Abstract:Glioblastoma is a highly aggressive form of brain cancer characterized by rapid progression and poor prognosis. Despite advances in treatment, the underlying genetic mechanisms driving this aggressiveness remain poorly understood. In this study, we employed multimodal deep learning approaches to investigate glioblastoma heterogeneity using joint image/RNA-seq analysis. Our results reveal novel genes associated with glioblastoma. By leveraging a combination of whole-slide images and RNA-seq, as well as introducing novel methods to encode RNA-seq data, we identified specific genetic profiles that may explain different patterns of glioblastoma progression. These findings provide new insights into the genetic mechanisms underlying glioblastoma heterogeneity and highlight potential targets for therapeutic intervention.

What problem does this paper attempt to address?

The paper attempts to address the issue of uncovering the genetic basis and heterogeneity of Glioblastoma (GB). Despite advances in treatment, the highly invasive nature and poor prognosis of glioblastoma remain poorly understood. Specifically, this study explores the heterogeneity of glioblastoma through a multimodal deep learning approach, combining Whole Slide Images (WSI) and RNA sequencing data (RNA-seq). ### Main Objectives: 1. **Identify potential genetic markers associated with glioblastoma invasiveness**: Through a multimodal deep learning approach, researchers aim to find new genes and genetic features that may explain the progression patterns of different glioblastomas. 2. **Develop new RNA-seq encoding techniques**: Researchers propose a pathway-based RNA-seq encoding method to improve the performance of deep learning models. 3. **Combine WSI and RNA-seq data**: Researchers develop a new method to integrate WSI and RNA-seq data for deep learning models to better understand the genetic heterogeneity of glioblastoma. ### Background: - **Glioblastoma**: A highly invasive brain tumor characterized by rapid progression and poor prognosis. - **Existing treatments**: Include surgery, chemoradiotherapy, and recently Tumor Treating Fields, but the median overall survival is only 20.9 months, with a progression-free survival of about 7 months. - **Tumor Stem Cells (GSCs)**: Believed to be the main driver of glioblastoma's biological heterogeneity, potentially leading to high recurrence rates. ### Methods: - **Dataset**: Utilized public data from The Cancer Genome Atlas (TCGA) and WSI and RNA-seq data from the STEMRI trial. - **RNA-seq encoding**: Based on Protein-to-Protein Interaction (PPI) maps, genes are clustered and then encoded. - **WSI encoding**: Images are segmented into non-overlapping patches and encoded using contrastive learning methods. - **Multimodal training**: Combined RNA-seq and WSI data, using a multimodal contrastive loss function to train the model. ### Results: - **Genetic markers**: Identified known and new genes associated with glioblastoma invasiveness. - **Encoding techniques**: The proposed pathway-based RNA-seq encoding method showed better performance. - **Multimodal model**: Successfully integrated WSI and RNA-seq data, enhancing the understanding of glioblastoma's genetic heterogeneity. ### Significance: - **Personalized medicine**: These findings provide new targets for developing personalized treatment strategies for glioblastoma patients. - **Continued research**: Emphasizes the importance of applying AI algorithms in cancer biology, providing direction for further research. In summary, this paper reveals the genetic basis and heterogeneity of glioblastoma through a multimodal deep learning approach, combining image and gene expression data, offering new insights for future treatments.

Uncovering the Genetic Basis of Glioblastoma Heterogeneity through Multimodal Analysis of Whole Slide Images and RNA Sequencing Data

Spatial transcriptomics inferred from pathology whole-slide images links tumor heterogeneity to survival in breast and lung cancer

Integrating imaging and genomic data for the discovery of distinct glioblastoma subtypes: a joint learning approach

Unveiling the Dynamics behind Glioblastoma Multiforme Single-Cell Data Heterogeneity

Single-cell Sequencing Data Analysis Reveals Heterogeneity of Glioblastoma Multiforme

Metabolic Insight into Glioma Heterogeneity: Mapping Whole Exome Sequencing to In Vivo Imaging with Stereotactic Localization and Deep Learning

Nimg-60. Subtyping Glioblastoma From Joint Clustering Of Imaging And Genomic Data

Epco-03. Metabolic Insight Into Glioma Heterogeneity: Mapping Whole Exome Sequencing Toin Vivo Imaging With Stereotactic Localization And Deep Learning

Spatial and molecular exploration of glioblastoma heterogeneity

Exploring glioma heterogeneity through omics networks: from gene network discovery to causal insights and patient stratification

Invariant Delineation of Nuclear Architecture in Glioblastoma Multiforme for Clinical and Molecular Association

Genomic Landscapes by Multiregion Sequencing Combined with Circulation Tumor DNA Detection Contribute to Molecular Diagnosis in Glioblastomas

Identifying cancer subtypes in glioblastoma by combining genomic, transcriptomic and epigenomic data

NIMG-66. RADIOGENOMICS-BASED HETEROGENEITY ANALYSIS OF GLIOBLASTOMA REVEALS MR IMAGING PHENOTYPICAL CHARACTERIZATION OF EGFR MUTATIONS

Tumor Heterogeneity in Gastrointestinal Cancer Based on Multimodal Data Analysis

An Intergraded Bioinformatics Approach for Identifying Novel Gene Networks in Glioblastoma

NIMG-22. AN AI-BASED COORDINATE SYSTEM ELUCIDATES RADIOGENOMIC HETEROGENEITY OF GLIOBLASTOMA VIA DEEP LEARNING AND MANIFOLD EMBEDDINGS

Highly Multiplexed Spatially Resolved Proteomic and Transcriptional Profiling of the Glioblastoma Microenvironment Using Archived Formalin-Fixed Paraffin-Embedded Specimens

Exploring Brain Tumor Heterogeneity for Survival Time Prediction

Unveiling divergent treatment prognoses in IDHwt-GBM subtypes through multiomics clustering: a swift dual MRI-mRNA model for precise subtype prediction

RADIOHEAD: Radiogenomic Analysis Incorporating Tumor Heterogeneity in Imaging Through Densities