Artificial intelligence (AI) aims to mimic human cognitive functions. It is bringing a paradigm shift to healthcare, powered by increasing availability of healthcare data and rapid progress of analytics techniques. We survey the current status of AI applications in healthcare and discuss its future. AI can be applied to various types of healthcare data (structured and unstructured). Popular AI techniques include machine learning methods for structured data, such as the classical support vector machine and neural network, and the modern deep learning, as well as natural language processing for unstructured data. Major disease areas that use AI tools include cancer, neurology and cardiology. We then review in more details the AI applications in stroke, in the three major areas of early detection and diagnosis, treatment, as well as outcome prediction and prognosis evaluation. We conclude with discussion about pioneer AI systems, such as IBM Watson, and hurdles for real-life deployment of AI.

https://svn.bmj.com/content/svnbmj/2/4/230.full.pdf

Artificial intelligence in healthcare: past, present and future

Phosphatidylinositol-3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR) signaling is one of the most important intracellular pathways, which can be considered as a master regulator for cancer. Enormous efforts have been dedicated to the development of drugs targeting PI3K signaling, many of which are currently employed in clinical trials evaluation, and it is becoming increasingly clear that PI3K inhibitors are effective in inhibiting tumor progression. PI3K inhibitors are subdivided into dual PI3K/mTOR inhibitors, pan-PI3K inhibitors and isoform-specific inhibitors. In this review, we performed a critical review to summarize the role of the PI3K pathway in tumor development, recent PI3K inhibitors development based on clinical trials, and the mechanisms of resistance to PI3K inhibition.

/pdf/targeting-pi3k-in-cancer-mechanisms-and-advances-in-clinical-4ituoao2ez.pdf

Targeting PI3K in cancer: mechanisms and advances in clinical trials

Noncoding RNAs (ncRNAs) constitute the majority of the human transcribed genome. This largest class of RNA transcripts plays diverse roles in a multitude of cellular processes, and has been implicated in many pathological conditions, especially cancer. The different subclasses of ncRNAs include microRNAs, a class of short ncRNAs; and a variety of long ncRNAs (lncRNAs), such as lincRNAs, antisense RNAs, pseudogenes, and circular RNAs. Many studies have demonstrated the involvement of these ncRNAs in competitive regulatory interactions, known as competing endogenous RNA (ceRNA) networks, whereby lncRNAs can act as microRNA decoys to modulate gene expression. These interactions are often interconnected, thus aberrant expression of any network component could derail the complex regulatory circuitry, culminating in cancer development and progression. Recent integrative analyses have provided evidence that new computational platforms and experimental approaches can be harnessed together to distinguish key ceRNA interactions in specific cancers, which could facilitate the identification of robust biomarkers and therapeutic targets, and hence, more effective cancer therapies and better patient outcome and survival.

Noncoding RNA:RNA Regulatory Networks in Cancer.

The incidence and mortality rate of cancer has been quickly increasing in the past decades. At present, cancer has become the leading cause of death worldwide. Most of the cancers cannot be effectively diagnosed at the early stage. Although there are multiple therapeutic treatments, including surgery, radiotherapy, chemotherapy, and targeted drugs, their effectiveness is still limited. The overall survival rate of malignant cancers is still low. It is necessary to further study the mechanisms for malignant cancers, and explore new biomarkers and targets that are more sensitive and effective for early diagnosis, treatment, and prognosis of cancers than traditional biomarkers and methods. Long non-coding RNAs (lncRNAs) are a class of RNA transcripts with a length greater than 200 nucleotides. Generally, lncRNAs are not capable of encoding proteins or peptides. LncRNAs exert diverse biological functions by regulating gene expressions and functions at transcriptional, translational, and post-translational levels. In the past decade, it has been demonstrated that the dysregulated lncRNA profile is widely involved in the pathogenesis of many diseases, including cancer, metabolic disorders, and cardiovascular diseases. In particular, lncRNAs have been revealed to play an important role in tumor growth and metastasis. Many lncRNAs have been shown to be potential biomarkers and targets for the diagnosis and treatment of cancers. This review aims to briefly discuss the latest findings regarding the roles and mechanisms of some important lncRNAs in the pathogenesis of certain malignant cancers, including lung, breast, liver, and colorectal cancers, as well as hematological malignancies and neuroblastoma.

https://www.mdpi.com/2073-4409/8/9/1015/pdf

Long Non-Coding RNA in the Pathogenesis of Cancers.

Numerous studies have highlighted that long non-coding RNAs (lncRNAs) can bind to microRNA (miRNA) sites as competing endogenous RNAs (ceRNAs), thereby affecting and regulating the expression of mRNAs and target genes. These lncRNA-associated ceRNAs have been theorized to play a significant role in cancer initiation and progression. However, the roles and functions of the lncRNA-miRNA-mRNA ceRNA network in squamous cell carcinoma of the tongue (SCCT) are still unclear. The miRNA, mRNA and lncRNA expression profiles from 138 patients with SCCT were downloaded from The Cancer Genome Atlas database. We identified the differential expression of miRNAs, mRNAs, and lncRNAs using the limma package of R software. We used the clusterProfiler package for GO and KEGG pathway annotations. The survival package was used to estimate survival analysis according to the Kaplan-Meier curve. Finally, the GDCRNATools package was used to construct the lncRNA-miRNA-mRNA ceRNA network. In total, 1943 SCCT-specific mRNAs, 107 lncRNAs and 100 miRNAs were explored. Ten mRNAs (CSRP2, CKS2, ADGRG6, MB21D1, GMNN, RIPOR3, RAD51, PCLAF, ORC1, NAGS), 9 lncRNAs (LINC02560, HOXC13 − AS, FOXD2 − AS1, AC105277.1, AC099850.3, STARD4 − AS1, SLC16A1 − AS1, MIR503HG, MIR100HG) and 8 miRNAs (miR − 654, miR − 503, miR − 450a, miR − 379, miR − 369, miR − 190a, miR − 101, and let−7c) were found to be significantly associated with overall survival (log-rank p < 0.05). Based on the analysis of the lncRNA-miRNA-mRNA ceRNA network, one differentially expressed (DE) lncRNA, five DEmiRNAs, and three DEmRNAs were demonstrated to be related to the pathogenesis of SCCT. In this study, we described the gene regulation by the lncRNA-miRNA-mRNA ceRNA network in the progression of SCCT. We propose a new lncRNA-associated ceRNA that could help in the diagnosis and treatment of SCCT.

/pdf/integrated-analysis-of-lncrna-mirna-mrna-cerna-network-in-3ifehqwz1x.pdf

Integrated analysis of lncRNA-miRNA-mRNA ceRNA network in squamous cell carcinoma of tongue

Accumulating evidence has highlighted the important roles of long non-coding RNAs (lncRNAs) acting as competing endogenous RNAs (ceRNAs) in tumor biology. However, the roles of cancer specific lncRNAs in lncRNA-related ceRNA network of lung adenocarcinoma (LUAD) are still unclear. In the present study, the 465 RNA sequencing profiles in LUAD patients were obtained from the cancer genome atlas (TCGA) database, which provides large sample RNA sequencing data free of charge, and 41 cancer specific lncRNAs, 25 miRNAs and 1053 mRNAs (fold change >2, p<0.05) were identified. Then, the lncRNA-miRNA-mRNA ceRNA network of LUAD was constructed with 29 key lncRNAs, 24 miRNAs and 72 mRNAs. Subsequently, we selected these 29 key lncRNAs to analyze their correlation with clinical features, and 21 of them were aberrantly expressed with tumor pathological stage, TNM staging system, lymph node metastasis and patient outcome assessment, respectively. Furthermore, there were 5 lncRNAs (BCRP3, LINC00472, CHIAP2, BMS1P20 and UNQ6494) positively correlated with overall survival (OS, log-rank p<0.05). Finally, 7 cancer specific lncRNAs were randomly selected to verify the expression in 53 newly diagnosed LUAD patients using qRT-PCR. The expression results between TCGA and qRT-PCR were 100% in agreement. The correlation between AFAP1-AS1 and LINC00472 and clinical features were also confirmed. Thus, our results showed the lncRNA expression profiles and we constructed an lncRNA-miRNA-mRNA ceRNA network in LUAD. The present study provides novel insight for better understanding of lncRNA-related ceRNA network in LUAD and facilitates the identification of potential biomarkers for diagnosis and prognosis.

/pdf/integrated-analysis-of-long-non-coding-rna-associated-cerna-39wt0gd61a.pdf

Integrated analysis of long non-coding RNA‑associated ceRNA network reveals potential lncRNA biomarkers in human lung adenocarcinoma.

Abnormal expression of long non-coding RNAs (lncRNAs) have been shown to play an important role in tumor biology. The Cancer Genome Atlas (TCGA) platform is a large sample sequencing database of lncRNAs, and further analysis of the associations between these data and patients' clinical related information can provide new approaches to find the functions of lncRNA. In the present study, 361 RNA sequencing profiles of gastric cancer (GC) patients were selected from TCGA. Then, we constructed the lncRNA-miRNA-mRNA competitive endogenous RNA (ceRNA) network of GC. There were 25 GC specific lncRNAs (fold change >2, p<0.05) identified, 19 of them were included in ceRNA network. Subsequently, we selected these 19 key lncRNAs and analyzed the correlations with clinical features and overall survival, 14 of them were discriminatively expressed with tumor size, tumor grade, TNM stage and lymphatic metastasis (p<0.05). In addition, eight lncRNAs (RPLP0P2, FOXD2-AS1, H19, TINCR, SLC26A4-AS1, SMIM10L2A, SMIM10L2B and SNORD116-4) were found to be significantly associated with overall survival (log-rank p<0.05). Finally, two key lncRNAs HOTAIR and UCA1 were selected for validation of their expression levels in 82 newly diagnosed GC patients by qRT-PCR. Results showed that the fold changes between TCGA and qRT-PCR were 100% in agreement. In addition, we also found that HOTAIR was significantly correlated with tumor size and lymphatic metastasis (p<0.05), and UCA1 was significantly correlated with tumor size, TNM stage and lymphatic metastasis (p<0.05). The clinical relevance of the two lncRNAs and the bioinformatics analysis results were almost the same. Overall, our study showed the GC specific lncRNAs expression patterns and a ceRNA network in GC. Clinical features related to GC specific lncRNAs also suggested these lncRNAs are worthwhile for further study as novel candidate biomarkers for the clinical diagnosis of GC and potential indicators for prognosis.

/pdf/integrated-analysis-of-long-non-coding-rna-competing-rb78qe3dm9.pdf

Integrated analysis of long non-coding RNA competing interactions reveals the potential role in progression of human gastric cancer

The long non-coding RNA (lncRNA) urothelial carcinoma-associated 1 (UCA1) has been recently shown to be dysregulated during disease occurrence and to play an important role in the progression of several cancers. However, the biological role and potential regulation mechanism of UCA1 in the carcinogenesis of gastric cancer remain unclear. In the present study, we found that UCA1 was aberrantly upregulated in gastric cancer tissues and gastric cancer cell lines, and was associated with TNM stage and metastasis. UCA1 silencing significantly inhibited gastric cancer BGC-823 cell proliferation and increased its apoptosis. We also found that UCA1 played an important role in the migration and invasion of gastric cancer cells in vitro and in vivo. The molecular mechanism of UCA1 suggested that UCA1 regulates the PI3K-Akt-mTOR signaling proteins and their downstream mediators, to alter gastric cancer progression in vitro and in vivo. Collectively, the results showed a pivotal role of UCA1 in the tumorigenesis of gastric cancer. In addition, the study characterized a novel lncRNA-mRNA regulatory network, which may lead to a better understanding of the pathogenesis of gastric cancer and assist in lncRNA-directed diagnosis and therapy for this malignancy.

/pdf/dysregulated-lncrna-uca1-contributes-to-the-progression-of-428n9esxol.pdf

Dysregulated lncRNA-UCA1 contributes to the progression of gastric cancer through regulation of the PI3K-Akt-mTOR signaling pathway

Lung cancer is one of the most lethal malignancies worldwide. To reduce the high morbidity and mortality of the disease, sensitive and specific biomarkers for early detection are urgently needed. Tumor-specific microRNAs (miRNAs) seem to be potential biomarkers for the early diagnosis and treatment of cancer. In this study, the microarray of miRNAs and mRNAs on the same samples was performed and the intersection taken with The Cancer Genome Atlas (TCGA) lung cancer miRNA/RNAseq dataset. Then, miRNA-mRNA regulatory network was constructed to identify miRNA candidates associated with lung cancer through integrating gene expression and miRNA-target prediction. Furthermore, the expression levels of miRNA candidates were validated by stem-loop real-time reverse transcription PCR (qRT-PCR) in larger lung cancer population. The relationship between signature miRNAs and the risk of lung cancer were assessed by conditional logistic regression analysis. Diagnostic value of these miRNAs was determined by areas under receiver operating characteristic curves (ROC). The Affymetrix microarray analysis identified a total of 116 miRNAs and 502 mRNAs that could distinguish lung tumor tissues from adjacent non-tumor tissues, of which 70 miRNAs and 136 mRNAs were upregulated, while 46 miRNAs and 366 mRNAs were downregulated, respectively. In combination with TCGA analysis, we identified 32 miRNAs and 377 mRNAs related to lung cancer. Then, 28 key miRNAs related to 61 inter-section mRNAs were identified by miRNA-mRNA network analysis. The miRNA function analysis was indicative of that 18 upregulated and 10 downregulated miRNAs involved in signaling pathways related to Environmental Information Processing and Human Diseases. Population result showed that the expression of 7 miRNAs (miR-205-5p, miR-3917, miR-30a-3p, miR-30a-5p, miR-30c-2-3p, miR-30d-5p and miR-27a-5p) was consistent with the analysis result of microarray and TCGA. In addition, upregulation of miR-205-5p, miR-3917 and downregulation of miR-30a-3p, miR-30a-5p, miR-30c-2-3p, miR-30d-5p, miR-27a-5p increased the risk of lung cancer by conditional logistic regression analysis. The diagnostic accuracy of miR-205-5p, miR-3917, miR-27a-5p, miR-30a-3p, miR-30a-5p, miR-30c-2-3p, miR-30d-5p showed that their corresponding AUCs were 0.728, 0.661, 0.637, 0.758, 0.772, 0.734, 0.776, respectively. Therefore, there are a set of signature miRNAs which may be promising biomarkers for the early screening of high-risk populations and early diagnosis of lung cancer.

/pdf/differential-expression-profiles-of-micrornas-as-potential-37cxwvjbow.pdf

Differential expression profiles of microRNAs as potential biomarkers for the early diagnosis of lung cancer

To investigate the potential candidate microRNA (miRNA) biomarkers for the clinical diagnosis, classification, and prognosis of gastric cancer (GC). We use bioinformatics overlapping subclasses analysis to find the tumor grade and lymphatic metastasis-related GC specific miRNAs from the Cancer Genome Atlas (TCGA) database. Then, we further investigated these GC specific miRNAs distributions in different GC clinical features and their correlations overall survival on the basis of GC patients’ information and their related RNA sequencing profile from TCGA. Finally, we randomly selected some of key miRNAs use qRT-PCR to confirm the reliability and validity. 22 GC specific key miRNAs were identified (Fold-change >2, P < 0.05), 11 of them were discriminatively expressed with tumor size, grade, TNM stage and lymphatic metastasis (P < 0.05). In addition, nine miRNAs (miR-196b-5p, miR-135b-5p, miR-183-5p, miR-182-5p, miR-133a-3p, miR-486-5p, miR-144-5p, miR-129-5p and miR-145-5p) were found to be significantly associated with overall survival (log-rank P < 0.05). Finally, four key miRNAs (miR-183-5p, miR-486-5p, miR-30c-2-3p and miR-133a-3p) were randomly selected to validation and their expression levels in 53 newly diagnosed GC patients by qRT-PCR. Results showed that the fold-changes between TCGA and qRT-PCR were 100 % in agreement. We also found miR-183-5p and miR-486-5p were significantly correlated with tumor TNM stage (P < 0.05), and miR-30c-2-3p and miR-133a-3p were associated with tumor differentiation degree and lymph-node metastasis (P < 0.05). These verified miRNAs clinically relevant, and the bioinformatics analysis results were almost the same. These key miRNAs may functions as potential candidate biomarkers for the clinical diagnosis, classification and prognosis for GC.

Wei-Wei Hong

Papers

Integrated analysis of long non-coding RNA‑associated ceRNA network reveals potential lncRNA biomarkers in human lung adenocarcinoma.

Integrated analysis of long non-coding RNA competing interactions reveals the potential role in progression of human gastric cancer

Dysregulated lncRNA-UCA1 contributes to the progression of gastric cancer through regulation of the PI3K-Akt-mTOR signaling pathway

Differential expression profiles of microRNAs as potential biomarkers for the early diagnosis of lung cancer

Identification and functional characterization of microRNAs reveal a potential role in gastric cancer progression