Elucidating Cancer Subtypes by Using the Relationship between DNA Methylation and Gene Expression.

Advancements in the field of next generation sequencing (NGS) have generated vast amounts of data for the same set of subjects. The challenge that arises is how to combine and reconcile results from different omics studies, such as epigenome and transcriptome, to improve the classification of disease subtypes. In this study, we introduce sCClust (sparse canonical correlation analysis with clustering), a technique to combine high-dimensional omics data using sparse canonical correlation analysis (sCCA), such that the correlation between datasets is maximized. This stage is followed by clustering the integrated data in a lower-dimensional space. We apply sCClust to gene expression and DNA methylation data for three cancer genomics datasets from the Cancer Genome Atlas (TCGA) to distinguish between underlying subtypes. We evaluate the identified subtypes using Kaplan-Meier plots and hazard ratio analysis on the three types of cancer-GBM (glioblastoma multiform), lung cancer and colon cancer. Comparison with subtypes identified by both single- and multi-omics studies implies improved clinical association. We also perform pathway over-representation analysis in order to identify up-regulated and down-regulated genes as tentative drug targets. The main goal of the paper is twofold: the integration of epigenomic and transcriptomic datasets followed by elucidating subtypes in the latent space. The significance of this study lies in the enhanced categorization of cancer data, which is crucial to precision medicine.

Comprehensive understanding of glioblastoma molecular phenotypes: classification, characteristics, and transition.

Among central nervous system-associated malignancies, glioblastoma (GBM) is the most common and has the highest mortality rate. The high heterogeneity of GBM cell types and the complex tumor microenvironment frequently lead to tumor recurrence and sudden relapse in patients treated with temozolomide. In precision medicine, research on GBM treatment is increasingly focusing on molecular subtyping to precisely characterize the cellular and molecular heterogeneity, as well as the refractory nature of GBM toward therapy. Deep understanding of the different molecular expression patterns of GBM subtypes is critical. Researchers have recently proposed tetra fractional or tripartite methods for detecting GBM molecular subtypes. The various molecular subtypes of GBM show significant differences in gene expression patterns and biological behaviors. These subtypes also exhibit high plasticity in their regulatory pathways, oncogene expression, tumor microenvironment alterations, and differential responses to standard therapy. Herein, we summarize the current molecular typing scheme of GBM and the major molecular/genetic characteristics of each subtype. Furthermore, we review the mesenchymal transition mechanisms of GBM under various regulators.

Two-headed UNetEfficientNets for parallel execution of segmentation and classification of brain tumors: incorporating postprocessing techniques with connected component labelling.

PURPOSE: The purpose of this study is to develop accurate and automated detection and segmentation methods for brain tumors, given their significant fatality rates, with aggressive malignant tumors like Glioblastoma Multiforme (GBM) having a five-year survival rate as low as 5 to 10%. This underscores the urgent need to improve diagnosis and treatment outcomes through innovative approaches in medical imaging and deep learning techniques. METHODS: In this work, we propose a novel approach utilizing the two-headed UNetEfficientNets model for simultaneous segmentation and classification of brain tumors from Magnetic Resonance Imaging (MRI) images. The model combines the strengths of EfficientNets and a modified two-headed Unet model. We utilized a publicly available dataset consisting of 3064 brain MR images classified into three tumor classes: Meningioma, Glioma, and Pituitary. To enhance the training process, we performed 12 types of data augmentation on the training dataset. We evaluated the methodology using six deep learning models, ranging from UNetEfficientNet-B0 to UNetEfficientNet-B5, optimizing the segmentation and classification heads using binary cross entropy (BCE) loss with Dice and BCE with focal loss, respectively. Post-processing techniques such as connected component labeling (CCL) and ensemble models were applied to improve segmentation outcomes. RESULTS: The proposed UNetEfficientNet-B4 model achieved outstanding results, with an accuracy of 99.4% after postprocessing. Additionally, it obtained high scores for DICE (94.03%), precision (98.67%), and recall (99.00%) after post-processing. The ensemble technique further improved segmentation performance, with a global DICE score of 95.70% and Jaccard index of 91.20%. CONCLUSION: Our study demonstrates the high efficiency and accuracy of the proposed UNetEfficientNet-B4 model in the automatic and parallel detection and segmentation of brain tumors from MRI images. This approach holds promise for improving diagnosis and treatment planning for patients with brain tumors, potentially leading to better outcomes and prognosis.

A deep convolutional neural network for the automatic segmentation of glioblastoma brain tumor: Joint spatial pyramid module and attention mechanism network.

This study proposes a deep convolutional neural network for the automatic segmentation of glioblastoma brain tumors, aiming sat replacing the manual segmentation method that is both time-consuming and labor-intensive. There are many challenges for automatic segmentation to finely segment sub-regions from multi-sequence magnetic resonance images because of the complexity and variability of glioblastomas, such as the loss of boundary information, misclassified regions, and subregion size. To overcome these challenges, this study introduces a spatial pyramid module and attention mechanism to the automatic segmentation algorithm, which focuses on multi-scale spatial details and context information. The proposed method has been tested in the public benchmarks BraTS 2018, BraTS 2019, BraTS 2020 and BraTS 2021 datasets. The Dice score on the enhanced tumor, whole tumor, and tumor core were respectively 79.90 %, 89.63 %, and 85.89 % on the BraTS 2018 dataset, respectively 77.14 %, 89.58 %, and 83.33 % on the BraTS 2019 dataset, and respectively 77.80 %, 90.04 %, and 83.18 % on the BraTS 2020 dataset, and respectively 83.48 %, 90.70 %, and 88.94 % on the BraTS 2021 dataset offering performance on par with that of state-of-the-art methods with only 1.90 M parameters. In addition, our approach significantly reduced the requirements for experimental equipment, and the average time taken to segment one case was only 1.48 s; these two benefits rendered the proposed network intensely competitive for clinical practice.

Analysis of the Prognosis of High-Grade gliomas in the View of New Immunohistochemistry Markers and 2016 WHO Classification.

AIM: To evaluate isocitrate dehydrogenase (IDH) mutation status and Ki67 percentages of tumors that were treated in our institution to determine whether these markers affected the initial diagnosis and survival rates. MATERIAL AND METHODS: High-grade glioma patients, who were operated in our department between 2013 and 2018, were enrolled in the study and retrospectively reviewed. New immunohistochemistry staining studies were conducted and survival analyses were performed. RESULTS: Of 135 patients and 136 tumors, 117 were glioblastoma multiforme (GBM), 8 were grade III-IV glioma, 4 were anaplastic astrocytoma and 7 were anaplastic oligodendroglioma. One patient had two different lesions, which were GBM and anaplastic astrocytoma respectively. Mean age was 55 (7-85) years, and 88 (65%) were male and 47 (35%) were female. The most common complaint was motor deficit (56%). Fourteen patients underwent reoperation due to recurrent disease. Tumors were most commonly found in the frontal lobe (53, 39%). Magnetic resonance imaging (MRI) features showed that existence of necrosis is strongly related to GBM (p < 0.01). Approximately 126 patients were found to be IDH-wildtype, which changed 6 patients? diagnosis to GBM, IDH wildtype from grade III-IV glioma. Five patients, who were diagnosed with anaplastic astrocytoma and anaplastic oligodendroglioma initially were found to be IDH wildtype. IDH mutation status, extend of resection, and age were found to affect survival. CONCLUSION: IDH mutation status is important in classifying high-grade gliomas, as well as its effects on prognosis. This mutation is related to several radiological features of tumors. Extent of resection and patient age are also profoundly affect survival. Detailing the diagnosis with molecular features will help physicians to shape targeted adjuvant therapies, which would better outcomes.

Local detection of microvessels in IDH-wildtype glioblastoma using relative cerebral blood volume: an imaging marker useful for astrocytoma grade 4 classification.

BACKGROUND: The microvessels area (MVA), derived from microvascular proliferation, is a biomarker useful for high-grade glioma classification. Nevertheless, its measurement is costly, labor-intense, and invasive. Finding radiologic correlations with MVA could provide a complementary non-invasive approach without an extra cost and labor intensity and from the first stage. This study aims to correlate imaging markers, such as relative cerebral blood volume (rCBV), and local MVA in IDH-wildtype glioblastoma, and to propose this imaging marker as useful for astrocytoma grade 4 classification. METHODS: Data from 73 tissue blocks belonging to 17 IDH-wildtype glioblastomas and 7 blocks from 2 IDH-mutant astrocytomas were compiled from the Ivy GAP database. MRI processing and rCBV quantification were carried out using ONCOhabitats methodology. Histologic and MRI co-registration was done manually with experts' supervision, achieving an accuracy of 88.8% of overlay. Spearman's correlation was used to analyze the association between rCBV and microvessel area. Mann-Whitney test was used to study differences of rCBV between blocks with presence or absence of microvessels in IDH-wildtype glioblastoma, as well as to find differences with IDH-mutant astrocytoma samples. RESULTS: Significant positive correlations were found between rCBV and microvessel area in the IDH-wildtype blocks (p < 0.001), as well as significant differences in rCBV were found between blocks with microvascular proliferation and blocks without it (p < 0.0001). In addition, significant differences in rCBV were found between IDH-wildtype glioblastoma and IDH-mutant astrocytoma samples, being 2-2.5 times higher rCBV values in IDH-wildtype glioblastoma samples. CONCLUSIONS: The proposed rCBV marker, calculated from diagnostic MRIs, can detect in IDH-wildtype glioblastoma those regions with microvessels from those without it, and it is significantly correlated with local microvessels area. In addition, the proposed rCBV marker can differentiate the IDH mutation status, providing a complementary non-invasive method for high-grade glioma classification.

Deep learning identified glioblastoma subtypes based on internal genomic expression ranks.

BACKGROUND: Glioblastoma (GBM) can be divided into subtypes according to their genomic features, including Proneural (PN), Neural (NE), Classical (CL) and Mesenchymal (ME). However, it is a difficult task to unify various genomic expression profiles which were standardized with various procedures from different studies and to manually classify a given GBM sample into a subtype. METHODS: An algorithm was developed to unify the genomic profiles of GBM samples into a standardized normal distribution (SND), based on their internal expression ranks. Deep neural networks (DNN) and convolutional DNN (CDNN) models were trained on original and SND data. In addition, expanded SND data by combining various The Cancer Genome Atlas (TCGA) datasets were used to improve the robustness and generalization capacity of the CDNN models. RESULTS: The SND data kept unimodal distribution similar to their original data, and also kept the internal expression ranks of all genes for each sample. CDNN models trained on the SND data showed significantly higher accuracy compared to DNN and CDNN models trained on primary expression data. Interestingly, the CDNN models classified the NE subtype with the lowest accuracy in the GBM datasets, expanded datasets and in IDH wide type GBMs, consistent with the recent studies that NE subtype should be excluded. Furthermore, the CDNN models also recognized independent GBM datasets, even with small set of genomic expressions. CONCLUSIONS: The GBM expression profiles can be transformed into unified SND data, which can be used to train CDNN models with high accuracy and generalization capacity. These models suggested NE subtype may be not compatible with the 4 subtypes classification system.

The inverse paradigm and the ancestral cell of IDH-wildtype glioblastoma.

Rethinking IDH-wildtype glioblastoma through its unique features can help researchers find innovative and effective treatments. It is currently emerging that, after decades of therapeutic impasse, some traditional concepts regarding IDH-wildtype glioblastoma need to be supplemented and updated to overcome therapeutic resistance. Indeed, multiple clinical aspects and recent indirect and direct experimental data are providing evidence that the supratentorial brain parenchyma becomes entirely and quiescently micro-infiltrated long before primary tumor bulk growth. Furthermore, they are indicating that the known micro-infiltration that occurs during the IDH-wildtype glioblastoma growth and evolution is not at the origin of distant relapses. It follows that the ubiquitous supratentorial brain parenchyma micro-infiltration as a source for the development of widespread distant recurrences is actually due to the silent stage that precedes tumor growth rather than to the latter. All this implies that, in addition to the heterogeneity of the primary bulk, there is a second crucial cause of therapeutic resistance that has never hitherto been identified and challenged. In this regard, the ancestral founder cancer stem cell (CSC) appears as the key cell that can link the two causes of resistance.

MFmap: A semi-supervised generative model matching cell lines to tumours and cancer subtypes.

Translating in vitro results from experiments with cancer cell lines to clinical applications requires the selection of appropriate cell line models. Here we present MFmap (model fidelity map), a machine learning model to simultaneously predict the cancer subtype of a cell line and its similarity to an individual tumour sample. The MFmap is a semi-supervised generative model, which compresses high dimensional gene expression, copy number variation and mutation data into cancer subtype informed low dimensional latent representations. The accuracy (test set F1 score >90%) of the MFmap subtype prediction is validated in ten different cancer datasets. We use breast cancer and glioblastoma cohorts as examples to show how subtype specific drug sensitivity can be translated to individual tumour samples. The low dimensional latent representations extracted by MFmap explain known and novel subtype specific features and enable the analysis of cell-state transformations between different subtypes. From a methodological perspective, we report that MFmap is a semi-supervised method which simultaneously achieves good generative and predictive performance and thus opens opportunities in other areas of computational biology.

Protein kinase Cι and SRC signaling define reciprocally related subgroups of glioblastoma with distinct therapeutic vulnerabilities.

We report that atypical protein kinase Cι (PKCι) is an oncogenic driver of glioblastoma (GBM). Deletion or inhibition of PKCι significantly impairs tumor growth and prolongs survival in murine GBM models. GBM cells expressing elevated PKCι signaling are sensitive to PKCι inhibitors, whereas those expressing low PKCι signaling exhibit active SRC signaling and sensitivity to SRC inhibitors. Resistance to the PKCι inhibitor auranofin is associated with activated SRC signaling and response to a SRC inhibitor, whereas resistance to a SRC inhibitor is associated with activated PKCι signaling and sensitivity to auranofin. Interestingly, PKCι- and SRC-dependent cells often co-exist in individual GBM tumors, and treatment of GBM-bearing mice with combined auranofin and SRC inhibitor prolongs survival beyond either drug alone. Thus, we identify PKCι and SRC signaling as distinct therapeutic vulnerabilities that are directly translatable into an improved treatment for GBM.

Diffuse midline gliomas, H3 K27M-mutant are associated with less peritumoral edema and contrast enhancement in comparison to glioblastomas, H3 K27M-wildtype of midline structures.

PURPOSE: The entity 'diffuse midline glioma, H3 K27M-mutant (DMG)' was introduced in the revised 4th edition of the 2016 WHO classification of brain tumors. However, there are only a few reports on magnetic resonance imaging (MRI) of these tumors. Thus, we conducted a retrospective survey focused on MRI features of DMG compared to midline glioblastomas H3 K27M-wildtype (mGBM-H3wt). METHODS: We identified 24 DMG cases and 19 mGBM-H3wt patients as controls. After being retrospectively evaluated for microscopic evidence of microvascular proliferations (MVP) and tumor necrosis by two experienced neuropathologists to identify the defining histological criteria of mGBM-H3wt, the samples were further analyzed by two experienced readers regarding imaging features such as shape, peritumoral edema and contrast enhancement. RESULTS: The DMG were found in the thalamus in 37.5% of cases (controls 63%), in the brainstem in 50% (vs. 32%) and spinal cord in 12.5% (vs. 5%). In MRI and considering MVP, DMG were found to be by far less likely to develop peritumoral edema (OR: 0.13; 95%-CL: 0.02-0.62) (p = 0.010). They, similarly, were associated with a significantly lower probability of developing strong contrast enhancement compared to mGBM-H3wt (OR: 0.10; 95%-CL: 0.02-0.47) (P = 0.003). CONCLUSION: Despite having highly variable imaging features, DMG exhibited markedly less edema and lower contrast enhancement in MRI compared to mGBM-H3wt. Of these features, the enhancement level was associated with evidence of MVP.

The impact of the molecular classification of glioblastoma on the interpretation of therapeutic clinical trial results.

In 2016, the World Health Organization (WHO) released the most recent update to the classification of central nervous system tumors. This update has led to the reshaping of tumor identification and subsequently changed current understanding of treatment options for patients. Moreover, the restructuring of the classification of central nervous system tumors to include molecular markers has led to the need to re-evaluate how to interpret pivotal trials. These trials originally enrolled patients purely based upon histologic diagnoses without the use of adjunctive, and frequently diagnostic molecular testing. With this new paradigm also comes the need to assess how one should incorporate molecular markers into current trials as well as shape future trials. First, we will discuss updates on the molecular classification of glioblastoma (GBM) (and its histologic mimics). This will be followed by a review of key pivotal trials which have defined our standard of care for glioblastoma within the context of molecular classification of their study populations. This will be followed by preliminary results of ongoing phase 3 cooperative group trials for high-grade gliomas that were initiated prior to routine molecular classification of tumors and how one could interpret these results in light of advances in molecular classification. Finally, we will end with suggestions for future clinical trial design with a focus on enrollment based upon molecular diagnostics.

The making of the glioblastoma classification.

Classification of cancer should lead to informative patients' stratification and selective therapeutic vulnerabilities. A pathway-based classification of glioblastoma uncovered a mitochondrial subtype with a unique sensitivity to inhibitors of oxidative phosphorylation. Precision targeting of cancer metabolism could provide therapeutic opportunities to a lethal neoplasm and be translated to other tumour types.

Identifying Cancer Subtypes Using a Residual Graph Convolution Model on a Sample Similarity Network.

Cancer subtype classification helps us to understand the pathogenesis of cancer and develop new cancer drugs, treatment from which patients would benefit most. Most previous studies detect cancer subtypes by extracting features from individual samples, ignoring their associations with others. We believe that the interactions of cancer samples can help identify cancer subtypes. This work proposes a cancer subtype classification method based on a residual graph convolutional network and a sample similarity network. First, we constructed a sample similarity network regarding cancer gene co-expression patterns. Then, the gene expression profiles of cancer samples as initial features and the sample similarity network were passed into a two-layer graph convolutional network (GCN) model. We introduced the initial features to the GCN model to avoid over-smoothing during the training process. Finally, the classification of cancer subtypes was obtained through a softmax activation function. Our model was applied to breast invasive carcinoma (BRCA), glioblastoma multiforme (GBM) and lung cancer (LUNG) datasets. The accuracy values of our model reached 82.58%, 85.13% and 79.18% for BRCA, GBM and LUNG, respectively, which outperformed the existing methods. The survival analysis of our results proves the significant clinical features of the cancer subtypes identified by our model. Moreover, we can leverage our model to detect the essential genes enriched in gene ontology (GO) terms and the biological pathways related to a cancer subtype.

Molecular Classification of Gliomas is Associated with Seizure Control: A Retrospective Analysis.

Classically, histologic grading of gliomas has been used to predict seizure association, with low-grade gliomas associated with an increased incidence of seizures compared to high-grade gliomas. In 2016, WHO reclassified gliomas based on histology and molecular characteristics. We sought to determine whether molecular classification of gliomas is associated with preoperative seizure presentation and/or post-operative seizure control across multiple glioma subtypes. All gliomas operated at our institution from 2007 to 2017 were identified based on ICD 9 and 10 billing codes and were retrospectively assessed for molecular classification of the IDH1 mutation, and 1p/19q codeletion. Logistic regression models were performed to assess associations of seizures at presentation as well as post-operative seizures with IDH status and the new WHO integrated classification. Our study included 376 patients: 82 IDH mutant and 294 IDH wildtype. The presence of IDH mutation was associated with seizures at presentation [OR 3.135 (1.818-5.404), p < 0.001]. IDH-mutant glioblastomas presented with seizures less often than other IDH-mutant glioma subtypes grade II and III [OR 0.104 (0.032-0.340), p < 0.001]. IDH-mutant tumors were associated with worse post-operative seizure outcomes, demonstrated by Engel Class [OR 2.666 (1.592-4.464), p < 0.001]. IDH mutation in gliomas is associated with an increased risk of seizure development and worse post-operative seizure control, in all grades except for GBM.

Normal tissue content impact on the GBM molecular classification.

Molecular classification of glioblastoma has enabled a deeper understanding of the disease. The four-subtype model (including Proneural, Classical, Mesenchymal and Neural) has been replaced by a model that discards the Neural subtype, found to be associated with samples with a high content of normal tissue. These samples can be misclassified preventing biological and clinical insights into the different tumor subtypes from coming to light. In this work, we present a model that tackles both the molecular classification of samples and discrimination of those with a high content of normal cells. We performed a transcriptomic in silico analysis on glioblastoma (GBM) samples (n = 810) and tested different criteria to optimize the number of genes needed for molecular classification. We used gene expression of normal brain samples (n = 555) to design an additional gene signature to detect samples with a high normal tissue content. Microdissection samples of different structures within GBM (n = 122) have been used to validate the final model. Finally, the model was tested in a cohort of 43 patients and confirmed by histology. Based on the expression of 20 genes, our model is able to discriminate samples with a high content of normal tissue and to classify the remaining ones. We have shown that taking into consideration normal cells can prevent errors in the classification and the subsequent misinterpretation of the results. Moreover, considering only samples with a low content of normal cells, we found an association between the complexity of the samples and survival for the three molecular subtypes.

Glioblastoma epigenome profiling identifies SOX10 as a master regulator of molecular tumour subtype.

Glioblastoma frequently exhibits therapy-associated subtype transitions to mesenchymal phenotypes with adverse prognosis. Here, we perform multi-omic profiling of 60 glioblastoma primary tumours and use orthogonal analysis of chromatin and RNA-derived gene regulatory networks to identify 38 subtype master regulators, whose cell population-specific activities we further map in published single-cell RNA sequencing data. These analyses identify the oligodendrocyte precursor marker and chromatin modifier SOX10 as a master regulator in RTK I-subtype tumours. In vitro functional studies demonstrate that SOX10 loss causes a subtype switch analogous to the proneural-mesenchymal transition observed in patients at the transcriptomic, epigenetic and phenotypic levels. SOX10 repression in an in vivo syngeneic graft glioblastoma mouse model results in increased tumour invasion, immune cell infiltration and significantly reduced survival, reminiscent of progressive human glioblastoma. These results identify SOX10 as a bona fide master regulator of the RTK I subtype, with both tumour cell-intrinsic and microenvironmental effects.

miRNA signature in glioblastoma: Potential biomarkers and therapeutic targets.

MicroRNAs (miRNAs) are transcripts with sizes of about 22 nucleotides, which are produced through a multistep process in the nucleus and cytoplasm. These transcripts modulate the expression of their target genes through binding with certain target regions, particularly 3' suntranslated regions. They are involved in the pathogenesis of several kinds of cancers, such as glioblastoma. Several miRNAs, including miR-10b, miR-21, miR-17-92-cluster, and miR-93, have been up-regulated in glioblastoma cell lines and clinical samples. On the other hand, expression of miR-7, miR-29b, miR-32, miR-34, miR-181 family members, and a number of other miRNAs have been decreased in this type of cancer. In the current review, we explain the role of miRNAs in the pathogenesis of glioblastoma through providing a summary of studies that reported dysregulation of these epigenetic effectors in this kind of brain cancer.

Glioblastoma TCGA Mesenchymal and IGS 23 Tumors are Identifiable by IHC and have an Immune-phenotype Indicating a Potential Benefit from Immunotherapy.

PURPOSE: Molecular subtype classifications in glioblastoma may detect therapy sensitivities. IHC would potentially allow the identification of molecular subtypes in routine clinical practice. EXPERIMENTAL DESIGN: Formalin-fixed, paraffin-embedded tumor samples of 124 uniformly treated, newly diagnosed patients with glioblastoma were submitted to RNA sequencing, IHC, and immune-phenotyping to identify differences in molecular subtypes associated with treatment sensitivities. RESULTS: We detected high molecular and IHC overlapping of the The Cancer Genome Atlas (TCGA) mesenchymal subtype with instrinsic glioma subtypes (IGS) cluster 23 and of the TCGA classical subtype with IGS cluster 18. IHC patterns, gene fusion profiles, and immune-phenotypes varied across subtypes. IHC revealed that the TCGA classical subtype was identified by high expression of EGFR and low expression of PTEN, while the mesenchymal subtype was identified by low expression of SOX2 and high expression of two antibodies, SHC1 and TCIRG1, selected on the basis of RNA differential transcriptomic expression. The proneural subtype was identified by frequent positive IDH1 expression and high Olig2 and Ki67 expression. Immune-phenotyping showed that mesenchymal and IGS 23 tumors exhibited a higher positive effector cell score, a higher negative suppressor cell score, and lower levels of immune checkpoint molecules. The cell-type deconvolution analysis revealed that these tumors are highly enriched in M2 macrophages, resting memory CD4+ T cells, and activated dendritic cells, indicating that they may be ideal candidates for immunotherapy, especially with anti-M2 and/or dendritic cell vaccination. CONCLUSIONS: There is a subset of tumors, frequently classified as mesenchymal or IGS cluster 23, that may be identified with IHC and could well be optimal candidates for immunotherapy.