Dr. David Gutman: Development and Validation of Radiology Descriptors in Gliomas
1 like888 views
This document discusses the development and validation of imaging descriptors for glioblastomas (GBM) to correlate imaging findings with genetic and clinical data. It highlights the use of datasets from The Cancer Genome Atlas (TCGA) and the Cancer Imaging Archive (TCIA) to analyze tumor characteristics and explore the relationships between imaging features and patient outcomes. The project aims to create a comprehensive set of image biomarkers and features that can aid in understanding tumor behavior and improving diagnostic processes.
Dr. David Gutman: Development and Validation of Radiology Descriptors in Gliomas
- 1. Development and Validation of Qualitative and Quantitative Descriptors in Gliomas David A Gutman MD PhD Department of Biomedical Informatics Emory University
- 2. Quick overview of Glioblastoma (GBM) Center for Comprehensive Informatics • Most common primary brain tumor in adults • Median survival 50 weeks • ISBTRC Goals: – To leverage rich datasets to understand the mechanisms of glioma progression through In Silico analysis – To manage, explore and share semantically complex data among researchers
- 3. The Cancer Genome Atlas (TCGA) • Characterize 500 tumors for each of a variety of cancers • Clinical records Center for Comprehensive Informatics • Genomics: gene, miRNA expression, copy number, sequence, DNA methylation • Imaging: pathology and radiology
- 4. TCGA and Imaging Data: Radiology and Pathology • The Cancer Imaging Archive (TCIA) now contains Center for Comprehensive Informatics radiology data on ~ 150 patients from the TCGA GBM data set • Pathology data is also available on ~ 200 patients • Our extended group’s goal is to “mine” radiology and pathology data for phenotypes that correlate with genetic and clinical characteristics of the patients • Dr. Cooper presented some of our work correlating pathology with genomics and outcomes • Parallel effort has been underway for radiology data sets
- 5. Overall question… • Do tumors that “look” different behave differently? Center for Comprehensive Informatics – e.g. different outcome – Different genetic profiles Problems… – Need for a standardized method to describe what the tumors look like…
- 6. Genetic signatures can define tumor subtypes Center for Comprehensive Informatics
- 7. Clustering identifies three morphological groups • Analyzed 200 million nuclei from 162 TCGA GBMs (462 slides) Center for Comprehensive Informatics • Named for functions of associated genes: Cell Cycle (CC), Chromatin Modification (CM), Protein Biosynthesis (PB) • Prognostically-significant (logrank p=4.5e-4)
- 8. Representative nuclei Center for Comprehensive Informatics Large, Small light nuclei, Intermediate hyperchromatic Eosinophilic cyoplasm L Cooper nuclei
- 9. How Does One Effectively Marry Imaging Findings of a Tumor to its Genomics? Center for Comprehensive Informatics X Genetic Microarray A Flanders
- 10. VASARI Feature Set • A set of 30 imaging characteristics to describe high Center for Comprehensive Informatics grade gliomas (GBM) using standardized vocabulary that is reproducible and understandable by neuroradiologists • Effort led by Adam Flanders and Carl Jaffe involving coordinating “reads” and feature set development by ~ 8 neuroradiologists
- 11. Defining a Rich Set of Qualitative and Quantitative Image Biomarkers • This has been a community-driven ontology development Center for Comprehensive Informatics project to create a comprehensive set of imaging observations for GBM – Collaboration with ASNR • Collaborators were asked to provide a list of clinical or literature observations that could be used to describe MRI features of GBM • Imaging features (26 features / 4 categories) – Location of lesion – Morphology of lesion margin (definition, thickness, enhancement, diffusion) – Morphology of lesion substance (enhancement, PS characteristics, focality/multicentricity, necrosis, cysts, midline invasion, cortical involvement, T1/FLAIR ratio) – Alterations in vicinity of lesion (edema, edema crossing midline, hemorrhage, pial invasion, ependymal invasion, satellites, deep WM invasion, calvarial remodeling)
- 12. F5 – Proportion Enhancing Center for Comprehensive Informatics Visually, when scanning through the entire tumor volume, what proportion of the entire tumor would you estimate is enhancing? (Assuming that the entire abnormality may be comprised of: (1) an enhancing component, (2) a non-enhancing component, (3) a necrotic component and (4) a edema component.)
- 13. F7 – Proportion Necrosis Center for Comprehensive Informatics Visually, when scanning through the entire tumor volume, what proportion of the tumor is estimated to represent necrosis? Necrosis is defined as a region within the tumor that does not enhance or shows markedly diminished enhancement, is high on T2W and proton density images, is low on T1W images, and has an irregular border). (Assuming that the entire abnormality may be comprised of: (1) an enhancing component, (2) a non-enhancing component, (3) a necrotic component and (4) a edema component.)
- 14. Capturing structured annotations and markups/AIM Data Service Center for Comprehensive Informatics
- 15. For validation, focused on semi-quantitative features • Compared various outcome and genomic measures Center for Comprehensive Informatics with these features • Also did comparisons between qualitative and quantitative volumetric measurements performed at MGH by Colen et. al using 3D slicer, and measurements done at Emory using the Velocity Platform
- 16. Correlating between quantitative and qualitative features: Man vs Machine Center for Comprehensive Informatics Results of univariate linear regression for agreement between VASARI measurements and measurements derived from quantitative volumetric analyses.
- 17. Agreement between qualitative and quantitative feature set Center for Comprehensive Informatics
- 18. Inter-rater agreement of relevant imaging features between radiologists scores according to VASARI standard Center for Comprehensive Informatics
- 19. 3d Slicer Volume Segmentation (R. Colen/MGH) Center for Comprehensive Informatics Visualization of quantitative volumetric segmentation methodology. Region corresponding to edema/tumor infiltration (blue) was segmented from FLAIR sequences whereas contrast enhancement (yellow) and necrosis (orange) have been segmented from T1 post contrast weighted images
- 20. Center for Comprehensive Informatics Machine vs Machine?
- 21. Center for Comprehensive Informatics Cleaning up the raw data from TCIA
- 22. Developed some tooling to help with image validation & QA Center for Comprehensive Informatics
- 23. Slicer Volumes vs Velocity Derived Volumes Center for Comprehensive Informatics
- 24. Center for Comprehensive Informatics Do image features predict outcome?
- 25. Combination of clinical and imaging features Center for Comprehensive Informatics
- 26. Are imaging features equally distributed across Verhaak classification subtypes? Center for Comprehensive Informatics
- 27. Correlation of Volumetric Data with Outcome Center for Comprehensive Informatics
- 28. Future Work • Working on extracting features from volumetric Center for Comprehensive Informatics images and doing pathway analysis • Also Rajan Jain (TJU) and Scott Hwang (Emory) have begun doing feature extraction/markups of perfusion and DTI data • Continue to collect imaging data from TCGA GBM contributors (as we track them down) • Continue to revise/simplify feature set • Consider extending feature set to lower grade cases
- 29. In Silico Brain Tumor Research Center Team • Emory University • Henry Ford Hospital Center for Comprehensive Informatics – Lee Cooper – Tom Mikkelsen – Joel Saltz – Lisa Scarpace – Daniel Brat – Carlos Moreno • Thomas Jefferson University – Chad Holder – Adam Flanders – Scott Hwang – Doris Gao • SAIC Frederick – William Dunn – John Freymann – Tarun Aurora – Justin Kirby • NCI – Eric Huang – Carl Jaffe • MGH – Rivka Colen