scholarly journals Deep Neural Network Analysis of Pathology Images With Integrated Molecular Data for Enhanced Glioma Classification and Grading

2021 ◽  
Vol 11 ◽  
Author(s):  
Linmin Pei ◽  
Karra A. Jones ◽  
Zeina A. Shboul ◽  
James Y. Chen ◽  
Khan M. Iftekharuddin
Keyword(s):  
Neural Network ◽  
Deep Neural Network ◽  
Tumor Classification ◽  
The Cancer Genome Atlas ◽  
World Health ◽  
Lower Grade ◽  
Cns Tumor ◽  
Molecular Features ◽  
First Time ◽  
Glioma Classification

Gliomas are primary brain tumors that originate from glial cells. Classification and grading of these tumors is critical to prognosis and treatment planning. The current criteria for glioma classification in central nervous system (CNS) was introduced by World Health Organization (WHO) in 2016. This criteria for glioma classification requires the integration of histology with genomics. In 2017, the Consortium to Inform Molecular and Practical Approaches to CNS Tumor Taxonomy (cIMPACT-NOW) was established to provide up-to-date recommendations for CNS tumor classification, which in turn the WHO is expected to adopt in its upcoming edition. In this work, we propose a novel glioma analytical method that, for the first time in the literature, integrates a cellularity feature derived from the digital analysis of brain histopathology images integrated with molecular features following the latest WHO criteria. We first propose a novel over-segmentation strategy for region-of-interest (ROI) selection in large histopathology whole slide images (WSIs). A Deep Neural Network (DNN)-based classification method then fuses molecular features with cellularity features to improve tumor classification performance. We evaluate the proposed method with 549 patient cases from The Cancer Genome Atlas (TCGA) dataset for evaluation. The cross validated classification accuracies are 93.81% for lower-grade glioma (LGG) and high-grade glioma (HGG) using a regular DNN, and 73.95% for LGG II and LGG III using a residual neural network (ResNet) DNN, respectively. Our experiments suggest that the type of deep learning has a significant impact on tumor subtype discrimination between LGG II vs. LGG III. These results outperform state-of-the-art methods in classifying LGG II vs. LGG III and offer competitive performance in distinguishing LGG vs. HGG in the literature. In addition, we also investigate molecular subtype classification using pathology images and cellularity information. Finally, for the first time in literature this work shows promise for cellularity quantification to predict brain tumor grading for LGGs with IDH mutations.

scholarly journals Significance of tumor protein p53 mutation in cellular process and drug selection in brain lower grade (WHO grades II and III) glioma

2020 ◽  
Vol 14 (12) ◽  
pp. 1139-1150
Author(s):  
Chang-feng Guo ◽  
Yugang Zhuang ◽  
Yuanzhuo Chen ◽  
Sheng Chen ◽  
Hu Peng ◽  
...  
Keyword(s):  
Differentially Expressed Genes ◽  
Tp53 Mutation ◽  
The Cancer Genome Atlas ◽  
Functional Enrichment ◽  
Differentially Expressed ◽  
World Health ◽  
Outcome Analysis ◽  
Lower Grade ◽  
Protein P53 ◽  
Tumor Protein P53

Aim: Tumor protein p53 ( TP53) mutant is one of the most frequently mutated genes in glioma. Results: The Cancer Genome Atlas data has shown that TP53 mutation is present in 49% of lower grade (World Health Organization [WHO] grades II and III) glioma patients. Data from The Genomics of Drug Sensitivity in Cancer database showed that three drugs: (5Z)-7-oxozeaenol, dabrafenib and nutlin-3a (−), have shown more resistance in patients with TP53 mutation. We identified 1100 differentially expressed genes. Functional enrichment analysis showed that the differentially expressed genes are mainly concentrated in the transport of ionic and cancer-related pathways. The top ten hub genes were identified and an outcome analysis revealed the most critical genes related to prognosis. Conclusion: Our results identified the key genes and pathways that might provide the basic proof to improve individualized treatment in patients with glioma.

Treatment of Adult Lower-Grade Glioma in the Era of Genomic Medicine

2016 ◽  
pp. 75-81 ◽  
Author(s):  
Susan M. Chang ◽  
Daniel P. Cahill ◽  
Kenneth D. Aldape ◽  
Minesh P. Mehta
Keyword(s):  
Molecular Genetic ◽  
Genomic Medicine ◽  
Genetic Alterations ◽  
World Health ◽  
Low Grade ◽  
Lower Grade ◽  
Cell Of Origin ◽  
Low Grade Gliomas ◽  
Molecular Features ◽  
Anaplastic Gliomas

By convention, gliomas are histopathologically classified into four grades by the World Health Organization (WHO) legacy criteria, in which increasing grade is associated with worse prognosis and grades also are subtyped by presumed cell of origin. This classification has prognostic value but is limited by wide variability of outcome within each grade, so the classification is rapidly undergoing dramatic re-evaluation in the context of a superior understanding of the biologic heterogeneity and molecular make-up of these tumors, such that we now recognize that some low-grade gliomas behave almost like malignant glioblastoma, whereas other anaplastic gliomas have outcomes comparable to favorable low-grade gliomas. This clinical spectrum is partly accounted for by the dispersion of several molecular genetic alterations inherent to clinical tumor behavior. These molecular biomarkers have become important not only as prognostic factors but also, more critically, as predictive markers to drive therapeutic decision making. Some of these, in the near future, will likely also serve as potential therapeutic targets. In this article, we summarize the key molecular features of clinical significance for WHO grades II and III gliomas and underscore how the therapeutic landscape is changing.

scholarly journals Development and evaluation of a deep neural network for histologic classification of renal cell carcinoma on biopsy and surgical resection slides

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Mengdan Zhu ◽  
Bing Ren ◽  
Ryland Richards ◽  
Matthew Suriawinata ◽  
Naofumi Tomita ◽  
...  
Keyword(s):  
Neural Network ◽  
Renal Cell Carcinoma ◽  
Cell Carcinoma ◽  
Surgical Resection ◽  
Renal Cell ◽  
Deep Neural Network ◽  
False Negative ◽  
The Cancer Genome Atlas ◽  
Diagnostic Model

AbstractRenal cell carcinoma (RCC) is the most common renal cancer in adults. The histopathologic classification of RCC is essential for diagnosis, prognosis, and management of patients. Reorganization and classification of complex histologic patterns of RCC on biopsy and surgical resection slides under a microscope remains a heavily specialized, error-prone, and time-consuming task for pathologists. In this study, we developed a deep neural network model that can accurately classify digitized surgical resection slides and biopsy slides into five related classes: clear cell RCC, papillary RCC, chromophobe RCC, renal oncocytoma, and normal. In addition to the whole-slide classification pipeline, we visualized the identified indicative regions and features on slides for classification by reprocessing patch-level classification results to ensure the explainability of our diagnostic model. We evaluated our model on independent test sets of 78 surgical resection whole slides and 79 biopsy slides from our tertiary medical institution, and 917 surgical resection slides from The Cancer Genome Atlas (TCGA) database. The average area under the curve (AUC) of our classifier on the internal resection slides, internal biopsy slides, and external TCGA slides is 0.98 (95% confidence interval (CI): 0.97–1.00), 0.98 (95% CI: 0.96–1.00) and 0.97 (95% CI: 0.96–0.98), respectively. Our results suggest that the high generalizability of our approach across different data sources and specimen types. More importantly, our model has the potential to assist pathologists by (1) automatically pre-screening slides to reduce false-negative cases, (2) highlighting regions of importance on digitized slides to accelerate diagnosis, and (3) providing objective and accurate diagnosis as the second opinion.

scholarly journals Predicting prognosis and IDH mutation status for patients with lower-grade gliomas using whole slide images

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Shuai Jiang ◽  
George J. Zanazzi ◽  
Saeed Hassanpour
Keyword(s):  
Performance Metrics ◽  
Small Sample Size ◽  
Small Sample ◽  
The Cancer Genome Atlas ◽  
World Health ◽  
Lower Grade ◽  
Who Grade ◽  
Mutation Status ◽  
Idh Mutations ◽  
Whole Slide Images

AbstractWe developed end-to-end deep learning models using whole slide images of adults diagnosed with diffusely infiltrating, World Health Organization (WHO) grade 2 gliomas to predict prognosis and the mutation status of a somatic biomarker, isocitrate dehydrogenase (IDH) 1/2. The models, which utilize ResNet-18 as a backbone, were developed and validated on 296 patients from The Cancer Genome Atlas (TCGA) database. To account for the small sample size, repeated random train/test splits were performed for hyperparameter tuning, and the out-of-sample predictions were pooled for evaluation. Our models achieved a concordance- (C-) index of 0.715 (95% CI: 0.569, 0.830) for predicting prognosis and an area under the curve (AUC) of 0.667 (0.532, 0.784) for predicting IDH mutations. When combined with additional clinical information, the performance metrics increased to 0.784 (95% CI: 0.655, 0.880) and 0.739 (95% CI: 0.613, 0.856), respectively. When evaluated on the WHO grade 3 gliomas from the TCGA dataset, which were not used for training, our models predicted survival with a C-index of 0.654 (95% CI: 0.537, 0.768) and IDH mutations with an AUC of 0.814 (95% CI: 0.721, 0.897). If validated in a prospective study, our method could potentially assist clinicians in managing and treating patients with diffusely infiltrating gliomas.

scholarly journals Extracting a Biologically Relevant Latent Space from Cancer Transcriptomes with Variational Autoencoders

2017 ◽  
Author(s):  
Gregory P. Way ◽  
Casey S. Greene
Keyword(s):  
Neural Network ◽  
Gene Expression ◽  
Deep Neural Network ◽  
Expression Profiles ◽  
Expression Patterns ◽  
The Cancer Genome Atlas ◽  
Neural Network Models ◽  
Neural Network Approach ◽  
Biologically Relevant ◽  
Latent Space

The Cancer Genome Atlas (TCGA) has profiled over 10,000 tumors across 33 different cancer-types for many genomic features, including gene expression levels. Gene expression measurements capture substantial information about the state of each tumor. Certain classes of deep neural network models are capable of learning a meaningful latent space. Such a latent space could be used to explore and generate hypothetical gene expression profiles under various types of molecular and genetic perturbation. For example, one might wish to use such a model to predict a tumor’s response to specific therapies or to characterize complex gene expression activations existing in differential proportions in different tumors. Variational autoencoders (VAEs) are a deep neural network approach capable of generating meaningful latent spaces for image and text data. In this work, we sought to determine the extent to which a VAE can be trained to model cancer gene expression, and whether or not such a VAE would capture biologically-relevant features. In the following report, we introduce a VAE trained on TCGA pan-cancer RNA-seq data, identify specific patterns in the VAE encoded features, and discuss potential merits of the approach. We name our method “Tybalt” after an instigative, cat-like character who sets a cascading chain of events in motion in Shakespeare’s “Romeo and Juliet”. From a systems biology perspective, Tybalt could one day aid in cancer stratification or predict specific activated expression patterns that would result from genetic changes or treatment effects.

scholarly journals Brain Tumor Classification Using Deep Neural Network

2020 ◽  
Vol 5 (5) ◽  
pp. 765-769
Author(s):  
Gökalp Çınarer ◽  
Bülent Gürsel Emiroğlu ◽  
Recep Sinan Arslan ◽  
Ahmet Haşim Yurttakal
Keyword(s):  
Neural Network ◽  
Brain Tumor ◽  
Deep Neural Network ◽  
Tumor Classification

scholarly journals High FREM2 Gene and Protein Expression Are Associated with Favorable Prognosis of IDH-WT Glioblastomas

Cancers ◽  
2019 ◽  
Vol 11 (8) ◽  
pp. 1060 ◽  
Author(s):  
Ivana Jovčevska ◽  
Alja Zottel ◽  
Neja Šamec ◽  
Jernej Mlakar ◽  
Maxim Sorokin ◽  
...  
Keyword(s):  
Protein Expression ◽  
The Cancer Genome Atlas ◽  
World Health ◽  
Low Grade ◽  
Lower Grade ◽  
Tissue Samples ◽  
Expression Levels ◽  
Favorable Prognosis ◽  
Gene And Protein Expression ◽  
Reference Samples

World Health Organization grade IV diffuse gliomas, known as glioblastomas, are the most common malignant brain tumors, and they show poor prognosis. Multimodal treatment of surgery followed by radiation and chemotherapy is not sufficient to increase patient survival, which is 12 to 18 months after diagnosis. Despite extensive research, patient life expectancy has not significantly improved over the last decade. Previously, we identified FREM2 and SPRY1 as genes with differential expression in glioblastoma cell lines compared to nonmalignant astrocytes. In addition, the FREM2 and SPRY1 proteins show specific localization on the surface of glioblastoma cells. In this study, we explored the roles of the FREM2 and SPRY1 genes and their proteins in glioblastoma pathology using human tissue samples. We used proteomic, transcriptomic, and bioinformatics approaches to detect changes at different molecular levels. We demonstrate increased FREM2 protein expression levels in glioblastomas compared to reference samples. At the transcriptomic level, both FREM2 and SPRY1 show increased expression in tissue samples of different glioma grades compared to nonmalignant brain tissue. To broaden our experimental findings, we analyzed The Cancer Genome Atlas glioblastoma patient datasets. We discovered higher FREM2 and SPRY1 gene expression levels in glioblastomas compared to lower grade gliomas and reference samples. In addition, we observed that low FREM2 expression was associated with progression of IDH-mutant low-grade glioma patients. Multivariate analysis showed positive association between FREM2 and favorable prognosis of IDH-wild type glioblastoma. We conclude that FREM2 has an important role in malignant progression of glioblastoma, and we suggest deeper analysis to determine its involvement in glioblastoma pathology.

scholarly journals Survival of diffuse astrocytic glioma, IDH1/2 wildtype, with molecular features of glioblastoma, WHO grade IV: a confirmation of the cIMPACT-NOW criteria

2019 ◽  
Vol 22 (4) ◽  
pp. 515-523 ◽  
Author(s):  
C Mircea S Tesileanu ◽  
Linda Dirven ◽  
Maarten M J Wijnenga ◽  
Johan A F Koekkoek ◽  
Arnaud J P E Vincent ◽  
...  
Keyword(s):  
Proportional Hazards ◽  
Proportional Hazards Model ◽  
Cox Proportional Hazards ◽  
Cox Proportional Hazards Model ◽  
World Health ◽  
Lower Grade ◽  
Isocitrate Dehydrogenase 1 ◽  
Who Grade ◽  
Astrocytic Glioma ◽  
Molecular Features

Abstract Background The Consortium to Inform Molecular and Practical Approaches to CNS Tumor Taxonomy (cIMPACT-NOW) has recommended that isocitrate dehydrogenase 1 and 2 wildtype (IDH1/2wt) diffuse lower-grade gliomas (LGGs) World Health Organization (WHO) grade II or III that present with (i) a telomerase reverse transcriptase promoter mutation (pTERTmt), and/or (ii) gain of chromosome 7 combined with loss of chromosome 10, and/or (iii) epidermal growth factor receptor (EGFR) amplification should be reclassified as diffuse astrocytic glioma, IDH1/2 wildtype, with molecular features of glioblastoma, WHO grade IV (IDH1/2wt astrocytomas WHO IV). This paper describes the overall survival (OS) of IDH1/2wt astrocytoma WHO IV patients, and more in detail patients with tumors with pTERTmt only. Methods In this retrospective multicenter study, we compared the OS of 71 IDH1/2wt astrocytomas WHO IV patients, with radiological characteristics of LGGs, with the OS of 197 IDH1/2wt glioblastoma patients. Moreover, we compared the OS of 22 pTERTmt only astrocytoma patients with the OS of the IDH1/2wt glioblastoma patients. Results Median OS was similar for IDH1/2wt astrocytoma WHO IV patients (23.8 mo) and IDH1/2wt glioblastoma patients (19.2 mo) (Cox proportional hazards model: hazard ratio [HR] 1.27, 95% CI: 0.85–1.88, P = 0.242). OS was also similar in patients with IDH1/2wt astrocytomas WHO IV, pTERTmt only, and IDH1/2wt glioblastomas (HR 1.15, 95% CI: 0.64–2.10, P = 0.641). Conclusions The presented data confirm the cIMPACT-NOW recommendation and we propose that IDH1/2wt astrocytomas WHO IV in the absence of other qualifying mutations should be classified as IDH1/2wt glioblastomas.

Sign in / Sign up

Export Citation Format

Share Document