scholarly journals A functional artificial neural network for noninvasive pretreatment evaluation of glioblastoma patients

2021 ◽  
Author(s):  
Eric Zander ◽  
Andrew Ardeleanu ◽  
Ryan Singleton ◽  
Barnabas Bede ◽  
Yilin Wu ◽  
...  
Keyword(s):  
Copy Number ◽  
Significant Risk ◽  
Image Data ◽  
Copy Number Variations ◽  
Image Features ◽  
The Cancer Genome Atlas ◽  
Karnofsky Performance Scale ◽  
Genetic Profiling ◽  
Artificial Neural ◽  
Genome Atlas

Abstract Background Pretreatment assessments for glioblastoma (GBM) patients, especially elderly or frail patients, are critical for treatment planning. However, genetic profiling with intracranial biopsy carries a significant risk of permanent morbidity. We previously demonstrated that the CUL2 gene, encoding the scaffold cullin2 protein in the cullin2-RING E3 ligase (CRL2), can predict GBM radiosensitivity and prognosis. CUL2 expression levels are closely regulated with its copy number variations (CNVs). This study aims to develop artificial neural networks (ANNs) for pretreatment evaluation of GBM patients with inputs obtainable without intracranial surgical biopsies. Methods Public datasets including Ivy-GAP, The Cancer Genome Atlas Glioblastoma (TCGA-GBM), the Chinese Glioma Genome Atlas (CGGA) were used for training and testing of the ANNs. T1 images from corresponding cases were studied using automated segmentation for features of heterogeneity and tumor edge contouring. A ratio comparing the surface area of tumor borders vs. the total volume (SvV) was derived from the DICOM-SEG conversions of segmented tumors. The edges of these borders were detected using the canny edge detector. Packages including Keras, Pytorch, and TensorFlow were tested to build the ANNs. A 4-layered ANN (8-8-8-2) with a binary output was built with optimal performance after extensive testing. Results The 4-layered deep learning ANN can identify a GBM patient’s overall survival (OS) cohort with 80-85% accuracy. The ANN requires 4 inputs, including CUL2 copy number, patients’ age at GBM diagnosis, Karnofsky Performance Scale (KPS), and SvV ratio. Conclusion Quantifiable image features can significantly improve the ability of ANNs to identify a GBM patients’ survival cohort. Features such as clinical measures, genetic data, and image data, can be integrated into a single ANN for GBM pretreatment evaluation.

scholarly journals A functional artificial neural network for noninvasive presurgical evaluation of glioblastoma multiforme prognosis and radiosensitivity profiling

2020 ◽  
Author(s):  
Eric Zander ◽  
Andrew Ardeleanu ◽  
Ryan Singleton ◽  
Barnabas Bede ◽  
Yilin Wu ◽  
...  
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Glioblastoma Multiforme ◽  
Copy Number ◽  
Significant Risk ◽  
Growth Factor Receptor ◽  
Copy Number Variations ◽  
The Cancer Genome Atlas ◽  
Artificial Neural ◽  
Genome Atlas

Background and Purpose: Genetic profiling for glioblastoma multiforme (GBM) patients with intracranial biopsy carries a significant risk of permanent morbidity. We previously demonstrated that the CUL2 gene, encoding the scaffold cullin2 protein in the cullin2-RING E3 ligase (CRL2), can predict GBM radiosensitivity and prognosis mainly due to the functional involvement of CRL2 in mediating hypoxia-inducible factor 1 (HIF-1) alpha; and epidermal growth factor receptor (EGFR) degradation. Because CUL2 expression levels are closely regulated with its copy number variations (CNVs), this study aims to develop an artificial neural network (ANN) that can predict GBM prognosis and help optimize personalized GBM treatment planning. Materials and Methods: Datasets including Ivy-GAP, The Cancer Genome Atlas Glioblastoma Multiforme (TCGA-GBM), the Chinese Glioma Genome Atlas (CGGA) were analyzed. T1 images from corresponding cases were studied using automated segmentation for features of heterogeneity and tumor edge contouring. Results: We developed a 4-layer neural network that can consistently predict GBM prognosis with 80-85% accuracy with 3 inputs including CUL2 copy number, patient age at GBM diagnosis, and surface vs. volume (SvV) ratio. Conclusion: A functional 4-layer neural network was constructed that can predict GBM prognosis and potential radiosensitivity.

scholarly journals Multi-omics Prognostic Analysis of Lysine Acetylation Regulators in Glioma

2020 ◽  
Author(s):  
Zewei Tu ◽  
Shigang Lv ◽  
Lei Wu ◽  
Qing Hu ◽  
Chuming Tao ◽  
...  
Keyword(s):  
Copy Number ◽  
Protein Modification ◽  
Cox Regression ◽  
Copy Number Variations ◽  
The Cancer Genome Atlas ◽  
Lysine Acetylation ◽  
Consensus Clustering ◽  
Expression Levels ◽  
Public Data ◽  
Genome Atlas

Abstract Background: Lysine acetylation is a crucial kind of protein modification and is related to the malignant development of various cancers. But their roles in glioma are still unclear and needed concluded comprehensively. Methods: In this study, we comprehensively analyzed the expression levels of 33 lysine acetylation regulators (LARs) and prognostic roles by using public data, including the Chinese Glioma Genome Atlas (CGGA) and The Cancer Genome Atlas (TCGA). The prognostic roles of LARs were judged by univariate Cox regression. Consensus clustering was applied to result in three stratified glioma subtypes (LA1, 2, and 3) with different clinical outcomes. We also constructed a risk signature for predicting the overall survival of glioma patients by using least absolute shrinkage and selection operator regression (LASSO regression). Besides, copy number variations (CNVs) and single nucleotide polymorphism (SNP) of LARs were also analyzed in our study. Results: We found the mRNA expression levels of most of LARs were dysregulated in gliomas and associated with the prognosis of glioma patients. The risk signature constructed by 14 LARs presented an independent prognostic role in both the CGGA (HR:1.96, 95%CI:1.33-2.90) and TCGA (HR:1.48, 95%CI:1.08-2.03) datasets and robust predictive effects in the ROC curves with all of area under curves more than 0.800. Moreover, the copy number variations of LARs were also significantly related to the prognosis of glioma patients, in which HDAC1 (1p) was one of the oncogenes lost in 1p/19q codeletion events, while SIRT2 (19q) and EP300 (22q) may act as tumor suppressors in gliomas with 19q or 22q deletions, respectively. Conclusion: LARs are potential biomarkers for the malignant progression of gliomas, and our results could be useful for predicting the OS of glioma patients and provide some clues in searching the functions of LARs in glioma progression. Keywords: glioma, lysine acetylation regulator, epigenetic, prognostic signature, biomarker.

Identification of T-cell-inflamed gastric adenocarcinoma in The Cancer Genome Atlas (TCGA).

2017 ◽  
Vol 35 (7_suppl) ◽  
pp. 16-16 ◽  
Author(s):  
Steven Brad Maron ◽  
Jason John Luke ◽  
Raymond Hovey ◽  
Riyue Bao ◽  
Thomas Gajewski ◽  
...  
Keyword(s):  
T Cell ◽  
Objective Response ◽  
Copy Number Variations ◽  
Cancer Genome ◽  
The Cancer Genome Atlas ◽  
Molecular Characteristics ◽  
Chi Square ◽  
Primary Tumors ◽  
Cancer Genome Atlas ◽  
Genome Atlas

16 Background: Gastroesophageal adenocarcinoma (GEC) is a significant global health problem. KEYNOTE-012 demonstrated a 22% objective response rate in patients with PD-L1 expressing GEC that received pembrolizumab. A subset of patients (pts) tumors express a T cell “inflamed” (TCI) phenotype, which can be measured using an IFN-γ-based immune signature and may prove more predictive of clinical benefit. Using a 160 gene RNA-Seq immune expression profile, we sought to characterize the molecular environments of TCI versus non-TCI GEC patients in The Cancer Genome Atlas (TCGA). Methods: 395 GEC pts with primary tumors in TCGA were clustered into TCI, non-TCI, and intermediate subtypes using both unsupervised hierarchical and K-means clustering (k = 3). Molecular characteristics were categorized using data acquired via CbioPortal and the UCSC Xena repository. Only non-silent somatic mutations and copy number variations (CNVs) reaching GISTIC2 -2 or +2 thresholds were considered. Statistical comparisons were performed using chi-square, ANOVA, and t-test. Results: The TCI subtype contained patients from all TCGA-defined subtypes - EBV-associated (56%), MSI-high (16%), chromosomal unstable (6%), and genomically stable (27%). No significant differences were seen between TCI and non-TCI for tumor site or stage. Mutations in PTEN, PIK3CA, CDH1, and RHOA were more frequent in TCI patients. ERBB2, CCNE1, and KRAS CNVs were infrequent in TCI patients as were PDE4D deletions ( p< 0.05 ). TCI tumors had higher expression of both co-inhibitory (PD-1, PD-L1/L2, CD28/80, BTLA, LAG3) and co-stimulatory (CD137/40/27, OX40, GITR, ICOS) checkpoint molecules ( p< 10-7). Total mutation burden was no different between TCI and non-TCI pts when excluding MSI-high pts nor when assessing MSI-high alone. Conclusions: The IFN immune phenotype encompassed GEC patients from all TCGA subsets. Correlation of clinical outcome with checkpoint blockade is necessary to confirm these molecular associations and the independent predictive utility of this immune-profile stratification.

scholarly journals Integrative Analysis of Methylation and Copy Number Variations of Prostate Adenocarcinoma Based on Weighted Gene Co-expression Network Analysis

2021 ◽  
Vol 11 ◽  
Author(s):  
Yaxin Hou ◽  
Junyi Hu ◽  
Lijie Zhou ◽  
Lilong Liu ◽  
Ke Chen ◽  
...  
Keyword(s):  
Network Analysis ◽  
Clinical Outcomes ◽  
Copy Number ◽  
Rho Gtpase ◽  
Specific Antigen ◽  
The United States ◽  
Prostate Adenocarcinoma ◽  
Copy Number Variations ◽  
The Cancer Genome Atlas ◽  
Prognostic Indicators

Prostate adenocarcinoma (PRAD) is the most pervasive carcinoma diagnosed in men with over 170,000 new cases every year in the United States and is the second leading cause of death from cancer in men despite its indolent clinical course. Prostate-specific antigen testing, which is the most commonly used non-invasive diagnostic method for PRAD, has improved early detection rates in the past decade, but its effectiveness for monitoring disease progression and predicting prognosis is controversial. To identify novel biomarkers for these purposes, we carried out weighted gene co-expression network analysis of the top 10,000 variant genes in PRAD from The Cancer Genome Atlas in order to identify gene modules associated with clinical outcomes. Methylation and copy number variation analysis were performed to screen aberrantly expressed genes, and the Kaplan–Meier survival and gene set enrichment analyses were conducted to evaluate the prognostic value and potential mechanisms of the identified genes. Cyclin E2 (CCNE2), rhophilin Rho GTPase-binding protein (RHPN1), enhancer of zeste homolog 2 (EZH2), tonsoku-like DNA repair protein (TONSL), epoxide hydrolase 2 (EPHX2), fibromodulin (FMOD), and solute carrier family 7 member (SLC7A4) were identified as potential prognostic indicators and possible therapeutic targets as well. These findings can improve diagnosis and disease monitoring to achieve better clinical outcomes in PRAD.

scholarly journals A Timely Update of Immunohistochemistry and Molecular Classification in the Diagnosis and Risk Assessment of Endometrial Carcinomas

2021 ◽  
Vol 145 (11) ◽  
pp. 1367-1378
Author(s):  
Minhua Wang ◽  
Pei Hui
Keyword(s):  
Copy Number ◽  
Treatment Strategies ◽  
The United States ◽  
Cancer Genome ◽  
The Cancer Genome Atlas ◽  
Molecular Profile ◽  
Cancer Genome Atlas ◽  
Endometrial Carcinomas ◽  
Genome Atlas

Context.— Endometrial carcinoma is the most common gynecologic malignancy in the United States and has been traditionally classified based on histology. However, the distinction of certain histologic subtypes based on morphology is not uncommonly problematic, and as such, immunohistochemical study is often needed. Advances in comprehensive tumor sequencing have provided novel molecular profiles of endometrial carcinomas. Four distinct molecular subtypes with different prognostic values have been proposed by The Cancer Genome Atlas program: polymerase epsilon ultramutated, microsatellite instability hypermutated, copy number low (microsatellite stable or no specific molecular profile), and copy number high (serouslike, p53 mutant). Objective.— To discuss the utilities of commonly used immunohistochemical markers for the classification of endometrial carcinomas and to review the recent advancements of The Cancer Genome Atlas molecular reclassification and their potential impact on treatment strategies. Data Sources.— Literature review and authors' personal practice experience. Conclusions.— The current practice of classifying endometrial cancers is predominantly based on morphology. The use of ancillary testing, including immunohistochemistry, is helpful in the identification, differential diagnosis, and classification of these cancers. New developments such as molecular subtyping have provided insightful prognostic values for endometrial carcinomas. The proposed The Cancer Genome Atlas classification is poised to gain further prominence in guiding the prognostic evaluation for tailored treatment strategies in the near future.

scholarly journals Identification of factors regulating the gene expression of alcohol dehydrogenases in hepatocellular carcinoma

2021 ◽  
Author(s):  
Jinghe Xie ◽  
Yaqi Qiu ◽  
Shuai Zhang ◽  
Keqing Ma ◽  
Yimeng Ou ◽  
...  
Keyword(s):  
Gene Expression ◽  
Hepatocellular Carcinoma ◽  
Dna Methylation ◽  
Copy Number ◽  
Gene Copy Number ◽  
Copy Number Variations ◽  
The Cancer Genome Atlas ◽  
Gene Copy ◽  
Excessive Alcohol Consumption ◽  
Alcohol Dehydrogenases

Abstract Background Excessive alcohol consumption has been documented to increase the risk of liver hepatocellular carcinoma (HCC) development. Accordingly, a broad interest pointed to alcohol dehydrogenases (ADHs), which display essential roles in alcohol metabolism. Despite the relevance of ADHs expression and the prognosis of HCC has been estimated, so far, limited research concerning the factors that are responsible for the regulation of ADHs expression has been reported. Methods In this study, using The Cancer Genome Atlas (TCGA) and RegNetwork database, we predicted potential factors consisting of DNA methylation, gene copy number variations, transcription factors (TFs) and microRNAs (miRNAs) that might impact ADHs gene expression in HCC. Results We found that DNA methylation induced the down-regulated expression of ADH1B. Of note, our results implicated that gene copy number variation might not have effects on ADHs expression. Regarding TFs, we speculated that NFYA modulated ADH1C, E2F1 and TFAP2A regulated ADH6 expression based on their expression and prognostic value. Moreover, miR-185 and miR-561 might elicit the repression of ADH4, and miR-105 might impair ADH6 expression. Conclusion This study revealed that multiple factors, including DNA methylation, TFs and microRNAs, affect the expression of ADH family members, which provided new insights into discovering promising HCC-suppressive targets.

scholarly journals A comprehensive genome-wide profiling comparison between HBV and HCV infected hepatocellular carcinoma

2019 ◽  
Vol 12 (1) ◽  
Author(s):  
Suofeng Sun ◽  
Yuan Li ◽  
Shuangyin Han ◽  
Hongtao Jia ◽  
Xiuling Li ◽  
...  
Keyword(s):  
Gene Expression ◽  
Hepatocellular Carcinoma ◽  
Copy Number ◽  
Survival Rates ◽  
Gene Expression Pattern ◽  
Asian Population ◽  
Copy Number Variations ◽  
The Cancer Genome Atlas ◽  
Integrated Analysis ◽  
Immune Related Genes

Abstract Background Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide, especially in East Asia. Even with the progress in therapy, 5-year survival rates remain unsatisfied. Chronic infection with the hepatitis B virus (HBV) or hepatitis C virus (HCV) has been epidemiologically associated with HCC and is the major etiology in the East Asian population. The detailed mechanism, especially the changes of DNA methylation and gene expression between the two types of virus-related HCC, and their contributions to the HCC development, metastasis, and recurrence remain largely unknown. Methods In this integrated analysis, we characterized genome-scale profiles of HBV and HCV infected HCC by comparing their gene expression pattern, methylation profiles, and copy number variations from the publicly accessible data of The Cancer Genome Atlas Program (TCGA). Results The HLA-A, STAT1, and OAS2 genes were highly enriched and up-regulated discovered in the HCV-infected HCC. Hypomethylation but not copy number variations might be the major factor for the up-regulation of these immune-related genes in HCV-infected HCC. Conclusions The results indicated the different epigenetic changes of HBV/HCV related hepatocarcinogenesis. The top up-regulated genes in HCV group were significantly clustered in the immune-related and defense response pathways. These findings will help us to understand the pathogenesis of HBV/HCV associated hepatocellular carcinoma.

Sign in / Sign up

Export Citation Format

Share Document