ISSN# 1545-4428 | Published date: 19 April, 2024
You must be logged in to view entire program, abstracts, and syllabi
At-A-Glance Session Detail
   
Radiomics & Imaging Biomarkers in Brain Tumors
Oral
Neuro
Monday, 06 May 2024
Nicoll 1
08:15 -  10:15
Moderators: Antonella Castellano & Manabu Kinoshita
Session Number: O-34
CME Credit

08:150018.
Automatic Infiltration Risk Prior Generation with Modified Triplet Loss for Pre-Operative Glioblastoma Infiltration Prediction
Walter Zhao1, Sree Gongala2, Eunate Alzaga Goñi1, Xiaofeng Wang3, Shengwen Deng2, Charit Tippareddy2, Hamed Akbari4, Anahita Fathi Kazerooni5, Christos Davatzikos6, Marta Couce7, Andrew E. Sloan8, Chaitra Badve2, and Dan Ma1
1Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, 2Department of Radiology, University Hospitals Cleveland Medical Center, Cleveland, OH, United States, 3Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland, OH, United States, 4Department of Bioengineering, Santa Clara University, Santa Clara, OH, United States, 5Center for Data Driven Discovery in Biomedicine, Children's Hospital of Pennsylvania, Philadelphia, OH, United States, 6Center for Biomedical Image Computing and Analytics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, OH, United States, 7Department of Pathology, University Hospitals Cleveland Medical Center, Cleveland, OH, United States, 8Piedmont Physicians Neurosurgery Atlanta, Piedmont Healthcare, Atlanta, GA, United States

Keywords: Tumors (Pre-Treatment), Tumor, Glioblastoma

Motivation: Pre-operative glioblastoma (GBM) infiltration prediction models rely on manual infiltration risk (IR) prior segmentation which is tedious, requires expert input, and is highly variable. 

Goal(s): Automation is needed for fast segmentation. A data-driven method would account for GBM heterogeneity and be independent of specific MRI input for applicability to clinical protocols.

Approach: IR priors are grown using modified triplet loss with inter-prior and intra-prior terms to ensure priors are distinct from each other and maintain similarity within individual priors.

Results: TripleSeq generated more consistent IR priors compared to manual segmentation. TripleSeq-trained models showed good classification (> 85% mean accuracy) of ground truth infiltration.

Impact: Glioblastoma (GBM) infiltration inevitably leads to tumor recurrence and progression. We introduce an automatic method to generate infiltration risk priors for improved GBM infiltration machine learning prediction, which applied pre-operatively can identify at-risk peritumoral regions for targeted neurosurgery and radiotherapy.

08:270019.
Predicting Brain Age of Healthy Adults Based on Morphological MRI Parcellation Using Radiomics
Eros Montin1,2, Marco Muccio1,2, Chenyang Li1,2,3, Zhe Sun1,2,3, Yulin Ge1,2, and Riccardo Lattanzi1,2
1Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology,, New York University Grossman School of Medicine, New York, New York, USA, new york, NY, United States, 2Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University Grossman School of Medicine, New York, New York, USA, new york, NY, United States, 3Vilcek Institute of Graduate Biomedical Sciences, New York University Grossman School of Medicine, New York, New York, USA, new york, NY, United States

Keywords: Aging, Aging, aging, structural imaging, radiomics, neuro

Motivation: A machine learning model capable of accurately estimating brain age could have a large clinical impact.

Goal(s): To apply radiomics analysis to morphological MR images and train a machine learning model capable of accurately estimating subjects’ age from radiomics features.

Approach: T1- and T2-weighted brain images of 725 healthy adults were used to extract 18324 radiomics features from bilateral caudate, putamen, and hippocampus, and used to train a stacking regressor machine learning model.

Results: Our machine learning model accurately estimated the subjects’ age with a mean absolute error of 4.77±0.35 years using radiomics features from T1-(45%) and T2-weighted(55%).

Impact: Investigating advanced machine learning methods to accurately estimate brain aging based on commonly used clinical MR images provides vital insights to further improve our understanding of brain changes in both healthy aging and neurodegeneration.

08:390020.
Radio-pathomic signatures within and beyond FLAIR hyperintensity predict prognosis in glioblastoma following gross total resection
Savannah Duenweg1, Michael Flatley2, Aleksandra Winiarz2, Samuel Bobholz2, Allison Lowman2, Biprojit Nath2, Fitzgerald Kyereme2, Jennifer Connelly2, Dylan Coss2, Max Krucoff2, Anjishnu Banerjee2, and Peter LaViolette2
1Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States, 2Medical College of Wisconsin, Milwaukee, WI, United States

Keywords: Tumors (Pre-Treatment), Tumor, glioma, neuro-oncology

Motivation: Glioblastoma (GBM), a highly lethal brain tumor, poses a significant threat to patient survival, even after gross total resection (GTR).

Goal(s): This study explored whether radio-pathomic features from autopsy-trained models could predict survival in GTR-treated GBM patients.

Approach: The relationship between cell density and tumor probability (TPM) beyond the FLAIR hyperintense (FH) region, as well as a habitat-based labeling within FH was investigated. Cox regressions evaluated the impact of habitat volume and radio-pathomic characteristics within FH on survival.

Results: The study revealed that radio-pathomic features of FH predicted overall survival, suggesting the ability to identify infiltrative tumor ultimately missed by surgery.

Impact: In GTR-treated GBM patients, the presence of infiltrative tumor cells within and beyond FLAIR hyperintensity may predict patient prognosis and could be used for optimizing treatment.

08:510021.
Integrated MRI radiomics, tumor microenvironment, and clinical risk factors improving survival prediction in patients with glioblastoma
Qing Zhou1 and Junlin Zhou1
1Lanzhou University Second Hospital, Lanzhou, China

Keywords: Tumors (Post-Treatment), Neuro

Motivation: The patients with glioblastoma (Gb) with poor prognosis and quality of life.

Goal(s): To construct a comprehensive model for predicting the prognosis of patients with Gb using a radiomics method and integrating tumor microenvironment .

Approach: In total, 149 patients with isocitrate dehydrogenase wild-type Gb were enrolled retrospectively. Selected the best feature combination related to Gb overall survival. Clinical-radiomics-TME models were established. 

Results: Lasso-Cox analyses were used to screen the factors related to OS in patients with Gb, including age, peritumoral edema, tumor purity, tumor-associated macrophages, and the 21 radiomics features. The clinical-radiomics-TME model had the best survival prediction ability, the C-indices was 0.727. 

Impact: Considering the poor prognosis of IDH wild-type Gb, we explored additional prognostic risk factors and established a survival prediction model. The clinical-radiomics-TME comprehensive model showed a significant improvement compared to other models and was most effective in predicting patient survival.

09:030022.
CO2 and O2 reactivity in brain gliomas
Oluwateniola Sophia Akinwale1, Yang Li1,2, Peiying Liu1, Xirui Hou1, Shanshan Jiang1, Doris Lin1,3, Jay J. Pillai1,3, and Hanzhang Lu1
1Department of Radiology, Johns Hopkins University School of Medicine, Baltimore, MD, United States, 2Graduate School of Biomedical Sciences, UT Southwestern Medical Center, Dallas, TX, United States, 3Department of Neurosurgery, Johns Hopkins University School of Medicine, Baltimore, MD, United States

Keywords: Tumors (Pre-Treatment), Tumor, Cerebrovascular reactivity; venous cerebral blood volume; bolus arrival time; hypercapnia; hyperoxia

Motivation: Current clinical practice assesses baseline vascular features and cerebrovascular reactivity with multiple techniques that involve the use of injected contrast and radioactive tracers. Obtaining this information requires numerous tests and visits, which increases patient stress and healthcare costs.

Goal(s): Our goal was to determine whether a multiparametric scan could conveniently and economically assess glioma hemodynamics with no exogenous contrast.

Approach: The technique involves sequential manipulation of CO2 and O2 in inspired gas while collecting BOLD MRI images to obtain CVR, vCBV, and BAT maps.

Results: Multiparametric maps correctly differentiated tumor and normal tissue with characteristics that may inform tumor classification.

Impact: We showed that an efficient multiparametric scan can map different vascular properties. These maps allow for tumor and healthy tissue differentiation and show qualitative traits that potentially informs tumor characteristics which could aid in the diagnostic evaluation of glioma patients.

09:150023.
Is edema of malignant glioma different from edema of other tumors? Analysis of time dependent diffusion image using ternary plot method
Toshiaki Taoka1, Rintaro Ito1, Rei Nakamichi2, Toshiki Nakane2, Kazushige Ichikawa3, Takaya Mori4, Ozaki Masanori4, Nobuyasu Ichinose4, Yoshiki Tanaka5, and Shinji Naganawa2
1Department of Innovative Biomedical Visualization (iBMV), Nagoya University, Nagoya, Japan, 2Department of Radiology, Nagoya University, Nagoya, Japan, 3Department of Radiological Technology, Nagoya University, Nagoya, Japan, 4Canon Medical Systems Corporation, Otawara, Japan, 5SORD Corporation, Tokyo, Japan

Keywords: Tumors (Pre-Treatment), Diffusion/other diffusion imaging techniques

Motivation: The edema around a malignant glioma contains infiltrating tumor cells. The motivation for this study was to determine the characteristics of the edema of malignant glioma.

Goal(s): The goal is to evaluate the characteristics of edema around malignant gliomas using a combination of the oscillating gradient spin echo and pulsed gradient spin echo.

Approach: The ternary plot method is used to evaluate the characteristics of edema by using a plot of existing tissue as an internal reference.

Results: Edema of malignant gliomas showed a different distribution in relation to the internal reference in the ternary plot method compared to edema of other tumors.

Impact: Ternary plot method was used to present pixel values obtained from oscillating gradient spin echo and pulsed gradient spin echo, and existing tissues were evaluated as internal references, which was thought to enable evaluation of the histological properties of edema.

09:270024.
Combined DW-MRI and DW-MRS sensitivity to glioma tumour microenvironment: a preliminary study
Marco Palombo1,2, Samuel Rot3,4, Matteo Figini5, Elizabeth Powell6, Bhavana Solanky3,7, Chloe Najac8, Bernard Siow9,10, Jeremy Rees11, Ciaran Hill11, Eleftheria Panagiotaki5, Itamar Ronen12, and Harpreet Hyare13
1Cardiff University Brain Research Imaging Centre (CUBRIC), School of Psychology, Cardiff University, Cardiff, United Kingdom, 2School of Computer Science, Cardiff University, Cardiff, United Kingdom, 3NMR Research Unit, Queen Square MS Centre, Department of Neuroinflammation, UCL Queen Square Institute of Neurology, Faculty of Brain Sciences, University College London, London, United Kingdom, 4Department of Medical Physics and Biomedical Engineering, University College London, London, United Kingdom, 5Centre for Medical Image Computing, Department of Computer Science, University College London, London, United Kingdom, 6Centre for Medical Image Computing, Medical Physics and Biomedical Engineering, University College London, London, United Kingdom, 7Centre for Medical Image Computing, Medical Physics & Biomedical Engineering, University College London, London, United Kingdom, 8Department of Radiology, C.J. Gorter MRI Center, Leiden University Medical Center, Leiden, Netherlands, 9In Vivo Imaging, The Francis Crick Institute, London, United Kingdom, 10Centre for Medical Image Computing, University College London, London, United Kingdom, 11NMR UCL Queen Square Institute of Neurology, University College London, London, United Kingdom, 12Clinical Imaging Sciences Centre, Brighton and Sussex Medical School, University of Sussex, Brighton, United Kingdom, 13NMR Research Unit, Queen Square UCL Queen Square Institute of Neurology, University College London, London, United Kingdom

Keywords: Tumors (Pre-Treatment), Cancer, Microstructure, Diffusion, Spectroscopy

Motivation: Understanding the complexity of the tumour microenvironment is critical for understanding glioma progression and developing effective therapies. 

Goal(s): To develop a non-invasive imaging pipeline for characterization of the glioma tumor microenvironment.

Approach: Combining DW-MRI and DW-MRS to enhance the characterization of a spectrum of gliomas as a feasibility study.

Results: Changes in neuronal (tNAA) and glial (tCho) metabolites apparent diffusion coefficients suggest neuronal atrophy and glial activation/reaction in tumor core, respectively. Changes in intracellular and extracellular volume fractions and extracellular diffusivity quantified by DW-MRI support and complement metabolite DW-MRS result and further suggest a more infiltrative marging in IDH mutant tumors.    

Impact: Combination DW-MRI and DW-MRS has potential to characterize the glioma microenvironment for improved understanding of radio-resistance and developing more effective therapies. 

09:390025.
Advanced Imaging investigations of Fractal Dimension and Lacunarity measures of Glioma Subcomponents as Discriminator of IDH Status
Neha Yadav1, Ankit Mohanty1, and Vivek Tiwari1
1Department of Biological Sciences, Indian Institute of Science Education and Research Berhampur, Berhampur, India

Keywords: Tumors (Pre-Treatment), Machine Learning/Artificial Intelligence, Fractal Dimension, Radiogenomic, Lacunarity, Glioma

Motivation: The presence of structural and geometric variations within gliomas, even among those with similar histologic grades, potentially reflect the phenotypic heterogeneity because of the distinct genetic and epigenetic landscape.

Goal(s): To develop a non-invasive radiogenomic platform to identify IDH and MGMT status using the geometry of glioma subcomponent.

Approach: Fractal dimension and Lacunarity, non-Euclidean geometric measures of glioma subcomponents, were estimated using MR images and wrapped in artificial intelligence-based models to discriminate IDH status and MGMT status.

Results: The combination of fractal dimension or lacunarity of enhancing and nonenhancing glioma subcomponent is the definitive discriminator of IDH status as wildtype or mutant.

Impact: Fractal Dimension and Lacunarity of Glioma subcomponents are unique for IDH-Mutant and IDH-Wildtype gliomas. Fractal-geometry analysis can serve as an effective non-invasive tool for identifying IDH-status prior to biopsy and surgical interventions, thereby improving the clinical management of glioma patients.

09:510026.
Preoperative prediction of MGMT methylation status in high-grade glioma based on MRI radiomics signature of habitat analysis
Binju Yang1, Yueluan Jiang2, Song Yang3, Miao Chang1, and Guoguang Fan1
1The First Affiliated Hospital of China Medical University, Shenyang, Liaoning Province, China, 2MR Research Collaboration, Siemens Healthineers, Beijing, China, 3MR Research Collaboration, Siemens Healthineers, Shanghai, China

Keywords: Radiomics, Neuro, MGMT,habitat analysis,radiomics

Motivation: To predict the oxygen 6-methylguanine-DNA methyltransferase (MGMT) methylation status in high-grade gliomas (HGG) before surgery by using conventional MRI radiomics features within tumor habitat. 

Goal(s): To better understand the molecular characteristics of HGG.

Approach: In 105 HGG patients, the whole tumor was segmented into 3subregions by Kmeans clusters on T2 and T1 contrast-enhanced images. Radiomic features were extracted from each subregion and the predictive performance of radiomics signature was compared with clinical data.

Results: The efficiency of 3 subregions segmentation using Kmeans clustering with habitats analysis was the highest. The AUC of the model validation set was as high as 0.878.

Impact: We developed a radiomic signature model that can be used to predict MGMT methylation status in HGG patients. This can be used as a tool to help clinicians assess MGMT methylation status in HGG patients and guide individualized treatment.

10:030027.
MRI-based habitats to quantify tumor microenvironment normalization in glioblastoma: validation with histology and transcriptomics
Junfeng Zhang1 and Hao Wu2
1Radiology, General Hospital of Western Theater Command of PLA, Chengdu, China, 2Radiology, The First Affiliated Hospital of Chongqing Medical University, Chongqing Medical University, Chongqing, China

Keywords: Tumors (Post-Treatment), Quantitative Imaging, Habitat imaging

Motivation: The lack of in vivo and noninvasive biomarkers to quantify tumor microenvironment (TME) normalization hinders the evaluation of bevacizumab (BEV) therapy response in glioblastoma (GBM). 

Goal(s): To quantify TME normalization during BEV therapy in GBM by conventional and multiparametric MRI (mpMRI).

Approach: The MRI-based habitats were generated by Gaussian mixture model in patient-derived GBM models. Spatial-paired analyses of MRI, histology, and single-cell RNA sequencing were performed to validate the effetiveness of habitats.

Results: A total of eight habitats were generated to quantify TME normalization spatiotemporally. Habitat7 was strongly correlated with TME normalization-associated phenotypes including pericyte coverage, hypoxia and immune cell infiltration.

Impact: We developed and validated a quantitative mpMRI-based biomarker to characterize TME normalization in GBM. This may provide a new in vivo approach for precise evaluation of BEV therapy response in GBM noninvasively.