A large cohort study identifying a novel prognosis prediction model for lung adenocarcinoma through machine learning strategies

Artificial intelligence in healthcare: a mastery

There is a vast development of artificial intelligence (AI) in recent years. Computational technology, digitized data collection and enormous advancement in this field have allowed AI applications to penetrate the core human area of specialization. In this review article, we describe current progress achieved in the AI field highlighting constraints on smooth development in the field of medical AI sector, with discussion of its implementation in healthcare from a commercial, regulatory and sociological standpoint. Utilizing sizable multidimensional biological datasets that contain individual heterogeneity in genomes, functionality and milieu, precision medicine strives to create and optimize approaches for diagnosis, treatment methods and assessment. With the arise of complexity and expansion of data in the health-care industry, AI can be applied more frequently. The main application categories include indications for diagnosis and therapy, patient involvement and commitment and administrative tasks. There has recently been a sharp rise in interest in medical AI applications due to developments in AI software and technology, particularly in deep learning algorithms and in artificial neural network (ANN). In this overview, we enlisted the major categories of issues that AI systems are ideally equipped to resolve followed by clinical diagnostic tasks. It also includes a discussion of the future potential of AI, particularly for risk prediction in complex diseases, and the difficulties, constraints and biases that must be meticulously addressed for the effective delivery of AI in the health-care sector.

Machine learning-based model for CD4+ conventional T cell genes to predict survival and immune responses in colorectal cancer

Globally, CRC ranks as a principal cause of mortality, with projections indicating a substantial rise in both incidence and mortality by the year 2040. The immunological responses to cancer heavily rely on the function of CD4Tconv. Despite this critical role, prognostic studies on CRC-related CD4Tconv remain insufficient. In this investigation, transcriptomic and clinical data were sourced from TCGA and GEO. Initially, we pinpointed CD4TGs using single-cell datasets. Prognostic genes were then isolated through univariate Cox regression analysis. Building upon this, 101 machine learning algorithms were employed to devise a novel risk assessment framework, which underwent rigorous validation using Kaplan-Meier survival analysis, univariate and multivariate Cox regression, time-dependent ROC curves, nomograms, and calibration plots. Furthermore, GSEA facilitated the examination of these genes' potential roles. The RS derived from this model was also analyzed for its implications in the TME, and its potential utility in immunotherapy and chemotherapy contexts. A novel prognostic model was developed, utilizing eight CD4TGs that are significantly linked to the outcomes of patients with CRC. This model's RS showcased remarkable predictive reliability for the overall survival rates of CRC patients and strongly correlated with malignancy levels. RS serves as an autonomous prognostic indicator, capable of accurately forecasting patient prognoses. Based on the median value of RS, patients were categorized into subgroups of high and low risk. The subgroup with higher risk demonstrated increased immune infiltration and heightened activity of genes associated with immunity. This investigation's establishment of a CD4TGs risk model introduces novel biomarkers for the clinical evaluation of CRC risks. These biomarkers may enhance therapeutic approaches and, in turn, elevate the clinical outcomes for patients with CRC by facilitating an integrated treatment strategy.

Diagnostic value of immune-related biomarker FAM83A in differentiating malignant from benign pleural effusion in lung adenocarcinoma

BACKGROUND

Malignant pleural effusion (MPE) is frequently observed in patients with advanced lung adenocarcinoma (LUAD). Pleural fluid cytology is a less invasive procedure compared to pleural biopsy. Therefore, it is crucial to identify novel effective biomarkers for LUAD-associated pleural fluid cytology.

METHODS

The RNA sequencing (RNA-Seq) and clinical data of LUAD cases were downloaded from TCGA and OncoSG databases. Differential gene expression analysis, survival analysis and immune cell infiltration analysis were performed on the LUAD datasets. The expression levels of FAM83A, TFF-1, and NapsinA in 94 paired LUAD and adjacent normal tissues, and in the pleural effusion specimens of 40 LUAD and 21 non-neoplastic patients were evaluated by immunohistochemistry.

RESULTS

FAM83A expression levels were significantly different between the LUAD and normal tissue datasets, and correlated with overall or disease-free survival, and histological grade of the tumors. Furthermore, the in-situ expression of FAM83A was higher in 89/94 LUAD tissues compared to the paired normal tissues. FAM83A expression was significantly correlated with immune cell infiltration, and showed a positive association with macrophage infiltration. In addition, FAM83A staining was positive in 37 LUAD pleural effusion samples, and negative in 20 non-neoplastic pleural effusion samples. The expression pattern of FAM83A in the pleural effusion of LUAD patients was relatively consistent with that of TFF-1 and NapsinA, and even stronger in some specimens that were weakly positive or negative for TTF1/NapsinA.

CONCLUSIONS

FAM83A is a promising immune-related biomarker in LUAD biopsy specimens and pleural fluid, and can distinguish between malignant and benign pleural effusion.

UPP1 promotes lung adenocarcinoma progression through the induction of an immunosuppressive microenvironment

The complexity of the tumor microenvironment (TME) is a crucial factor in lung adenocarcinoma (LUAD) progression. To gain deeper insights into molecular mechanisms of LUAD, we perform an integrative single-cell RNA sequencing (scRNA-seq) data analysis of 377,574 cells from 117 LUAD patient samples. By linking scRNA-seq data with bulk gene expression data, we identify a cluster of prognostic-related UPP1high tumor cells. These cells, primarily situated at the invasive front of tumors, display a stronger association with the immunosuppressive components in the TME. Our cytokine array analysis reveals that the upregulation of UPP1 in tumor cells leads to the increased release of various immunosuppressive cytokines, with TGF-β1 being particularly prominent. Furthermore, this UPP1 upregulation also elevates the expression of PD-L1 through the PI3K/AKT/mTOR pathway, which contributes to the suppression of CD8 + T cells. Cytometry by time-of-flight (CyTOF) analysis provides additional evidence of the role of UPP1 in shaping the immunosuppressive nature of the TME. Using patient-derived organoids (PDOs), we discover that UPP1high tumors exhibit relatively increased sensitivity to Bosutinib and Dasatinib. Collectively, our study highlights the immunosuppressive role of UPP1 in LUAD, and these findings may provide insights into the molecular features of LUAD and facilitate the development of personalized treatment strategies.

AI/ML advances in non-small cell lung cancer biomarker discovery

Lung cancer is the leading cause of cancer deaths among both men and women, representing approximately 25% of cancer fatalities each year. The treatment landscape for non-small cell lung cancer (NSCLC) is rapidly evolving due to the progress made in biomarker-driven targeted therapies. While advancements in targeted treatments have improved survival rates for NSCLC patients with actionable biomarkers, long-term survival remains low, with an overall 5-year relative survival rate below 20%. Artificial intelligence/machine learning (AI/ML) algorithms have shown promise in biomarker discovery, yet NSCLC-specific studies capturing the clinical challenges targeted and emerging patterns identified using AI/ML approaches are lacking. Here, we employed a text-mining approach and identified 215 studies that reported potential biomarkers of NSCLC using AI/ML algorithms. We catalogued these studies with respect to BEST (Biomarkers, EndpointS, and other Tools) biomarker sub-types and summarized emerging patterns and trends in AI/ML-driven NSCLC biomarker discovery. We anticipate that our comprehensive review will contribute to the current understanding of AI/ML advances in NSCLC biomarker research and provide an important catalogue that may facilitate clinical adoption of AI/ML-derived biomarkers.

Improving Pancreatic Cyst Management: Artificial Intelligence-Powered Prediction of Advanced Neoplasms through Endoscopic Ultrasound-Guided Confocal Endomicroscopy

Despite the increasing rate of detection of incidental pancreatic cystic lesions (PCLs), current standard-of-care methods for their diagnosis and risk stratification remain inadequate. Intraductal papillary mucinous neoplasms (IPMNs) are the most prevalent PCLs. The existing modalities, including endoscopic ultrasound and cyst fluid analysis, only achieve accuracy rates of 65-75% in identifying carcinoma or high-grade dysplasia in IPMNs. Furthermore, surgical resection of PCLs reveals that up to half exhibit only low-grade dysplastic changes or benign neoplasms. To reduce unnecessary and high-risk pancreatic surgeries, more precise diagnostic techniques are necessary. A promising approach involves integrating existing data, such as clinical features, cyst morphology, and data from cyst fluid analysis, with confocal endomicroscopy and radiomics to enhance the prediction of advanced neoplasms in PCLs. Artificial intelligence and machine learning modalities can play a crucial role in achieving this goal. In this review, we explore current and future techniques to leverage these advanced technologies to improve diagnostic accuracy in the context of PCLs.

Predicting lung cancer survival based on clinical data using machine learning: A review

Machine learning has gained popularity in predicting survival time in the medical field. This review examines studies utilizing machine learning and data-mining techniques to predict lung cancer survival using clinical data. A systematic literature review searched MEDLINE, Scopus, and Google Scholar databases, following reporting guidelines and using the COVIDENCE system. Studies published from 2000 to 2023 employing machine learning for lung cancer survival prediction were included. Risk of bias assessment used the prediction model risk of bias assessment tool. Thirty studies were reviewed, with 13 (43.3%) using the surveillance, epidemiology, and end results database. Missing data handling was addressed in 12 (40%) studies, primarily through data transformation and conversion. Feature selection algorithms were used in 19 (63.3%) studies, with age, sex, and N stage being the most chosen features. Random forest was the predominant machine learning model, used in 17 (56.6%) studies. While the number of lung cancer survival prediction studies is limited, the use of machine learning models based on clinical data has grown since 2012. Consideration of diverse patient cohorts and data pre-processing are crucial. Notably, most studies did not account for missing data, normalization, scaling, or standardized data, potentially introducing bias. Therefore, a comprehensive study on lung cancer survival prediction using clinical data is needed, addressing these challenges.

Autoencoder-based multimodal prediction of non-small cell lung cancer survival

The ability to accurately predict non-small cell lung cancer (NSCLC) patient survival is crucial for informing physician decision-making, and the increasing availability of multi-omics data offers the promise of enhancing prognosis predictions. We present a multimodal integration approach that leverages microRNA, mRNA, DNA methylation, long non-coding RNA (lncRNA) and clinical data to predict NSCLC survival and identify patient subtypes, utilizing denoising autoencoders for data compression and integration. Survival performance for patients with lung adenocarcinoma (LUAD) and squamous cell carcinoma (LUSC) was compared across modality combinations and data integration methods. Using The Cancer Genome Atlas data, our results demonstrate that survival prediction models combining multiple modalities outperform single modality models. The highest performance was achieved with a combination of only two modalities, lncRNA and clinical, at concordance indices (C-indices) of 0.69 ± 0.03 for LUAD and 0.62 ± 0.03 for LUSC. Models utilizing all five modalities achieved mean C-indices of 0.67 ± 0.04 and 0.63 ± 0.02 for LUAD and LUSC, respectively, while the best individual modality performance reached C-indices of 0.64 ± 0.03 for LUAD and 0.59 ± 0.03 for LUSC. Analysis of biological differences revealed two distinct survival subtypes with over 900 differentially expressed transcripts.

The activity of cuproptosis pathway calculated by AUCell algorithm was employed to construct cuproptosis landscape in lung adenocarcinoma

Cuproptosis is a recently described copper-dependent cell death pathway. Consequently, there are still few studies on lung adenocarcinoma (LUAD)-related cuproptosis, and we aimed to deepen in this matter. In this study, data from 503 patients with lung cancer from the TCGA-LUAD cohort data collection and 11 LUAD single-cells from GSE131907 as well as from 10 genes associated with cuproptosis were analyzed. The AUCell R package was used to determine the copper-dependent cell death pathway activity for each cell subpopulation, calculate the CellChat score, and display cell communication for each cell subpopulation. The PROGENy score was calculated to show the scores of tumor-related pathways in different cell populations. GO and KEGG analyses were used to calculate pathway activity. Univariate COX and random forest analyses were used to screen prognosis-associated genes and construct models. The ssGSEA and xCell algorithms were used to calculate the immunocyte infiltration score. Based on data from the GDSC database, the drug sensitivity score was calculated using oncoPredict. Finally, in vitro experiments were performed to determine the role of TLE1, the most important gene in the prognostic model. The 11 LUAD single-cell samples were classified into 8 different cell populations, from which epithelial cells showed the highest copper-dependent cell death pathway activity. Epithelial cell subsets were significantly positively correlated with MAKP, hypoxia, and other pathways. In addition, cell subgroup communication showed highly active collagen and APP pathways. Using the Findmark algorithm, differentially expressed genes (DEGs) between epithelial and other cell types were identified. Combined with the bulk data in the TCGA-LUAD database, DEGs were enriched in pathways such as EGFR tyrosine kinase inhibitor resistance, Hippo signaling pathway, and tight junction. Subsequently, we selected 4 genes (out of 112) with prognostic significance, ANKRD29, RHOV, TLE1, and NPAS2, and used them to construct a prognostic model. The high- and low-risk groups, distinguished by the median risk score, showed significantly different prognoses. Finally, we chose TLE1 as a biomarker based on the relative importance score in the prognostic model. In vitro experiments showed that TLE1 promotes tumor proliferation and migration and inhibits apoptosis.

Identification of New Prognostic Genes and Construction of a Prognostic Model for Lung Adenocarcinoma

Lung adenocarcinoma (LUAD) is a rapidly progressive malignancy, and its mortality rate is very high. In this study, we aimed at finding novel prognosis-related genes and constructing a credible prognostic model to improve the prediction for LUAD patients. Differential gene expression, mutant subtype, and univariate Cox regression analyses were conducted with the dataset from the Cancer Genome Atlas (TCGA) database to screen for prognostic features. These features were employed in the following multivariate Cox regression analysis and the produced prognostic model included the stage and expression of SMCO2, SATB2, HAVCR1, GRIA1, and GALNT4, as well as mutation subtypes of TP53. The exactness of the model was confirmed by an overall survival (OS) analysis and disease-free survival (DFS) analysis, which indicated that patients in the high-risk group had a poorer prognosis compared to those in the low-risk group. The area under the receiver operating characteristic curve (AUC) was 0.793 in the training group and 0.779 in the testing group. The AUC of tumor recurrence was 0.778 in the training group and 0.815 in the testing group. In addition, the number of deceased patients increased as the risk scores raised. Furthermore, the knockdown of prognostic gene HAVCR1 suppressed the proliferation of A549 cells, which supports our prognostic model that the high expression of HAVCR1 predicts poor prognosis. Our work created a reliable prognostic risk score model for LUAD and provided potential prognostic biomarkers.

An eleven-gene risk model associated with lymph node metastasis predicts overall survival in lung adenocarcinoma

Lung adenocarcinoma (LUAD) occupies major causes of tumor death. Identifying potential prognostic risk genes is crucial to predict the overall survival of patients with LUAD. In this study, we constructed and proved an 11-gene risk signature. This prognostic signature divided LUAD patients into low- and high-risk groups. The model outperformed in prognostic accuracy at varying follow-up times (AUC for 3 years: 0.699, 5 years: 0.713, and 7 years: 0.716). Two GEO datasets also indicate the great accuracy of the risk signature (AUC = 782 and 771, respectively). Multivariate analysis identified 4 independent risk factors including stage N (HR 1.320, 95% CI 1.102-1.581, P = 0.003), stage T (HR 3.159, 95% CI 1.920-3.959, P < 0.001), tumor status (HR 5.688, 95% CI 3.883-8.334, P < 0.001), and the 11-gene risk model (HR 2.823, 95% CI 1.928-4.133, P < 0.001). The performance of the nomogram was good in the TCGA database (AUC = 0.806, 0.798, and 0.818 for 3-, 5- and 7-year survival). The subgroup analysis in different age, gender, tumor status, clinical stage, and recurrence stratifications indicated that the accuracy was high in different subgroups (all P < 0.05). Briefly, our work established an 11-gene risk model and a nomogram merging the model with clinicopathological characteristics to facilitate individual prediction of LUAD patients for clinicians.

Artificial Intelligence in the Diagnosis and Treatment of Pancreatic Cystic Lesions and Adenocarcinoma

Pancreatic cancer is projected to become the second leading cause of cancer-related mortality in the United States by 2030. This is in part due to the paucity of reliable screening and diagnostic options for early detection. Amongst known pre-malignant pancreatic lesions, pancreatic intraepithelial neoplasia (PanIN) and intraductal papillary mucinous neoplasms (IPMNs) are the most prevalent. The current standard of care for the diagnosis and classification of pancreatic cystic lesions (PCLs) involves cross-sectional imaging studies and endoscopic ultrasound (EUS) and, when indicated, EUS-guided fine needle aspiration and cyst fluid analysis. However, this is suboptimal for the identification and risk stratification of PCLs, with accuracy of only 65-75% for detecting mucinous PCLs. Artificial intelligence (AI) is a promising tool that has been applied to improve accuracy in screening for solid tumors, including breast, lung, cervical, and colon cancer. More recently, it has shown promise in diagnosing pancreatic cancer by identifying high-risk populations, risk-stratifying premalignant lesions, and predicting the progression of IPMNs to adenocarcinoma. This review summarizes the available literature on artificial intelligence in the screening and prognostication of precancerous lesions in the pancreas, and streamlining the diagnosis of pancreatic cancer.

A Transendothelial Leukocyte Transmigration Model Based on Computational Fluid Dynamics and BP Neural Network

The mechanism of immune infiltration involving immune cells is closely related to various diseases. A key issue in immune infiltration is the transendothelial transmigration of leukocytes. Previous studies have primarily interpreted the leukocyte infiltration of from biomedical perspective. The physical mechanism of leukocyte infiltration remains to be explored. By integrating the immune cell transmigration computational fluid dynamics (CFD) data, the paper builds a time-dependent leukocyte transmigration prediction model based on the bio-inspired methods, namely back propagation neural networks (BPNN) model. The model can efficiently predict the immune cell transmigration in a special microvascular environment, and obtain good prediction accuracy. The model accurately predicted the cell movement and flow field changes during the transmigration. In the test data set, it has high prediction accuracy for cell deformation, motion velocity and flow lift forces during downstream motion, and maintains a good prediction accuracy for drag force. The two prediction models achieved the prediction of leukocyte transmigration in a specific microvascular environment and maintained a high prediction accuracy, indicating the feasibility and robustness of the BPNN model applied to the prediction of immune cell infiltration. Compared with traditional CFD simulations, BPNN models avoid complex and time-dependent physical modeling and computational processes.

Effect of PPP1R14D gene high expression in lung adenocarcinoma knocked out on proliferation and apoptosis of DMS53 cell

PURPOSE

Globally, lung cancer remains the most commonly diagnosed cancer and the leading cause of cancer-related mortality. Lung adenocarcinoma (LUAD) is a common subtype of lung cancer and carries a poor prognosis. Treatment outcomes biomarkers in LUAD are critical, and there is currently a paucity of data; therefore, there is a need for novel biomarkers and newer therapeutic targets.

METHODS

Bayesian analysis was used to obtain the whole-genome t value of LUAD. Gene set enrichment analysis (GSEA) was conducted to obtain the normalized enrichment scores (NES) of the whole genome, and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway was analyzed using the Gene Set Analysis Toolkit. Herein, we investigated the PPP1R14D expression level at the protein level in LUAD and the impact of PPP1R14D knockdown on the proliferation and apoptosis of LUAD cells in vitro.

RESULTS

A total of 483 LUAD samples and 59 normal control samples were included, and 904 differentially expressed genes (DEGs) and 504 LUAD-related genes reported in the literature were obtained. The DEGs showed that PPP1R14D was the most significantly up-regulated gene. Western blot of 30 cases of LUAD tissue and adjacent normal tissue also found that PPP1R14D was significantly highly expressed in cancer tissues. Lentivirus-mediated shRNA strategy effectively inhibited PPP1R14D expression in human LUAD cells DMS53, while PPP1R14D knockdown induced apoptosis and cell proliferation in DMS53 cells.

CONCLUSION

Abnormally up-regulated PPP1R14D promotes the survival and proliferation of tumor cells in human LUAD and may serve as a therapeutic and diagnostic target for LUAD.

One Ferroptosis-Related Gene-Pair Signature Serves as an Original Prognostic Biomarker in Lung Adenocarcinoma

Lung adenocarcinoma is the most common histological subtype of lung cancer which causes the largest number of deaths worldwide. Exploring reliable prognostic biomarkers based on biological behaviors and molecular mechanisms is essential for predicting prognosis and individualized treatment strategies. Ferroptosis is a recently discovered type of regulated cell death. We downloaded ferroptosis-related genes from the literature and collected transcriptome profiles of lung adenocarcinoma from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) to construct ferroptosis-related gene-pair matrixes. Then, we performed the least absolute shrinkage and selection operator regression to build our prognostic ferroptosis-related gene-pair index (FRGPI) in TCGA training matrix. Our study validated FRGPI through ROC curves, Kaplan–Meier methods, and Cox hazard analyses in TCGA and GEO cohorts. The optimal cut-off 0.081 stratified patients into low- and high-FRGPI groups. Also, the low-FRGPI group had a significantly better prognosis than the high-FRGPI group. For further study, we analyzed differentially expressed ferroptosis-related genes between high- and low-FRGPI groups. Gene set enrichment analysis (GSEA) enrichment maps indicated that “cell cycle,” “DNA replication,” “proteasome,” and “the p53 signaling pathway” were significantly enriched in the high-FRGPI group. The high-FRGPI group also presented higher infiltration of M1 macrophages. Meanwhile, there were few differences in adaptive immune responses between high- and low-FRGPI groups. In conclusion, FRGPI was an independent prognostic biomarker which might be beneficial for guiding individualized tumor therapy.

Artificial intelligence-assisted decision making for prognosis and drug efficacy prediction in lung cancer patients: a narrative review

Objective

In this review, we aim to present frontier studies in patients with lung cancer as it related to artificial intelligence (AI)-assisted decision-making and summarize the latest advances, challenges and future trend in this field.

Background

Despite increasing survival rate in cancer patients over the last decades, lung cancer remains one of the leading causes of death worldwide. The early diagnosis, accurate evaluation and individualized treatment are vital approaches to improve the survival rate of patients with lung cancer. Thus, decision making based on these approaches requires accuracy and efficiency beyond manpower. Recent advances in AI and precision medicine have provided a fertile environment for the development of AI-based models. These models have the potential to assist radiologists and oncologists in detecting lung cancer, predicting prognosis and developing personalized treatment plans for better outcomes of the patients.

Methods

We searched literature from 2000 through July 31^th, 2021 in Medline/PubMed, the Web of Science, the Cochrane Library, ACM Digital Library, INSPEC and EMBASE. Key words such as “artificial intelligence”, “AI”, “deep learning”, “lung cancer”, “NSCLC”, “SCLC” were combined to identify related literatures. These literatures were then selected by two independent authors. Articles chosen by only one author will be examined by another author to determine whether this article was relative and valuable. The selected literatures were read by all authors and discussed to draw reliable conclusions.

Conclusions

AI, especially for those based on deep learning and radiomics, is capable of assisting clinical decision making from many aspects, for its quantitatively interpretation of patients’ information and its potential to deal with the dynamics, individual differences and heterogeneity of lung cancer. Hopefully, remaining problems such as insufficient data and poor interpretability may be solved to put AI-based models into clinical practice.

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.

Identification of CD8+ T Cell Related Biomarkers in Ovarian Cancer

Background: Immunotherapy is a promising strategy for ovarian cancer (OC), and this study aims to identify biomarkers related to CD8⁺ T cell infiltration to further discover the potential therapeutic target. Methods: Three datasets with OC transcriptomic data were downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. Two immunotherapy treated cohorts were obtained from the Single Cell Portal and Mariathasan's study. The infiltration fraction of immune cells was quantified using three different algorithms, Cell-type Identification by Estimating Relative Subsets of RNA Transcripts (CIBERSORT), and microenvironment cell populations counter (MCPcounter), and single-sample GSEA (ssGSEA). Weighted gene co-expression network analysis (WGCNA) was applied to identify the co-expression modules and related genes. The nonnegative matrix factorization (NMF) method was proposed for sample classification. The mutation analysis was conducted using the "maftools" R package. Key molecular markers with implications for prognosis were screened by univariate COX regression analysis and K-M survival analysis, which were further determined by the receiver operating characteristic (ROC) curve. Results: A total of 313 candidate CD8⁺ T cell-related genes were identified by taking the intersection from the TCGA-OV and GSE140082 cohorts. The NMF clustering analysis suggested that patients in the TCGA-OV cohort were divided into two clusters and the Cluster 1 group showed a worse prognosis. In contrast, Cluster 2 had higher amounts of immune cell infiltration, elevated ssGSEA scores in immunotherapy, and a higher mutation burden. CSMD3, MACF1, PDE4DIP, and OBSCN were more frequently mutated in Cluster 1, while SYNE2 was more frequently mutated in Cluster 2. CD38 and CXCL13 were identified by univariate COX regression analysis and K-M survival analysis in the TCGA-OV cohort, which were further externally validated in GSE140082 and GSE32062. Of note, patients with lower CXCL13 expression showed a worse prognosis and the CR/PR group had a higher expression of CXCL13 in two immunotherapy treated cohorts. Conclusion: OC patients with different CD8⁺ T cell infiltration had distinct clinical prognoses. CXCL13 might be a potential therapeutic target for the treatment of OC.

A narrative review of prognosis prediction models for non-small cell lung cancer: what kind of predictors should be selected and how to improve models?

Objective

To discover potential predictors and explore how to build better models by summarizing the existing prognostic prediction models of non-small cell lung cancer (NSCLC).

Background

Research on clinical prediction models of NSCLC has experienced explosive growth in recent years. As more predictors of prognosis are discovered, the choice of predictors to build models is particularly important, and in the background of more applications of next-generation sequencing technology, gene-related predictors are widely used. As it is more convenient to obtain samples and follow-up data, the prognostic model is preferred by researchers.

Methods

PubMed and the Cochrane Library were searched using the items “NSCLC”, “prognostic model”, “prognosis prediction”, and “survival prediction” from 1 January 1980 to 5 May 2021. Reference lists from articles were reviewed and relevant articles were identified.

Conclusions

The performance of gene-related models has not obviously improved. Relative to the innovation and diversity of predictors, it is more important to establish a highly stable model that is convenient for clinical application. Most of the prevalent models are highly biased and referring to PROBAST at the beginning of the study may be able to significantly control the bias. Existing models should be validated in a large external dataset to make a meaningful comparison.

Identifying Molecular Biomarkers for Diseases With Machine Learning Based on Integrative Omics

Molecular biomarkers are certain molecules or set of molecules that can be of help for diagnosis or prognosis of diseases or disorders. In the past decades, thanks to the advances in high-throughput technologies, a huge amount of molecular 'omics' data, e.g., transcriptomics and proteomics, have been accumulated. The availability of these omics data makes it possible to screen biomarkers for diseases or disorders. Accordingly, a number of computational approaches have been developed to identify biomarkers by exploring the omics data. In this review, we present a comprehensive survey on the recent progress of identification of molecular biomarkers with machine learning approaches. Specifically, we categorize the machine learning approaches into supervised, un-supervised and recommendation approaches, where the biomarkers including single genes, gene sets and small gene networks. In addition, we further discuss potential problems underlying bio-medical data that may pose challenges for machine learning, and provide possible directions for future biomarker identification.

Hypoxia-related prognostic model in bladder urothelial reflects immune cell infiltration

Hypoxia is a common feature of tumor microenvironment (TME). This study aims to establish the genetic features related to hypoxia in Bladder urothelial carcinoma (BLCA) and investigate the potential correlation with hypoxia in the TME and immune cells. We established a BLCA outcome model using the hypoxia-related genes from The Cancer Genome Atlas using regression analysis and verified the model using the Gene Expression Omnibus GSE32894 cohort. We measured the effect of each gene in the hypoxia-related risk model using the Human Protein Atlas website. The predictive abilities were compared using the area under the receiver operating characteristic curves. Gene Set Enrichment Analysis was utilized for indicating enrichment pathways. We analyzed immune cell infiltration between risk groups using the CIBERSORT method. The indicators related to immune status between the two groups were also analyzed. The findings indicated that the high-risk group had better outcomes than the low-risk group in the training and validation sets. Each gene in the model affected the survival of BLCA patients. Our hypoxia-related risk model had better performance compared to other hypoxia-related markers (HIF-1α and GLUT-1). The high-risk group was enriched in immune-related pathways. The expression of chemokines and immune cell markers differed significantly between risk groups. Immune checkpoints were more highly expressed in the high-risk group. These findings suggest that the hypoxia-related risk model predicts patients' outcomes and immune status in BLCA risk groups. Our findings may contribute to the treatment of BLCA.

Characteristic of molecular subtypes in lung adenocarcinoma based on m6A RNA methylation modification and immune microenvironment

BACKGROUND

Lung adenocarcinoma (LUAD) is a major subtype of lung cancer and closely associated with poor prognosis. N6-methyladenosine (m6A), one of the most predominant modifications in mRNAs, is found to participate in tumorigenesis. However, the potential function of m6A RNA methylation in the tumor immune microenvironment is still murky.

METHODS

The gene expression profile cohort and its corresponding clinical data of LUAD patients were downloaded from TCGA database and GEO database. Based on the expression of 21 m6A regulators, we identified two distinct subgroups by consensus clustering. The single-sample gene-set enrichment analysis (ssGSEA) algorithm was conducted to quantify the relative abundance of the fraction of 28 immune cell types. The prognostic model was constructed by Lasso Cox regression. Survival analysis and receiver operating characteristic (ROC) curves were used to evaluate the prognostic model.

RESULT

Consensus classification separated the patients into two clusters (clusters 1 and 2). Those patients in cluster 1 showed a better prognosis and were related to higher immune scores and more immune cell infiltration. Subsequently, 457 differentially expressed genes (DEGs) between the two clusters were identified, and then a seven-gene prognostic model was constricted. The survival analysis showed poor prognosis in patients with high-risk score. The ROC curve confirmed the predictive accuracy of this prognostic risk signature. Besides, further analysis indicated that there were significant differences between the high-risk and low-risk groups in stages, status, clustering subtypes, and immunoscore. Low-risk group was related to higher immune score, more immune cell infiltration, and lower clinical stages. Moreover, multivariate analysis revealed that this prognostic model might be a powerful prognostic predictor for LUAD. Ultimately, the efficacy of this prognostic model was successfully validated in several external cohorts (GSE30219, GSE50081 and GSE72094).

CONCLUSION

Our study provides a robust signature for predicting patients' prognosis, which might be helpful for therapeutic strategies discovery of LUAD.

Transcriptomics based multi-dimensional characterization and drug screen in esophageal squamous cell carcinoma

BACKGROUND

Esophageal squamous cell carcinoma (ESCC) remains one of the deadly cancer types. Comprehensively dissecting the molecular characterization and the heterogeneity of ESCC paves the way for developing more promising therapeutics.

METHODS

Expression profiles of multiple ESCC datasets were integrated. ATAC-seq and RNA-seq were combined to reveal the chromatin accessibility features. A prognosis-related subtype classifier (PrSC) was constructed, and its association with the tumor microenvironment (TME) and immunotherapy was assessed. The key gene signature was validated in clinical samples. Based on the TME heterogeneity of ESCC patients, potential subtype-specific therapeutic agents were screened.

FINDINGS

The common differentially expressed genes (cDEGs) in ESCC were identified. Up-regulated genes (HEATR1, TIMELESS, DTL, GINS1, RUVBL1, and ECT2) were found highly important in ESCC cell survival. The expression alterations of PRIM2, HPGD, NELL2, and TFAP2B were associated with chromatin accessibility changes. PrSC was a robust scoring tool that was not only associated with the prognosis of ESCC patients, but also could reflect the TME heterogeneity. TNS1^high fibroblasts were associated with immune exclusion. TG-101348 and Vinorelbine were identified as potential subtype-specific therapeutic agents. Besides, the application of PrSC into two immunotherapy cohorts indicated its potential value in assessing treatment response to immunotherapy.

INTERPRETATION

Our study depicted the multi-dimensional characterization of ESCC, established a robust scoring tool for the prognosis assessment, highlighted the role of TNS1^high fibroblasts in TME, and identified potential drugs for clinical use.

FUNDING

A full list of funding bodies that contributed to this study can be found in the Acknowledgements section.

Multiomic Analysis of Cereblon Expression and Its Prognostic Value in Kidney Renal Clear Cell Carcinoma, Lung Adenocarcinoma, and Skin Cutaneous Melanoma

Cereblon (CRBN) is a component of the E3 ubiquitin ligase complex that plays crucial roles in various cellular processes. However, no systematic studies on the expression and functions of CRBN in solid tumors have been conducted to date. Here, we analyzed CRBN expression and its clinical value using several bioinformatic databases. CRBN mRNA expression was downregulated in various cancer types compared to normal cells. Survival analysis demonstrated that overall survival was significantly positively correlated with CRBN expression in some cancer types including lung adenocarcinoma (LUAD), kidney renal clear cell carcinoma (KIRC), and skin cutaneous melanoma (SKCM). CRBN expression was downregulated regardless of clinicopathological characteristics in LUAD and KIRC. Analysis of genes that are commonly correlated with CRBN expression among KIRC, LUAD, and SKCM samples elucidated the potential CRBN-associated mechanisms of cancer progression. Overall, this study revealed the prognostic value of CRBN and its potential associated mechanisms, which may facilitate the development of anti-cancer therapeutic agents.

A prognostic nomogram based on LASSO Cox regression in patients with alpha-fetoprotein-negative hepatocellular carcinoma following non-surgical therapy

Background

Alpha-fetoprotein-negative hepatocellular carcinoma (AFP-NHCC) (< 8.78 ng/mL) have special clinicopathologic characteristics and prognosis. The aim of this study was to apply a new method to establish and validate a new model for predicting the prognosis of patients with AFP-NHCC.

Methods

A total of 410 AFP-negative patients with clinical diagnosed with HCC following non-surgical therapy as a primary cohort; 148 patients with AFP-NHCC following non-surgical therapy as an independent validation cohort. In primary cohort, independent factors for overall survival (OS) by LASSO Cox regression were all contained into the nomogram1; by Forward Stepwise Cox regression were all contained into the nomogram2. Nomograms performance and discriminative power were assessed with concordance index (C-index) values, area under curve (AUC), Calibration curve and decision curve analyses (DCA). The results were validated in the validation cohort.

Results

The C-index of nomogram1was 0.708 (95%CI: 0.673–0.743), which was superior to nomogram2 (0.706) and traditional modes (0.606–0.629). The AUC of nomogram1 was 0.736 (95%CI: 0.690–0.778). In the validation cohort, the nomogram1 still gave good discrimination (C-index: 0.752, 95%CI: 0.691–0.813; AUC: 0.784, 95%CI: 0.709–0.847). The calibration curve for probability of OS showed good homogeneity between prediction by nomogram1 and actual observation. DCA demonstrated that nomogram1 was clinically useful. Moreover, patients were divided into three distinct risk groups for OS by the nomogram1: low-risk group, middle-risk group and high-risk group, respectively.

Conclusions

Novel nomogram based on LASSO Cox regression presents more accurate and useful prognostic prediction for patients with AFP-NHCC following non-surgical therapy. This model could help patients with AFP-NHCC following non-surgical therapy facilitate a personalized prognostic evaluation.

Construction and Validation of a Prognostic Gene-Based Model for Overall Survival Prediction in Hepatocellular Carcinoma Using an Integrated Statistical and Bioinformatic Approach

Hepatocellular carcinoma (HCC) is one of the most common lethal cancers worldwide and is often related to late diagnosis and poor survival outcome. More evidence is demonstrating that gene-based prognostic models can be used to predict high-risk HCC patients. Therefore, our study aimed to construct a novel prognostic model for predicting the prognosis of HCC patients. We used multivariate Cox regression model with three hybrid penalties approach including least absolute shrinkage and selection operator (Lasso), adaptive lasso and elastic net algorithms for informative prognostic-related genes selection. Then, the best subset regression was used to identify the best prognostic gene signature. The prognostic gene-based risk score was constructed using the Cox coefficient of the prognostic gene signature. The model was evaluated by Kaplan-Meier (KM) and receiver operating characteristic curve (ROC) analyses. A novel four-gene signature associated with prognosis was identified and the risk score was constructed based on the four-gene signature. The risk score efficiently distinguished the patients into a high-risk group with poor prognosis. The time-dependent ROC analysis revealed that the risk model had a good performance with an area under the curve (AUC) of 0.780, 0.732, 0.733 in 1-, 2- and 3-year prognosis prediction in The Cancer Genome Atlas (TCGA) dataset. Moreover, the risk score revealed a high diagnostic performance to classify HCC from normal samples. The prognosis and diagnosis prediction performances of risk scores were verified in external validation datasets. Functional enrichment analysis of the four-gene signature and its co-expressed genes involved in the metabolic and cell cycle pathways was constructed. Overall, we developed a novel-gene-based prognostic model to predict high-risk HCC patients and we hope that our findings can provide promising insight to explore the role of the four-gene signature in HCC patients and aid risk classification.

Identification of theranostic factors for patients developing metastasis after surgery for early-stage lung adenocarcinoma

Rationale: Lung adenocarcinoma (LUAD) is an aggressive disease with high propensity of metastasis. Among patients with early-stage disease, more than 30% of them may relapse or develop metastasis. There is an unmet medical need to stratify patients with early-stage LUAD according to their risk of relapse/metastasis to guide preventive or therapeutic approaches. In this study, we identified 4 genes that can serve both therapeutic and diagnostic (theranostic) purposes.

Methods: Three independent datasets (GEO, TCGA, and KMPlotter) were used to evaluate gene expression profile of patients with LUAD by unbiased screening approach. Upon significant genes uncovered, functional enrichment analysis was carried out. The predictive power of their expression on patient prognosis were evaluated. Once confirmed their theranostic roles by integrated bioinformatics, we further conducted in vitro and in vivo validation.

Results: We found that four genes (ADAM9, MTHFD2, RRM2, and SLC2A1) were associated with poor patient outcomes with an increased hazard ratio in LUAD. Knockdown of them, both separately and simultaneously, suppressed lung cancer cell proliferation and migration ability in vitro and prolonged survival time in metastatic tumor mouse models. Moreover, these four biomarkers were found to be overexpressed in tumor tissues from LUAD patients, and the total immunohistochemical staining scores correlated with poor prognosis.

Conclusions: These results suggest that these four identified genes could be theranostic biomarkers for stratifying high-risk patients who develop relapse/metastasis in early-stage LUAD. Developing therapeutic approaches for the four biomarkers may benefit early-stage LUAD patients after surgery.

Identification of an immune prognostic 11-gene signature for lung adenocarcinoma

Background

The immunological tumour microenvironment (TME) has occupied a very important position in the beginning and progression of non-small cell lung cancer (NSCLC). Prognosis of lung adenocarcinoma (LUAD) remains poor for the local progression and widely metastases at the time of clinical diagnosis. Our objective is to identify a potential signature model to improve prognosis of LUAD.

Methods

With the aim to identify a novel immune prognostic signature associated with overall survival (OS), we analysed LUADs extracted from The Cancer Genome Atlas (TCGA). Immune scores and stromal scores of TCGA-LUAD were downloaded from Estimation of STromal and Immune cells in MAlignant Tumour tissues Expression using data (ESTIMATE). LASSO COX regression was applied to build the prediction model. Then, the prognostic gene signature was validated in the GSE68465 dataset.

Results

The data from TCGA datasets showed patients in stage I and stage II had higher stromal scores than patients in stage IV (P < 0.05), and for immune score patients in stage I were higher than patients in stage III and stage IV (P < 0.05). The improved overall survivals were observed in high stromal score and immune score groups. Patients in the high-risk group exhibited the inferior OS (P = 2.501e − 05). By validating the 397 LUAD patients from GSE68465, we observed a better OS in the low-risk group compared to the high-risk group, which is consistent with the results from the TCGA cohort. Nomogram results showed that practical and predicted survival coincided very well, especially for 3-year survival.

Conclusion

We obtained an 11 immune score related gene signature model as an independent element to effectively classify LUADs into different risk groups, which might provide a support for precision treatments. Moreover, immune score may play a potential valuable sole for estimating OS in LUADs.

Development of Machine Learning Models for Predicting Postoperative Delayed Remission in Patients With Cushing's Disease

CONTEXT

Postoperative hypercortisolemia mandates further therapy in patients with Cushing's disease (CD). Delayed remission (DR) is defined as not achieving postoperative immediate remission (IR), but having spontaneous remission during long-term follow-up.

OBJECTIVE

We aimed to develop and validate machine learning (ML) models for predicting DR in non-IR patients with CD.

METHODS

We enrolled 201 CD patients, and randomly divided them into training and test datasets. We then used the recursive feature elimination (RFE) algorithm to select features and applied 5 ML algorithms to construct DR prediction models. We used permutation importance and local interpretable model-agnostic explanation (LIME) algorithms to determine the importance of the selected features and interpret the ML models.

RESULTS

Eighty-eight (43.8%) of the 201 CD patients met the criteria for DR. Overall, patients who were younger, had a low body mass index, a Knosp grade of III-IV, and a tumor not found by pathological examination tended to achieve a lower rate of DR. After RFE feature selection, the Adaboost model, which comprised 18 features, had the greatest discriminatory ability, and its predictive ability was significantly better than using Knosp grading and postoperative immediate morning serum cortisol (PoC). The results obtained from permutation importance and LIME algorithms showed that preoperative 24-hour urine free cortisol, PoC, and age were the most important features, and showed the reliability and clinical practicability of the Adaboost model in DC prediction.

CONCLUSIONS

Machine learning-based models could serve as an effective noninvasive approach to predicting DR, and could aid in determining individual treatment and follow-up strategies for CD patients.

The definition of remission and recurrence of Cushing's disease

Accurate classification of postsurgical remission, and early recognition of recurrence are crucial to timely treat and prevent excess mortality in Cushing's Disease, yet the criteria used to define remission are variable and there is no consensus to define recurrence. Remission is defined as postsurgical hypocortisolemia, but delayed remission may occur. Recurrence is the return of clinical manifestations with biochemical evidence of hypercortisolism. The proper combination of tests and their timing are controversial. Reliable predicting tools may lead to earlier diagnosis upon recurrence. Many factors have been studied independently for prediction with variable performance. Novel artificial intelligence approaches seek to integrate these variables into risk calculators and machine-learning algorithms with an acceptable short-term predictive performance but lack longer-term accuracy. Prospective studies using these approaches are needed. This review summarizes the evidence behind the definitions of remission and recurrence and provide an overview of the available tools to predict and/or diagnose them.

Classify multicategory outcome in patients with lung adenocarcinoma using clinical, transcriptomic and clinico-transcriptomic data: machine learning versus multinomial models

Classification of multicategory survival-outcome is important for precision oncology. Machine learning (ML) algorithms have been used to accurately classify multi-category survival-outcome of some cancer-types, but not yet that of lung adenocarcinoma. Therefore, we compared the performances of 3 ML models (random forests, support vector machine [SVM], multilayer perceptron) and multinomial logistic regression (Mlogit) models for classifying 4-category survival-outcome of lung adenocarcinoma using the TCGA. Mlogit model overall performed similar to SVM and multilayer perceptron models (micro-average area under curve=0.82), while random forests model was inferior. Surprisingly, transcriptomic data alone and clinico-transcriptomic data appeared sufficient to accurately classify the 4-category survival-outcome in these patients, but no models using clinical data alone performed well. Notably, NDUFS5, P2RY2, PRPF18, CCL24, ZNF813, MYL6, FLJ41941, POU5F1B, and SUV420H1 were the top-ranked genes that were associated with alive without disease and inversely linked to other outcomes. Similarly, BDKRB2, TERC, DNAJA3, MRPL15, SLC16A13, CRHBP and ACSBG2 were associated with alive with progression and GAL3ST3, AD2, RAB41, HDC, and PLEKHG1 associated with dead with disease, respectively, while also inversely linked other outcomes. These cross-linked genes may be used for risk-stratification and future treatment development.

A Recurrence-Specific Gene-Based Prognosis Prediction Model for Lung Adenocarcinoma through Machine Learning Algorithm

Background

After curative surgical resection, about 30-75% lung adenocarcinoma (LUAD) patients suffer from recurrence with dismal survival outcomes. Identification of patients with high risk of recurrence to impose intense therapy is urgently needed.

Materials and Methods

Gene expression data of LUAD were obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. Differentially expressed genes (DEGs) were calculated by comparing the recurrent and primary tissues. Prognostic genes associated with the recurrence-free survival (RFS) of LUAD patients were identified using univariate analysis. LASSO Cox regression and multivariate Cox analysis were applied to extract key genes and establish the prediction model.

Results

We detected 37 DEGs between primary and recurrent LUAD tumors. Using univariate analysis, 31 DEGs were found to be significantly associated with RFS. We established the RFS prediction model including thirteen genes using the LASSO Cox regression. In the training cohort, we classified patients into high- and low-risk groups and found that patients in the high-risk group suffered from worse RFS compared to those in the low-risk group (P < 0.01). Concordant results were confirmed in the internal and external validation cohort. The efficiency of the prediction model was also confirmed under different clinical subgroups. The high-risk group was significantly identified as the risk factor of recurrence in LUAD by the multivariate Cox analysis (HR = 13.37, P = 0.01). Compared to clinicopathological features, our prediction model possessed higher accuracy to identify patients with high risk of recurrence (AUC = 96.3%). Finally, we found that the G2M checkpoint pathway was enriched both in recurrent tumors and primary tumors of high-risk patients.

Conclusions

Our recurrence-specific gene-based prognostic prediction model provides extra information about the risk of recurrence in LUAD, which is conducive for clinicians to conduct individualized therapy in clinic.

Molecular characterization, biological function, tumor microenvironment association and clinical significance of m6A regulators in lung adenocarcinoma

N6-methyladenosine (m6A) modification can regulate a variety of biological processes. However, the implications of m6A modification in lung adenocarcinoma (LUAD) remain largely unknown. Here, we systematically evaluated the m6A modification features in more than 2400 LUAD samples by analyzing the multi-omics features of 23 m6A regulators. We depicted the genetic variation features of m6A regulators, and found mutations of FTO and YTHDF3 were linked to worse overall survival. Many m6A regulators were aberrantly expressed in tumors, among which FTO, IGF2BP3, YTHDF1 and RBM15 showed consistent alteration features across 11 independent cohorts. Besides, the regulator-pathway interaction network demonstrated that m6A modification was associated with various biological pathways, including immune-related pathways. The correlation between m6A regulators and tumor microenvironment was also assessed. We found that LRPPRC was negatively correlated with most tumor-infiltrating immune cells. On the other hand, we established a scoring tool named m6Sig, which was positively correlated with PD-L1 expression and could reflect both the tumor microenvironment characterization and prognosis of LUAD patients. Comparison of CNV between high and low m6Sig groups revealed differences on chromosome 7. Application of m6Sig on an anti-PD-L1 immunotherapy cohort confirmed that the high m6Sig group demonstrated therapeutic advantages and clinical benefits. Our study indicated that m6A modification is involved in many aspects of LUAD and contributes to tumor microenvironment formation. A better understanding of m6A modification will provide more insights into the molecular mechanisms of LUAD and facilitate developing more effective personalized treatment strategies. A web application was built along with this study (http://www.bioinfo-zs.com/luadexpress/).

Interactive Verification Analysis of Multiple Sequencing Data for Identifying Potential Biomarker of Lung Adenocarcinoma

Background

Lung adenocarcinoma (LUAD) comprises around 40% of all lung cancers, and in about 70% of patients, it has spread locally or systemically when first detected leading to a worse prognosis.

Methods

We filtered out differentially expressed genes (DEGs) based on the RNA sequencing data in the Gene Expression Omnibus database and verified and deeply analyzed screened DEGs using a combined bioinformatics approach.

Results

Expressions of 11,143 genes in 694 nontumor lung tissues and LUAD cases from 8 independent laboratories were analyzed; 188 mRNAs were identified as differentially expressed genes (DEGs). A PPI network constructed with 188 DEGs screened out 8 hub DEGs (CDH5, PECAM1, VWF, CLDN5, COL1A1, MMP9, SPP1, and IL6) which highly interconnected with other nodes. The expression levels of 8 hub genes in LUAD and control were assessed in the Oncomine database, and the results were consistent. The survival curves of 8 hub genes showed that their expressions are significantly related to the prognosis of lung cancer and LUAD patients except for IL6. Since the expression of IL6 is nonspecific and highly sensitive, we choose the other 7 hub genes we had verified to do the next analysis. Mutual exclusivity or cooccurrence analysis of 7 hub genes identified a tendency towards cooccurrence between CDH5, PECAM1, and VWF in LUAD. The coexpression profiles of CDH5 in LUAD were identified, and we found that PECAM1 and VWF coexpressed with CDH5. Immunohistochemistry and RT-PCR analysis showed that higher levels of CDH5, PECAM1, and VWF were expressed in normal lung tissues but a low or undetectable level was found in LUAD tissues.

Conclusions

Taken together, we speculate that CDH5, PECAM1, and VWF played an important role in LUAD.

Development of a novel prognostic score combining clinicopathologic variables, gene expression, and mutation profiles for lung adenocarcinoma

BACKGROUND

Integrating phenotypic and genotypic information to improve prognostic prediction is under active investigation for lung adenocarcinoma (LUAD). In this study, we developed a new prognostic model for event-free survival (EFS) and recurrence-free survival (RFS) based on the combination of clinicopathologic variables, gene expression, and mutation data.

METHODS

We enrolled a total of 408 patients from the Cancer Genome Atlas Lung Adenocarcinoma (TCGA-LUAD) project for the study. We pre-selected gene expression or mutation features and constructed 14 different input feature sets for predictive model development. We assessed model performance with multiple evaluation metrics including the distribution of C-index on testing dataset, risk score significance, and time-dependent AUC under competing risks scenario. We stratified patients into higher- and lower-risk subgroups by the final risk score and further investigated underlying immune phenotyping variations associated with the differential risk.

RESULTS

The model integrating all three types of data achieved the best prediction performance. The resultant risk score provided a higher-resolution risk stratification than other models within pathologically defined subgroups. The score could account for extra EFS-related variations that were not captured by clinicopathologic scores. Being validated for RFS prediction under a competing risks modeling framework, the score achieved a significantly higher time-dependent AUC as compared to that of the conventional clinicopathologic variables-based model (0.772 vs. 0.646, p value < 0.001). The higher-risk patients were characterized with transcriptional aberrations of multiple immune-related genes, and a significant depletion of mast cells and natural killer cells.

CONCLUSIONS

We developed a novel prognostic risk score with improved prediction accuracy, using clinicopathologic variables, gene expression and mutation profiles as input, for LUAD. Such score was a significant predictor of both EFS and RFS.

TRIAL REGISTRATION

This study was based on public open data from TCGA and hence the study objects were retrospectively registered.

From tobacco smoking to cancer mutational signature: a mediation analysis strategy to explore the role of epigenetic changes

Background

Tobacco smoking is associated with a unique mutational signature in the human cancer genome. It is unclear whether tobacco smoking-altered DNA methylations and gene expressions affect smoking-related mutational signature.

Methods

We systematically analyzed the smoking-related DNA methylation sites reported from five previous casecontrol studies in peripheral blood cells to identify possible target genes. Using the mediation analysis approach, we evaluated whether the association of tobacco smoking with mutational signature is mediated through altered DNA methylation and expression of these target genes in lung adenocarcinoma tumor tissues.

Results

Based on data obtained from 21,108 blood samples, we identified 374 smoking-related DNA methylation sites, annotated to 248 target genes. Using data from DNA methylations, gene expressions and smoking-related mutational signature generated from ~ 7700 tumor tissue samples across 26 cancer types from The Cancer Genome Atlas (TCGA), we found 11 of the 248 target genes whose expressions were associated with smoking-related mutational signature at a Bonferroni-correction P < 0.001. This included four for head and neck cancer, and seven for lung adenocarcinoma. In lung adenocarcinoma, our results showed that smoking increased the expression of three genes, AHRR, GPR15, and HDGF, and decreased the expression of two genes, CAPN8, and RPS6KA1, which were consequently associated with increased smoking-related mutational signature. Additional evidence showed that the elevated expression of AHRR and GPR15 were associated with smoking-altered hypomethylations at cg14817490 and cg19859270, respectively, in lung adenocarcinoma tumor tissues. Lastly, we showed that decreased expression of RPS6KA1, were associated with poor survival of lung cancer patients.

Conclusions

Our findings provide novel insight into the contributions of tobacco smoking to carcinogenesis through the underlying mechanisms of the elevated mutational signature by altered DNA methylations and gene expressions.

Prediction for cardiovascular diseases based on laboratory data: An analysis of random forest model

Background

To establish a prediction model for cardiovascular diseases (CVD) in the general population based on random forests.

Methods

A retrospective study involving 498 subjects was conducted in Xi'an Medical University between 2011 and 2018. The random forest algorithm was used to screen out the variables that greatly affected the CVD prediction and to establish a prediction model. The important variables were included in the multifactorial logistic regression analysis. The area under the curve (AUC) was compared between logistic regression model and random forest model.

Results

The random forest model revealed the variables, including the age, body mass index (BMI), fasting blood glucose (FBG), diastolic blood pressure (DBP), triglyceride (TG), systolic blood pressure (SBP), total cholesterol (TC), waist circumference, and high‐density lipoprotein‐cholesterol (HDL‐C), were more significant for CVD prediction; the AUC was 0.802 in CVD prediction. Multifactorial logistic regression analysis indicated that the risk factors for CVD included the age [odds ratio (OR): 1.14, 95% confidence intervals (CI): 1.10‐1.17, P < .001], BMI (OR: 1.13, 95% CI: 1.06‐1.20, P < .001), TG (OR: 1.11, 95% CI: 1.02‐1.22, P = .023), and DBP (OR: 1.04, 95% CI: 1.02‐1.06, P = .001); the AUC was 0.843 in CVD prediction. The established logistic regression prediction model was Logit P = Log[P/(1 − P)] = −11.47 + 0.13 × age + 0.12 × BMI + 0.11 × TG + 0.04 × DBP; P = 1/[1 + exp(−Logit P)]. People were prone to develop CVD at the time of P > .51.

Conclusions

A prediction model for CVD is developed in the general population based on random forests, which provides a simple tool for the early prediction of CVD.

Comparative Analysis of Age- and Gender-Associated Microbiome in Lung Adenocarcinoma and Lung Squamous Cell Carcinoma

The intra-tumor microbiota has been increasingly implicated in cancer pathogenesis. In this study, we aimed to examine the microbiome in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) and determine its compositional differences with relation to age and gender. After grouping 497 LUAD and 433 LUSC patients by age and gender and removing potential contaminants, we identified differentially abundant microbes in each patient cohort vs. adjacent normal samples. We then correlated dysregulated microbes with patient survival rates, immune infiltration, immune and cancer pathways, and genomic alterations. We found that most age and gender cohorts in both LUAD and LUSC contained unique, significantly dysregulated microbes. For example, LUAD-associated Escherichia coli str. K-12 substr. W3110 was dysregulated in older female and male patients and correlated with both patient survival and genomic alterations. For LUSC, the most prominent bacterial species that we identified was Pseudomonas putida str. KT2440, which was uniquely associated with young LUSC male patients and immune infiltration. In conclusion, we found differentially abundant microbes implicated with age and gender that are also associated with genomic alterations and immune dysregulations. Further investigation should be conducted to determine the relationship between gender and age-associated microbes and the pathogenesis of lung cancer.

Development and Validation of a 12-Gene Immune Relevant Prognostic Signature for Lung Adenocarcinoma Through Machine Learning Strategies

Background: Although immunotherapy with checkpoint inhibitors is changing the face of lung adenocarcinoma (LUAD) treatments, only limited patients could benefit from it. Therefore, we aimed to develop an immune-relevant-gene-based signature to predict LUAD patients' prognosis and to characterize their tumor microenvironment thus guiding therapeutic strategy.

Methods and Materials: Gene expression data of LUAD patients from Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) were systematically analyzed. We performed Cox regression and random survival forest algorithm to identify immune-relevant genes with potential prognostic value. A risk score formula was then established by integrating these selected genes and patients were classified into high- and low-risk score group. Differentially expressed genes, infiltration level of immune cells, and several immune-associated molecules were further compared across the two groups.

Results: Nine hundred and fifty-four LUAD patients were enrolled in this study. After implementing the 2-steps machine learning screening methods, 12 immune-relevant genes were finally selected into the risk-score formula and the patients in high-risk group had significantly worse overall survival (HR = 10.6, 95%CI = 3.21–34.95, P < 0.001). We also found the distinct immune infiltration patterns in the two groups that several immune cells like cytotoxic cells and immune checkpoint molecules were significantly enriched and upregulated in patients from the high-risk group. These findings were further validated in two independent LUAD cohorts.

Conclusion: Our risk score formula could serve as a powerful and accurate tool for predicting survival of LUAD patients and may facilitate clinicians to choose the optimal therapeutic regimen more precisely.

Identification and validation of tumor environment phenotypes in lung adenocarcinoma by integrative genome-scale analysis

PURPOSE

To comprehensively elucidate the landscape of the tumor environment (TME) of lung adenocarcinoma (LUAD), which has a profound impact on prognosis and response to immunotherapy.

METHODS AND MATERIALS

Using a large dataset of LUAD patients from The Cancer Genome Atlas, Gene Expression Omnibus database (GEO), and our institution (n = 1411), we estimated the infiltration pattern of 24 immune cell populations in each sample and systematically correlated the TME phenotypes with genomic traits and clinicopathologic characteristics.

RESULTS

The LUAD microenvironment was classified into two distinct TME clusters (A and B), and a random forest classifier model was constructed. TMEcluster A was characterized by sparse distribution of immune cell infiltration, relatively low levels of immunomodulators and slightly higher mutation load. By contrast, enrichment of both cytotoxic T cells and immunosuppressor cells was observed in TMEcluster B. Moreover, several immune-related cytokines or markers including IFN-γ, TNF-β, and several immune checkpoint molecules such as PD-L1 were also upregulated in TMEcluster B. Multivariable Cox analysis revealed that the TMEcluster was an independent prognostic factor (TMEcluster B vs. A, hazard ratio = 0.68, 95% confidence interval = 0.50-0.91, p = 0.010). These findings were all externally validated in the data from the GEO database and our institution.

CONCLUSIONS

Our findings describe a comprehensive landscape of LUAD immune infiltration pattern and integrate several previously proposed biomarkers associated with distinct immunophenotypes, thus shedding light on how tumors interact with immune microenvironment. Our results may guide a more precise immune therapeutic strategy for LUAD patients.

Identification of Three Differentially Expressed miRNAs as Potential Biomarkers for Lung Adenocarcinoma Prognosis

OBJECTIVE

The aim of this study areto screen MicroRNAs (miRNAs) related to the prognosis of lung adenocarcinoma (LUAD) and to explore the possible molecular mechanisms.

METHODS

The data for a total of 535 patients with LUAD data were downloaded from The Cancer Genome Atlas (TCGA) database. The miRNAs for LUAD prognosis were screened by both Cox risk proportional regression model and Last Absolute Shrinkage and Selection Operator (LASSO) regression model. The performances of the models were verified by time-dependent Receiver Operating Characteristic (ROC) curve. The possible biological processes linked to the miRNAs' target genes were analyzed by Gene Ontology (GO), Kyoto gene and genome encyclopedia (KEGG).

RESULTS

Among 127 differentially expressed miRNAs identified from the screening analysis, there are 111 up-regulated and 16 down-regulated miRNAs. Three of them, hsa-miR-1293, hsa-miR-490 and hsa-miR- 5571, were also significantly associated with the survival of the LUAD patients. The targets of the three miRNAs are significantly enriched in systemic lupus erythematosus pathways.

CONCLUSION

Hsa-miR-1293, hsa-miR-490 and hsa-miR-5571 can be potentially used as novel biomarkers for the prognosis prediction of LUAD.

Development and Interpretation of Multiple Machine Learning Models for Predicting Postoperative Delayed Remission of Acromegaly Patients During Long-Term Follow-Up

Background: Some patients with acromegaly do not reach the remission standard in the short term after surgery but achieve remission without additional postoperative treatment during long-term follow-up; this phenomenon is defined as postoperative delayed remission (DR). DR may complicate the interpretation of surgical outcomes in patients with acromegaly and interfere with decision-making regarding postoperative adjuvant therapy. Objective: We aimed to develop and validate machine learning (ML) models for predicting DR in acromegaly patients who have not achieved remission within 6 months of surgery. Methods: We enrolled 306 acromegaly patients and randomly divided them into training and test datasets. We used the recursive feature elimination (RFE) algorithm to select features and applied six ML algorithms to construct DR prediction models. The performance of these ML models was validated using receiver operating characteristics analysis. We used permutation importance, SHapley Additive exPlanations (SHAP), and local interpretable model-agnostic explanation (LIME) algorithms to determine the importance of the selected features and interpret the ML models. Results: Fifty-five (17.97%) acromegaly patients met the criteria for DR, and five features (post-1w rGH, post-1w nGH, post-6m rGH, post-6m IGF-1, and post-6m nGH) were significantly associated with DR in both the training and the test datasets. After the RFE feature selection, the XGboost model, which comprised the 15 important features, had the greatest discriminatory ability (area under the curve = 0.8349, sensitivity = 0.8889, Youden's index = 0.6842). The XGboost model showed good discrimination ability and provided significantly better estimates of DR of patients with acromegaly compared with using only the Knosp grade. The results obtained from permutation importance, SHAP, and LIME algorithms showed that post-6m IGF-1 is the most important feature in XGboost algorithm prediction and showed the reliability and the clinical practicability of the XGboost model in DR prediction. Conclusions: ML-based models can serve as an effective non-invasive approach to predicting DR and could aid in determining individual treatment and follow-up strategies for acromegaly patients who have not achieved remission within 6 months of surgery.