A novel microRNA signature for pathological grading in lung adenocarcinoma based on TCGA and GEO data

Machine learning combining multi-omics data and network algorithms identifies adrenocortical carcinoma prognostic biomarkers

Background: Rare endocrine cancers such as Adrenocortical Carcinoma (ACC) present a serious diagnostic and prognostication challenge. The knowledge about ACC pathogenesis is incomplete, and patients have limited therapeutic options. Identification of molecular drivers and effective biomarkers is required for timely diagnosis of the disease and stratify patients to offer the most beneficial treatments. In this study we demonstrate how machine learning methods integrating multi-omics data, in combination with system biology tools, can contribute to the identification of new prognostic biomarkers for ACC. Methods: ACC gene expression and DNA methylation datasets were downloaded from the Xena Browser (GDC TCGA Adrenocortical Carcinoma cohort). A highly correlated multi-omics signature discriminating groups of samples was identified with the data integration analysis for biomarker discovery using latent components (DIABLO) method. Additional regulators of the identified signature were discovered using Clarivate CBDD (Computational Biology for Drug Discovery) network propagation and hidden nodes algorithms on a curated network of molecular interactions (MetaBase™). The discriminative power of the multi-omics signature and their regulators was delineated by training a random forest classifier using 55 samples, by employing a 10-fold cross validation with five iterations. The prognostic value of the identified biomarkers was further assessed on an external ACC dataset obtained from GEO (GSE49280) using the Kaplan-Meier estimator method. An optimal prognostic signature was finally derived using the stepwise Akaike Information Criterion (AIC) that allowed categorization of samples into high and low-risk groups. Results: A multi-omics signature including genes, micro RNA's and methylation sites was generated. Systems biology tools identified additional genes regulating the features included in the multi-omics signature. RNA-seq, miRNA-seq and DNA methylation sets of features revealed a high power to classify patients from stages I-II and stages III-IV, outperforming previously identified prognostic biomarkers. Using an independent dataset, associations of the genes included in the signature with Overall Survival (OS) data demonstrated that patients with differential expression levels of 8 genes and 4 micro RNA's showed a statistically significant decrease in OS. We also found an independent prognostic signature for ACC with potential use in clinical practice, combining 9-gene/micro RNA features, that successfully predicted high-risk ACC cancer patients. Conclusion: Machine learning and integrative analysis of multi-omics data, in combination with Clarivate CBDD systems biology tools, identified a set of biomarkers with high prognostic value for ACC disease. Multi-omics data is a promising resource for the identification of drivers and new prognostic biomarkers in rare diseases that could be used in clinical practice.

MOCSS: Multi-omics data clustering and cancer subtyping via shared and specific representation learning

Cancer is an extremely complex disease and each type of cancer usually has several different subtypes. Multi-omics data can provide more comprehensive biological information for identifying and discovering cancer subtypes. However, existing unsupervised cancer subtyping methods cannot effectively learn comprehensive shared and specific information of multi-omics data. Therefore, a novel method is proposed based on shared and specific representation learning. For each omics data, two autoencoders are applied to extract shared and specific information, respectively. To reduce redundancy and mutual interference, orthogonality constraint is introduced to separate shared and specific information. In addition, contrastive learning is applied to align the shared information and strengthen their consistency. Finally, the obtained shared and specific information for all samples are used for clustering tasks to achieve cancer subtyping. Experimental results demonstrate that the proposed method can effectively capture shared and specific information of multi-omics data and outperform other state-of-the-art methods on cancer subtyping.

Ferroptosis-Related Genes Are Potential Therapeutic Targets and the Model of These Genes Influences Overall Survival of NSCLC Patients

Background: Lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSCC) are two of the most common subtypes of non-small cell lung cancer (NSCLC), with high mortality rates and rising incidence worldwide. Ferroptosis is a mode of programmed cell death caused by lipid peroxidation, the accumulation of reactive oxygen species, and is dependent on iron. The recent discovery of ferroptosis has provided new insights into tumor development, and the clinical relevance of ferroptosis for tumor therapy is being increasingly appreciated. However, its role in NSCLC remains to be explored. Methods: The clinical and molecular data for 1727 LUAD and LUSCC patients and 73 control individuals were obtained from the Gene Expression Omnibus (GEO) database and the Cancer Genome Atlas (TCGA) database. Gene expression profiles, copy number variations and somatic mutations of 57 ferroptosis-related genes in 1727 tumor samples from the four datasets were used in a univariate Cox analysis and consensus clustering analysis. The biological signatures of each pattern were identified. A ferroptosis score was generated by combining the univariate Cox regression analysis and random forest algorithm followed by principal component analysis (PCA) and further investigated for its predictive and therapeutic value in LUAD and LUSCC. Results: The expression of 57 ferroptosis-related genes in NSCLC patients differed significantly from that of normal subjects. Based on unsupervised clustering of ferroptosis-related genes, we divided all patients into three ferroptosis expression pattern groups, which showed differences in ferroptosis-associated gene expression patterns, immune cell infiltration levels, prognostic characteristics and enriched pathways. Using the differentially expressed genes in the three ferroptosis expression patterns, a set of 17 ferroptosis-related gene prognostic models was established, which clustered all patients in the cohort into a low score group and a high score group, with marked differences in prognosis (p < 0.001). The high ferroptosis score was significantly associated with positive response to radiotherapy (p < 0.001), high T stage (p < 0.001), high N stage (p < 0.001) and high-grade tumor (p < 0.001) characteristics. Conclusions: The 17 ferroptosis-associated genes show great potential for stratifying LUAD and LUSCC patients into high and low risk groups. Interestingly, a high ferroptosis score in LUAD patients was associated with a good prognosis, whereas a similar high ferroptosis score in LUSCC patients was associated with a poor prognosis. Familiarity with the mechanisms underlying ferroptosis and its implications for the treatment of NSCLC, as well as its effect on OS and PFS, may provide guidance and insights in developing new therapeutic targets for NSCLC.

Machine-Learning-Based Late Fusion on Multi-Omics and Multi-Scale Data for Non-Small-Cell Lung Cancer Diagnosis

Differentiation between the various non-small-cell lung cancer subtypes is crucial for providing an effective treatment to the patient. For this purpose, machine learning techniques have been used in recent years over the available biological data from patients. However, in most cases this problem has been treated using a single-modality approach, not exploring the potential of the multi-scale and multi-omic nature of cancer data for the classification. In this work, we study the fusion of five multi-scale and multi-omic modalities (RNA-Seq, miRNA-Seq, whole-slide imaging, copy number variation, and DNA methylation) by using a late fusion strategy and machine learning techniques. We train an independent machine learning model for each modality and we explore the interactions and gains that can be obtained by fusing their outputs in an increasing manner, by using a novel optimization approach to compute the parameters of the late fusion. The final classification model, using all modalities, obtains an F1 score of 96.81±1.07, an AUC of 0.993±0.004, and an AUPRC of 0.980±0.016, improving those results that each independent model obtains and those presented in the literature for this problem. These obtained results show that leveraging the multi-scale and multi-omic nature of cancer data can enhance the performance of single-modality clinical decision support systems in personalized medicine, consequently improving the diagnosis of the patient.

A narrative review of artificial intelligence-assisted histopathologic diagnosis and decision-making for non-small cell lung cancer: achievements and limitations

Objective

To summarize the current evidence regarding the applications, workflow, and limitations of artificial intelligence (AI) in the management of patients pathologically-diagnosed with lung cancer.

Background

Lung cancer is one of the most common cancers and the leading cause of cancer-related deaths worldwide. AI technologies have been applied to daily medical workflow and have achieved an excellent performance in predicting histopathologic subtypes, analyzing gene mutation profiles, and assisting in clinical decision-making for lung cancer treatment. More advanced deep learning for classifying pathologic images with minimal human interactions has been developed in addition to the conventional machine learning scheme.

Methods

Studies were identified by searching databases, including PubMed, EMBASE, Web of Science, and Cochrane Library, up to February 2021 without language restrictions.

Conclusions

A number of studies have evaluated AI pipelines and confirmed that AI is robust and efficacious in lung cancer diagnosis and decision-making, demonstrating that AI models are a useful tool for assisting oncologists in health management. Although several limitations that pose an obstacle for the widespread use of AI schemes persist, the unceasing refinement of AI techniques is poised to overcome such problems. Thus, AI technology is a promising tool for use in diagnosing and managing lung cancer.

Development and Validation of Lactate Metabolism-Related lncRNA Signature as a Prognostic Model for Lung Adenocarcinoma

BACKGROUND

Lung cancer has been a prominent research focus in recent years due to its role in cancer-related fatalities globally, with lung adenocarcinoma (LUAD) being the most prevalent histological form. Nonetheless, no signature of lactate metabolism-related long non-coding RNAs (LMR-lncRNAs) has been developed for patients with LUAD. Accordingly, we aimed to develop a unique LMR-lncRNA signature to determine the prognosis of patients with LUAD.

METHOD

The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases were utilized to derive the lncRNA expression patterns. Identification of LMR-lncRNAs was accomplished by analyzing the co-expression patterns between lncRNAs and LMR genes. Subsequently, the association between lncRNA levels and survival outcomes was determined to develop an effective signature. In the TCGA cohort, Cox regression was enlisted to build an innovative signature consisting of three LMR-lncRNAs, which was validated in the GEO validation cohort. GSEA and immune infiltration analysis were conducted to investigate the functional annotation of the signature and the function of each type of immune cell.

RESULTS

Fourteen differentially expressed LMR-lncRNAs were strongly correlated with the prognosis of patients with LUAD and collectively formed a new LMR-lncRNA signature. The patients could be categorized into two cohorts based on their LMR-lncRNA signatures: a low-risk and high-risk group. The overall survival of patients with LUAD in the high-risk group was considerably lower than those in the low-risk group. Using Cox regression, this signature was shown to have substantial potential as an independent prognostic factor, which was further confirmed in the GEO cohort. Moreover, the signature could anticipate survival across different groups based on stage, age, and gender, among other variables. This signature also correlated with immune cell infiltration (including B cells, neutrophils, CD4⁺ T cells, CD8⁺ T cells, etc.) as well as the immune checkpoint blockade target CTLA-4.

CONCLUSION

We developed and verified a new LMR-lncRNA signature useful for anticipating the survival of patients with LUAD. This signature could give potentially critical insight for immunotherapy interventions in patients with LUAD.

An 11-Gene Signature Risk-Prediction Model Based on Prognosis-Related miRNAs and Their Target Genes in Lung Adenocarcinoma

Aberrant expression of microRNAs may affect tumorigenesis and progression by regulating their target genes. This study aimed to construct a risk model for predicting the prognosis of patients with lung adenocarcinoma (LUAD) based on differentially expressed microRNA-regulated target genes. The miRNA sequencing data, RNA sequencing data, and patients’ LUAD clinical data were downloaded from the The Cancer Genome Atlas (TCGA) database. Differentially expressed miRNAs and genes were screened out by combining differential analysis with LASSO regression analysis to further screen out miRNAs associated with patients’ prognosis, and target gene prediction was performed for these miRNAs using a target gene database. Overlapping gene screening was performed for target genes and differentially expressed genes. LASSO regression analysis and survival analysis were then used to identify key genes. Risk score equations for prognostic models were established using multifactorial COX regression analysis to construct survival prognostic models, and the accuracy of the models was evaluated using subject working characteristic curves. The groups were divided into high- and low-risk groups according to the median risk score, and the correlation with the clinicopathological characteristics of the patients was observed. A total of 123 up-regulated miRNAs and 22 down-regulated miRNAs were obtained in this study. Five prognosis-related miRNAs were screened using LASSO regression analysis and Kaplan-Meier method validation, and their target genes were screened with the overlap of differentially expressed genes before multifactorial COX analysis finally resulted in an 11-gene risk model for predicting patient prognosis. The area under the ROC curve proved that the model has high accuracy. The 11-gene risk-prediction model constructed in this study may be an effective predictor of prognosis.

A novel microRNA-based signature predicts prognosis among nasopharyngeal cancer patients

IMPACT STATEMENT

Nasopharyngeal cancer is one of the most common malignant tumors in the head and neck. Identification of promising miRNA biomarkers might benefit a lot to the detection of nasopharyngeal carcinoma. A three-miRNA signature (has-miR-142-3p, has-miR-29c, and has-miR-30e) was obviously associated with the overall survival of nasopharyngeal carcinoma patients. The model has better clinical independence and has better clinical prediction effect when combined with clinical characteristics. Our results revealed that a three-miRNA signature was a potential novel prognostic biomarker for nasopharyngeal carcinoma.

SARS-CoV-2 may regulate cellular responses through depletion of specific host miRNAs

Cold viruses have generally been considered fairly innocuous until the appearance of the severe acute respiratory coronavirus 2 (SARS-CoV-2) in 2019, which caused the coronavirus disease 2019 (COVID-19) global pandemic. Two previous viruses foreshadowed that a coronavirus could potentially have devastating consequences in 2002 [severe acute respiratory coronavirus (SARS-CoV)] and in 2012 [Middle East respiratory syndrome coronavirus (MERS-CoV)]. The question that arises is why these viruses are so different from the relatively harmless cold viruses. On the basis of an analysis of the current literature and using bioinformatic approaches, we examined the potential human miRNA interactions with the SARS-CoV-2's genome and compared the miRNA target sites in seven coronavirus genomes that include SARS-CoV-2, MERS-CoV, SARS-CoV, and four nonpathogenic coronaviruses. Here, we discuss the possibility that pathogenic human coronaviruses, including SARS-CoV-2, could modulate host miRNA levels by acting as miRNA sponges to facilitate viral replication and/or to avoid immune responses.