Effects of HAR1 on cognitive function in mice and the regulatory network of HAR1 determined by RNA sequencing and applied bioinformatics analysis.

Background: HAR1 is a 118-bp segment that lies in a pair of novel non-coding RNA genes. It shows a dramatic accelerated change with an estimated 18 substitutions in the human lineage since the human-chimpanzee ancestor, compared with the expected 0.27 substitutions based on the slow rate of change in this region in other amniotes. Mutations of HAR1 lead to a different HAR1 secondary structure in humans compared to that in chimpanzees. Methods: We cloned HAR1 into the EF-1α promoter vector to generate transgenic mice. Morris water maze tests and step-down passive avoidance tests were conducted to observe the changes in memory and cognitive abilities of mice. RNA-seq analysis was performed to identify differentially expressed genes (DEGs) between the experimental and control groups. Systematic bioinformatics analysis was used to confirm the pathways and functions that the DEGs were involved in. Results: Memory and cognitive abilities of the transgenic mice were significantly improved. The results of Gene Ontology (GO) analysis showed that Neuron differentiation, Dentate gyrus development, Nervous system development, Cerebral cortex neuron differentiation, Cerebral cortex development, Cerebral cortex development and Neurogenesis are all significant GO terms related to brain development. The DEGs enriched in these terms included Lhx2, Emx2, Foxg1, Nr2e1 and Emx1. All these genes play an important role in regulating the functioning of Cajal-Retzius cells (CRs). The DEGs were also enriched in glutamatergic synapses, synapses, memory, and the positive regulation of long-term synaptic potentiation. In addition, "cellular response to calcium ions" exhibited the second highest rich factor in the GO analysis. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of the DEGs showed that the neuroactive ligand-receptor interaction pathway was the most significantly enriched pathway, and DEGs also notably enriched in neuroactive ligand-receptor interaction, axon guidance, and cholinergic synapses. Conclusion: HAR1 overexpression led to improvements in memory and cognitive abilities of the transgenic mice. The possible mechanism for this was that the long non-coding RNA (lncRNA) HAR1A affected brain development by regulating the function of CRs. Moreover, HAR1A may be involved in ligand-receptor interaction, axon guidance, and synapse formation, all of which are important in brain development and evolution. Furthermore, cellular response to calcium may play an important role in those processes.

The role of miR-145-5p in esophageal squamous cell carcinoma tumor-associated macrophages and selection of immunochemotherapy.

Background: The impact of miR-145-5p in immune infiltration and the potential application in esophageal squamous cell carcinoma (ESCC) immunochemotherapy remains unknown. Methods: Transcriptomic data for ESCC tissues and normal tissues and clinical materials of patients with ESCC were obtained from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) databases. The differences in mRNA levels in cancer tissues and noncancerous tissues were analyzed, and we subsequently investigated the association between miR-145-5p expression and the key parameters of ESCC progression and prognosis. Additionally, cytological experiments were performed to evaluate the biological functions of miR-145-5p. Pathways potentially affected by miR-145-5p were analyzed by Gene Set Enrichment Analysis (GSEA) and REACTOME. We also analyzed the function of miR-145-5p in immune infiltration through the TIMER2 (Tumor Immune Estimation Resource) database. Results: The analysis of gene chip data from the TCGA database and GEO database (including GSE13937 and GSE43732) showed that the expression of miR-145-5p is downregulated in ESCC (P<0.05) and that patients with high miR-145-5p levels had lower survival rates (P<0.05). The expression of miR-145-5p was significantly correlated with the disease-free survival (DFS) rate (P<0.05) and M stage (P<0.05) in the TCGA database and was significantly correlated with the T stage (P<0.05) and TNM stage (P<0.05) in the GSE13937 database. Functional experiments showed that miR-145-5p attenuated proliferation (P<0.05), migration (P<0.01) and invasion (P<0.01) in the Eca109 cell line. Both GSEA gene enrichment and REACTOME gene enrichment revealed that miR-145-5p was associated with tumor signaling pathways and immune signaling pathways. Immune infiltration analysis revealed that the expression level of miR-145-5p was significantly correlated with the infiltration level of macrophages (P<0.05) and was positively correlated with the level of gene markers of M2 macrophages and tumor-associated macrophages (P<0.05). Conclusions: MiR-145-5p acts as a tumor suppressor microRNA in ESCC and is an important noncoding RNA in the high M2-like tumor-associated macrophage infiltration of ESCC. Assessing the miR-145-5p level in ESCC samples has translational meaning, which help illustrate the immune infiltration status, predict the prognostic outcome, and select the type of immunochemotherapy.

Mining the Prognostic Role of DNA Methylation Heterogeneity in Lung Adenocarcinoma.

Purpose: DNA methylation heterogeneity is a type of tumor heterogeneity in the tumor microenvironment, but studies on the identification of the molecular heterogeneity of the lung adenocarcinoma genome with respect to DNA methylation sites and their roles in lung cancer progression and prognosis are scarce. Methods: Prognosis-associated DNA methylation subtypes were filtered by the Cox proportional hazards model and then established by unsupervised cluster analysis. Association analysis of these subtypes with clinical features and functional analysis of annotated genes potentially affected by methylation sites were performed. The robustness of the model was further tested by a Bayesian network classifier. Results: Over 7 thousand methylation sites were associated with lung adenocarcinoma prognosis. We identified seven molecular methylation subtypes, including 630 methylation sites. The subtypes yielded the most stable results for differentiating methylation profiles, prognosis, and gene expression patterns. The annotated genes potentially affected by these methylation sites are enriched in biological processes such as morphogenesis and cell adhesion, but their individual impact on the tumor microenvironment and prognosis is multifaceted. Discussion. We revealed that DNA methylation heterogeneity could be clustered and associated with the clinical features and prognosis of lung adenocarcinoma, which could lead to the development of a novel molecular tool for clinical evaluation.

Potential Impact of Cancer Susceptibility Genes on Lung Cancer Metastasis.

Background: Studies of prognosis-related molecular markers are an important tool to uncover the mechanism of tumour metastasis. Cancer susceptibility gene testing is an important tool for genetic counselling of cancer risk. However, the impact of lung cancer susceptibility genes (LCSGs) on lung cancer metastasis and prognosis has not been well studied. Methods: The list of lung cancer susceptibility genes was retrospectively analysed and updated. After expression profiling and functional analysis, LCSG-based signatures for prognosis were identified by Cox regression and LASSO regression analyses. For translational purposes, nomograms integrating LCSGs and clinical characteristics were constructed. Results: A total of 301 LCSGs were employed for modelling. For lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC), 10-gene and 7-gene signatures were created and independently validated. The LCSG-based risk score could stratify LUAD survival (univariate: hazard ratio (HR) = 1.076, 95% confidence interval (CI) = 1.049-1.103, P < 0.001; multivariate: HR = 1.066, 95% CI = 1.037-1.095, P < 0.001) and LUSC survival (univariate: HR = 1.149, 95% CI = 1.066-1.239, P < 0.001; multivariate: HR = 1.129, 95% CI = 1.038-1.228, P = 0.005). One of the processes affected by differentially expressed genes in both LUAD and LUSC was the negative regulation of epithelial cell differentiation. Conclusions: Overall, novel LCSG-based gene signatures for LUAD and LUSC were constructed. These findings could expand the understanding of the impact of LCSG expression on cancer metastasis and prognosis.

5-methylcytosine RNA methylation regulators affect prognosis and tumor microenvironment in lung adenocarcinoma.

Background: Accumulating evidence has shown that 5-methylcytosinec (m5C) RNA methylation plays an essential role in tumorigenesis. However, the roles of m5C regulators in the prognosis, tumor microenvironment (TME), and immunotherapy responses of lung adenocarcinoma (LUAD) have not been fully analyzed. Methods: Based on 14 m5C RNA regulators, we evaluated the m5C RNA modification patterns in patients with LUAD (n=594) in The Cancer Genome Atlas (TCGA). Unsupervised clustering analysis was performed to confirm distinct m5C modification patterns. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed to investigate the biological functions of differentially expressed genes (DEGs) among different m5C RNA modification patterns. An m5C signature (m5Csig) was constructed using least absolute shrinkage and selection operator (LASSO) algorithms. The GSE72094 cohort (n=442) from the Gene Expression Omnibus (GEO) was used to validate m5Csig. A receiver operating characteristic (ROC) model was constructed to evaluate the sensitivity and specificity of m5Csig. Tumor-infiltrating immune cells (TIICs) between the high- and low-risk groups were estimated using the Cell Type Identification By Estimating Relative Subsets Of RNA Transcripts (CIBERSORT) algorithm. Results: We identified 3 m5C RNA modification clusters. Overall survival (OS) differed among the 3 clusters. The m5Csig, including TRDMT1, NSUN1, NSUN4, NSUN7, and ALYREF, was constructed to classify patients with LUAD into high- and low-risk groups. The high-risk group, with more immune cell infiltration, had a significantly poorer OS than that the low-risk group, which was associated with better response to immune checkpoint blockade therapy. Conclusions: The present study revealed that m5C RNA regulators play a significant role in TME regulation in LUAD. The m5Csig can predict the prognosis of patients with LUAD and might provide novel strategies for tumor immunotherapy.

Integrative pan cancer analysis reveals the importance of CFTR in lung adenocarcinoma prognosis.

Cystic fibrosis (CF) and cystic fibrosis transmembrane conductance regulator (CFTR) mutations have been shown to be associated with the risk of a variety of cancers. However, the clinical significance of aberrant CFTR gene expression in human tumors remains unknown. The expression profiles and prognostic landscapes of CFTR in human cancers were identified from the PubMed, OVID, CNKI, TCGA, ONCOMINE, PrognoScan, and GEPIA databases. Over 11, 000 cancer samples from the literature, GEPIA database, and PrognoScan database were included in this study. In general, CFTR has various expression and prognostic profiles in cancers, but the results from cross-database and meta-analyses revealed that CFTR is a robust biomarker for LUAD prognosis. Collectively, this study suggests that CFTR is an important prognostic biomarker for LUAD survival, implying that it could be used as a prognostic biomarker and therapeutic target for LUAD.

Trends, Symptoms, and Outcomes of Resectable Giant Mediastinal Tumors.

Describing the changes in surgical procedures and factors affecting the surgical outcome of patients who have undergone complete resection of giant mediastinal tumors (GMTs, diameter ≥ 10 centimeters) could improve preoperative decision-making and prognostic evaluations. We accessed data from three sources, which are case reports on surgical treatment of GMTs from PubMed, Web of Science, and EMBASE until June 1, 2019; patients with resected GMT from the Surveillance, Epidemiology, and End Results (SEER) database; and retrospective review of medical records in our institution from 2000 to 2019. The worldwide distribution, clinicopathological characteristics, symptom profile, prognosis of patients with GMT resection, and nomogram for surgical outcome prediction are reported. A total of 242 rare GMT cases from four continents (Asia, North America, South America, and Europe) were included. The median age of the patients was 40 (IQR: 27, range: 13-83) years, and the male-to-female ratio was 1.57:1. Dyspnea, shortness of breath, cough, and chest pain or discomfort were the major symptoms at presentation. The prognosis of benign and low-grade malignant GMTs was superior to that of high-grade malignant GMTs. Tumor malignancy played the most critical role in predicting postoperative survival, followed by longest tumor diameter and a posterior mediastinum location. The findings of this study suggest that the number of successful GMT surgeries has increased over the last decade and describe clinical features of GMTs. Physicians should prioritize tumor malignancy as a leading factor in predicting outcome rather than tumor size.

The first comprehensive database of germline pathogenic variants in East Asian cancer patients.

Pathogenic germline variants in cancer-associated genes are risk factors for cancer predisposition. However, systematic mining and summarizing of cancer pathogenic or likely pathogenic variants has not been performed for people of East Asian descent. This study aimed to investigate publicly available data to identify germline variants in East Asian cancer cohorts and compare them to variants in Caucasian cancer cohorts. Based on the data we retrieved, we built a comprehensive database, named COGVIC (Catalog of Germline Variants in Cancer). A total of 233 variants in the East Asian population were identified. The majority (87%) of genes with cancer-associated variants were not shared between the East Asian and Caucasian cohorts. This included pathogenic variants in BRCA2. Our study summarized the prevalence of germline variants in East Asian cancer cohorts and provides an easy-to-use online tool to explore germline mutations related to cancer susceptibility. DATABASE URL: http://www.cogvic.vip/.

Identification of Genetic Variants Associated with Sex-Specific Lung-Cancer Risk.

BACKGROUND: The incidence of lung cancer differs between men and women, suggesting the potential role of sex-specific influences in susceptibility to this cancer. While behavioural differences may account for some of the risk, another possibility is that X chromosome susceptibility genes may have an effect. Little is known about genetic variants on the X chromosome that contribute to sex-specific lung-cancer risk, so we investigated this in a previously characterized cohort. METHODS: We conducted a genetic association reanalysis of 518 lung cancer patients and 844 controls to test for lung cancer susceptibility variants on the X chromosome. Annotated gene expression, co-expression analysis, pathway, and immune infiltration analyses were also performed. RESULTS: 24 SNPs were identified as significantly associated with male, but not female, lung cancer cases. These resided in blocks near the annotated genes DMD, PTCHD1-AS, and AL008633.1. Of these, DMD was differentially expressed in lung cancer cases curated in The Cancer Genome Atlas. A functional enrichment and a KEGG pathway analysis of co-expressed genes revealed that differences in immune function could play a role in sex-specific susceptibility. CONCLUSIONS: Our analyses identified potential genetic variants associated with sex-specific lung cancer risk. Integrating GWAS and RNA-sequencing data revealed potential targets for lung cancer prevention.

Current Evidence of the Efficacy and Safety of Neoadjuvant EGFR-TKIs for Patients With Non-small Cell Lung Cancer.

PURPOSE: Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) have been indicated to be an effective treatment for advanced EGFR-mutant NSCLC. However, the neoadjuvant application of EGFR-TKIs in resectable NSCLC needs further investigation. Here, we aimed to evaluate the efficacy and safety of neoadjuvant EGFR-TKIs for lung cancer. METHODS: Published studies on neoadjuvant EGFR-TKIs in NSCLC were identified in PubMed, Web of Science, and EMBASE until June 1, 2020. Data on surgical rates, objective response rates (ORRs), pathologic responses, and adverse event (AE) rates were retrieved for proportional meta-analysis. RESULTS: In total, 7 enrolled studies involving 129 EGFR-TKI-sensitive NSCLC patients were included in this analysis. The overall surgical rate in these studies was 95% (95% CI: 83% to 100%), with an ORR of 48% (95% CI: 39% to 57%) in the population with EGFR-TKI-sensitive mutations, whereas the ORR including wild-type EGFR patients was 28% (95% CI: 14% to 44%). The rate of grade 1-2 AEs was 69% (95% CI: 41% to 91%) but with an acceptable rate of grade 3-4 AEs of 0% (95% CI: 0% to 5%). The pooled rates of rash and diarrhea were 56% (95% CI: 31% to 79%) and 25% (95% CI: 6% to 51%), respectively. The impact of neoadjuvant EGFR-TKIs on survival remains inconclusive. CONCLUSIONS: Neoadjuvant EGFR-TKIs showed objective responses in approximately half of EGFR-sensitive NSCLC patients with a tolerable adverse effect profile. The favorable impact of neoadjuvant EGFR-TKIs on NSCLC needs more evidence for validation, such as the comparison of survival improvement between EGFR-TKIs and chemotherapy. The efficacy of neoadjuvant next-generation EGFR-TKIs in clinical trials remains unclear.

The intra-class heterogeneity of immunophenotyping and immune landscape in oesophageal cancer and clinical implications.

BACKGROUND: The response rate and survival benefit of immunotherapy vary among patients, implying specific immune status of an individual could be associated with the effect of immunotherapy. However, in-depth studies of immune subtypes (ISs), immune landscape and tumour microenvironment of oesophageal cancer (ESCA) and their clinical implications are less reported. METHODS: We first accessed data from publicly available databases and preprocessed it based on a standard protocol. Then, ISs were identified by unsupervised learning. Thereafter, the association of these ISs and tumour mutation burden (TMB), biomarkers of chemotherapy-induced immune response, tumour markers were also assessed. In addition, the immune characteristics, immune landscape, co-expression network of immune genes, and clinical implications were visualized and analysed. RESULTS: We identified three immunoclusters based on immune-associated genes with intra-class heterogeneity and prognostic value. Cluster-specific associations with TMB, markers of chemotherapy-induced immune response, and tumour markers were revealed. A 4-gene signature (risk score= -0.16514291×BHLHE22-0.03964046×MXRA8-0.15242778×SLIT2-0.05553572×SPON1) based on co-expressed genes in the immunoclusters was developed and externally validated. CONCLUSIONS: In summary, we identified clinically relevant immunoclusters in both adenocarcinoma and squamous cell carcinoma of oesophagus, revealing the necessity of assessing the complexity and diversity of immune microenvironment for cancer immunotherapy.

Lung segmentectomy assisted by highly selective independent segmental ventilation: a series of three cases.

BACKGROUND: The identification of targeted intersegmental planes and resection with adequate surgical margins are among the crucial steps in anatomical pulmonary segmentectomy, and technical improvements are still needed. CASE PRESENTATION: We reported three cases of intersegmental plane identification using highly selective independent segmental ventilation during segmentectomy. All cases required cooperation with an anesthesiologist who was able to perform segmental ventilation and double confirmation of segmental bronchus branches by the surgeon. The surgical procedure provides a direct visualization of spare segment inflation and saves time in deflation over the conventional residual segment inflation method. CONCLUSIONS: Highly selective independent segmental ventilation could be considered a suitable option for pulmonary intersegmental plane identification and could be universally used for lung segmentectomy.

Multi-Omics Characterization and Origin Assessment of Bilateral Lung Adenoid Cystic Carcinoma: A Case Report.

BACKGROUND: Primary adenoid cystic carcinoma (ACC) of the lung, which arises from the bronchial gland and is rare, accounting for only 0.04-0.2% of all primary lung tumors. The genetic profiling of bilateral ACC of unknown primary site and application in postoperative decision-making are less reported. CASE PRESENTATION: A 57-year-old male with a smoking history of over 30 years and multiple nodules in both lungs was present to our department. After assessing the bilateral solid nodules in his Positron Emission Tomography-Computed Tomography (PET/CT) scan, malignant lesions at the left lower lung, right lower lung, and right middle lung are suspected. Sequential selective video-assisted thoracoscopic surgeries (VATS) were performed. A genetic alteration test of 425 cancer-related genes and global gene expression profile of the specimens revealed intrapulmonary metastasis existed. The patient was followed up for three years without recurrence and tissue mutations in liquid biopsy. CONCLUSION: We present a way of omics-based multiple pulmonary lesions origin assessment, facilitating post-operative differential diagnosis and treatment decision for difficult cases.

Cross-platform genomic identification and clinical validation of breast cancer diagnostic biomarkers.

INTRODUCTION: Circulating non-coding RNA is an ideal source to discover novel biomarkers for non-invasive screening. However, studies for the discovery of universal miRNAs in serum and exosomes for breast cancer early diagnosis are limited. METHODS: Based on bioinformatic analysis, in vitro and in vivo studies were performed to understand the role of identified hsa-miR-423-5p in cancer proliferation, migration, cancer stem cell properties. Next, global non-coding RNA expression profiles in blood serum and exosome were performed. hsa-miR-423-5p expression from a total of 356 peripheral blood samples was evaluated and the association of hsa-miR-423-5p expression with clinical characteristics, sensitivity and specificity for breast cancer diagnosis were assessed. RESULTS: The expression of serum and exosomal hsa-miR-423-5p is abnormally increased in breast cancer. Suppression of hsa-miR-423-5p inhibited cell proliferation and invasion in both T47D and MDA-MB-231 breast cancer cell lines, and tumor growth in vivo. Compared with 113 healthy women, quantification analysis of hsa-miR-423-5p in 224 breast cancer samples confirmed the abnormal expression. Serum hsa-miR-423-5p was significantly associated with the clinical stage (P=0.001) and Ki-67 level (P=0.004). CONCLUSIONS: A translational bioinformatics analysis procedure and validation by in vitro, in vivo, and clinical samples reveal that hsa-miR-423-5p could be used as a non-invasive breast cancer biomarker.

Novel PHOX2B germline mutation in childhood medulloblastoma: a case report.

BACKGROUND: Medulloblastoma is an aggressive brain tumor mostly found in children, few studies on pathogenic germline mutations predisposing this disease was reported. CASE PRESENTATION: We present an 11-year-old male with medulloblastoma, who harbors a de novo PHOX2B germline mutation as detected by whole exome sequencing (WES). Family history was negative. Sanger sequencing confirmed this mutation in peripheral blood, hair bulbs, urine and saliva. Identification of novel germline mutations is beneficial for childhood cancer screening. CONCLUSIONS: This case revealed a de novo PHOX2B germline mutation as a potential cause of medulloblastoma in a child and suggests familial germline variant screening is useful when an affected family is considering having a second child.

Clinicopathological features, survival outcomes, and appropriate surgical approaches for stage I acinar and papillary predominant lung adenocarcinoma.

BACKGROUND: Whether prognosis differs between lung acinar predominant adenocarcinoma (ACN) and papillary predominant adenocarcinoma (PAP) patients remains controversial. Furthermore, the appropriate surgical plan for each subtype is undetermined. METHODS: Data of stage I ACN or PAP patients from 2004 to 2015 were retrospectively reviewed by SEER*Stat 8.3.5. The primary outcome was overall survival (OS) and lung cancer specific survival (LCSS). RESULTS: 1531 patients (PAP, 484; ACN, 1047) were included. ACN patients had better OS (P = .001) and LCSS (P = .003) than PAP patients. Among stage I ACN patients, lobectomy with mediastinal lymph node dissection (Lob) (P = .001) or segmentectomy (Seg) (P = .003) provided a better OS than wedge resection (Wed). And ACN patients who received Lob had a equivalent LCSS, compared to those who received Seg (P = .895). For patients with PAP in stage I, those who received Lob tended to have a better prognosis than that received Seg (HR of OS, 0.605, 95% CI: 0.263-1.393; HR of LCSS, 0.541, 95% CI: 0.194-1.504) or Wed (HR of OS, 0.735, 95% CI: 0.481-1.123; HR of LCSS, 0.688, 95% CI: 0.402-1.180). CONCLUSIONS: Among patients with lung adenocarcinoma in stage I, those with ACN have a better OS and LCSS than that with PAP. For patients with stage I ACN, Seg and Lob, rather than Wed, seem to be an equivalent treatment choice; however, Seg is the prior option because it could preserve more lung function than Lob. For patients with PAP, Lob tends to be a better choice than Wed and Seg, although the prognostic difference between them is nonsignificant.

IRGS: an immune-related gene classifier for lung adenocarcinoma prognosis.

BACKGROUND: Tumour cells interfere with normal immune functions by affecting the expression of some immune-related genes, which play roles in the prognosis of cancer patients. In recent years, immunotherapy for tumours has been widely studied, but a practical prognostic model based on immune-related genes in lung adenocarcinoma comparable to existing model has not been established and reported. METHODS: We first obtained publicly accessible lung adenocarcinoma RNA expression data from The Cancer Genome Atlas (TCGA) for differential gene expression analysis and then filtered immune-related genes based on the ImmPort database. By using the lasso algorithm and multivariate Cox Proportional-Hazards (CoxPH) regression analysis, we identified candidate genes for model development and validation. The robustness of the model was further examined by comparing the model with three established gene models. RESULTS: Gene expression data from a total of 524 lung adenocarcinoma patients from TCGA were used for model development. We identified four biomarkers (MAP3K8, CCL20, VEGFC, and ANGPTL4) that could predict overall survival in lung adenocarcinoma (HR = 1.98, 95% CI 1.48 to 2.64, P = 4.19e-06) and this model could be used as a classifier for the evaluation of low-risk and high-risk groups. This model was validated with independent microarray data and was highly comparable with previously reported gene expression signatures for lung adenocarcinoma prognosis. CONCLUSIONS: In this study, we identified a practical and robust four-gene prognostic model based on an immune gene dataset with cross-platform compatibility. This model has potential value in improving TNM staging for survival predictions in patients with lung adenocarcinoma. IMPACT: The study provides a method of immune relevant gene prognosis model and the identification of immune gene classifier for the prediction of lung adenocarcinoma prognosis with RNA sequencing and microarray compatibility.

Machine Learning Models to Predict Primary Sites of Metastatic Cervical Carcinoma From Unknown Primary.

Metastatic cervical carcinoma from unknown primary (MCCUP) accounts for 1-4% of all head and neck tumors, and identifying the primary site in MCCUP is challenging. The most common histopathological type of MCCUP is squamous cell carcinoma (SCC), and it remains difficult to identify the primary site pathologically. Therefore, it seems necessary and urgent to develop novel and effective methods to determine the primary site in MCCUP. In the present study, the RNA sequencing data of four types of SCC and Pan-Cancer from the cancer genome atlas (TCGA) were obtained. And after data pre-processing, their differentially expressed genes (DEGs) were identified, respectively. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis indicated that these significantly changed genes of four types of SCC share lots of similar molecular functions and histological features. Then three machine learning models, [Random Forest (RF), support vector machine (SVM), and neural network (NN)] which consisted of ten genes to distinguish these four types of SCC were developed. Among the three models with prediction tests, the RF model worked best in the external validation set, with an overall predictive accuracy of 88.2%, sensitivity of 88.71%, and specificity of 95.42%. The NN model is the second in efficacy, with an overall accuracy of 82.02%, sensitivity of 81.23%, and specificity of 93.04%. The SVM model is the last, with an overall accuracy of 76.69%, sensitivity of 74.81%, and specificity of 90.84%. The present analysis of similarities and differences among the four types of SCC, and novel models developments for distinguishing four types of SCC with informatics methods shed lights on precision MCCUP diagnosis in the future.