better diagnostic tools and therapy, the five-year survival rate for individuals with advanced and metastatic ccRCC remains dismal. Unfortunately, ccRCC is less susceptible to radiation and chemotherapy. Consequently, targeted therapy and immunotherapy play a crucial role in the treatment of ccRCC. Enhancer RNAs (eRNAs) are noncoding RNAs transcribed by enhancers. Extensive research has shown that eRNAs are implicated in a variety of cancer signaling pathways. However, the biological functions of eRNAs have not been systematically investigated in ccRCC. In this study, we conducted a comprehensive investigation of the role of eRNAs in the onset and management of ccRCC. Patient prognosis-influencing eRNAs and target genes were chosen to construct a predictive signature. On the basis of the median riskscore, ccRCC patients were split into high- and low-risk subgroups. The prediction efficiency was assessed in several cohorts, and multi-omics analysis was carried out to investigate the differences and underlying mechanisms between the high- and low-risk groups. In addition, we investigated its potential to facilitate clinical treatment choices. The riskscore might be used to forecast a patient’s response to immunotherapy and targeted therapy, giving a revolutionary method for selecting treatment regimens with pinpoint accuracy.

The cuproptosis-associated 13 gene signature as a robust predictor for outcome and response to immune- and targeted-therapies in clear cell renal cell carcinoma

Cuproptosis, the newly identified form of regulatory cell death (RCD), results from mitochondrial proteotoxic stress mediated by copper and FDX1. Little is known about significances of cuproptosis in oncogenesis. Here we determined clinical implications of cuproptosis in clear cell renal cell carcinoma (ccRCC). Based on the correlation and survival analyses of cuproptosis-correlated genes in TCGA ccRCC cohort, we constructed a cuproptosis-associated 13 gene signature (CuAGS-13) score system. In both TCGA training and two validation cohorts, when patients were categorized into high- and low-risk groups according to a median score as the cutoff, the CuAGS-13 high-risk group was significantly associated with shorter overall survival (OS) and/or progression-free survival (PFS) independently (P<0.001 for all). The CuAGS-13 score assessment could also predict recurrence and recurrence-free survival of patients at stage I - III with a high accuracy, which outperformed the ccAccB/ClearCode34 model, a well-established molecular predictor for ccRCC prognosis. Moreover, patients treated with immune checkpoint inhibitors (ICIs) acquired complete/partial remissions up to 3-time higher coupled with significantly longer PFS in the CuAGS-13 low- than high-risk groups in both training and validation cohorts of ccRCCs (7.2 - 14.1 vs. 2.1 - 3.0 months, P<0.001). The combination of ICI with anti-angiogenic agent Bevacizumab doubled remission rates in CuAGS-13 high-risk patients while did not improve the efficacy in the low-risk group. Further analyses showed a positive correlation between CuAGS-13 and TIDE scores. We also observed that the CuAGS-13 score assessment accurately predicted patient response to Sunitinib, and higher remission rates in the low-risk group led to longer PFS (Low- vs. high-risk, 13.9 vs. 5.8 months, P = 5.0e-12). Taken together, the CuAGS-13 score assessment serves as a robust predictor for survival, recurrence, and response to ICIs, ICI plus anti-angiogenic drugs and Sunitinib in ccRCC patients, which significantly improves patient stratifications for precision medicine of ccRCC.

APOA1 mRNA and protein in kidney renal clear cell carcinoma correlate with the disease outcome

Renal cancer is one of the most common malignant tumors with high mortality, and kidney renal clear cell carcinoma (KIRC) is the most common type of renal cancer. We attempted to evaluate the clinical and prognostic significance of Apolipoprotein A1 (APOA1) mRNA and protein in KIRC patients. Clinical data along with RNA-sequencing data were downloaded from UCSC Xena. The Human Protein Atlas database was searched to reveal APOA1 protein expression profiles in KIRC and normal renal tissues. The TIMER database was applied to determine the correlations of APOA1 with immune cells and PD-1 and PD-L1 in KIRC. Ninety-one cases of KIRC patients and 93 healthy controls from our hospital were enrolled for clinical validation. Levels of APOA1 mRNA in KIRC tissues (N = 535) are not only lower than the levels in normal renal tissues (N = 117), but also in paired normal renal tissues (N = 72). High expression of APOA1 mRNA at the time of surgery was correlated with worse overall survival (OS) (HR 1.66; p = 0.037) and disease-free survival (DFS) (HR 1.65; p = 0.047), and APOA1 DNA methylation was linked to worse OS (HR 2.1; p = 0.001) rather than DFS (HR 1.12; p = 0.624) in KIRC patients. Concentrations of preoperative serum APOA1 protein were markedly decreased in KIRC patients compared to healthy controls (p < 0.01), and low levels of APOA1 protein predicted less favorable OS than those with high levels (HR = 2.84, p = 0.0407). APOA1 negatively correlated with various immune cell infiltrates and PD-L1 expression (r = − 0.283, p = 2.74e−11) according to the TIMER database. Low levels of APOA1 mRNA at the time of surgery predict favorable survival in KIRC patients. Our results provide insights to identify a novel prognostic index with great clinical utility.