Title:
Machine-learning-based classification of gliblastoma using dynamic susceptibility enhanced MR image derived delta-radiomic features

dc.contributor.author Jeong, Jiwoong
dc.contributor.committeeMember Elder, Eric
dc.contributor.committeeMember Liu, Tian
dc.contributor.department Mechanical Engineering
dc.date.accessioned 2019-05-29T13:56:22Z
dc.date.available 2019-05-29T13:56:22Z
dc.date.created 2018-05
dc.date.issued 2018-04-25
dc.date.submitted May 2018
dc.date.updated 2019-05-29T13:56:22Z
dc.description.abstract Purpose: Glioblastoma (GBM) is the most aggressive cancer with poor prognosis due to its heterogeneity. The purpose of this study is to improve the tissue characterization of these highly heterogeneous brain tumors using delta-radiomic signature of dynamic susceptibility contrast enhanced (DSC) MR images, which are commonly used to derive blood perfusion parameters to the tumor, with machine learning approaches. Methods: Multiparametric magnetic resonance (MR) images of 25 patients with histopathologically confirmed 13 high and 12 low grade GBM were taken using a standard brain tumor imaging protocol. All DSC images were registered to T2 FLAIR images. The tumor contours in T2 FLAIR images and its contralateral regions of the normal tissue were used to extract delta-radiomic features from each DSC image over the entire volume of DSC time course images before applying feature selection methods. The most informative and non-redundant features, or signature, were selected to train a random forest to differentiate high-grade (HG) and low-grade (LG) tumors while feature correlation limits were applied to remove redundancies. Then a leave-one-out cross-validation random forest was applied to the dataset to classify GBMs. To evaluate the performance of our proposed classification method, overall prediction accuracy, confidence, sensitivity and specificity were calculated. Results: Analysis of the predictions showed that our method consistently predicted the tumor grade of 24 out of 25 patients correctly (0.96). Based on the leave-one-out cross-validation, the mean prediction accuracy was 0.95±0.10 for HG and 0.85±0.25 for LG. The area under the receiver operating characteristic curve (AUC) was 0.9380. Conclusion: Our method performed well in classifying high and low grade GBMs based on the DSC MRI data. This study shows that delta-radiomic features of DSC MRI are correlated with GBM grades and may be use to improve imaging characterization of gliomas. The performance of our method in interrogating DSC MRI data will be explored further using temporal delta-radiomic features that take advantage of the differences in tumor contrast between the baseline and peak contrast images.
dc.description.degree M.S.
dc.format.mimetype application/pdf
dc.identifier.uri http://hdl.handle.net/1853/61094
dc.language.iso en_US
dc.publisher Georgia Institute of Technology
dc.subject Delta-radiomics
dc.subject Machine-learning
dc.subject DSC MRI
dc.subject Glioblastoma
dc.title Machine-learning-based classification of gliblastoma using dynamic susceptibility enhanced MR image derived delta-radiomic features
dc.type Text
dc.type.genre Thesis
dspace.entity.type Publication
local.contributor.corporatename George W. Woodruff School of Mechanical Engineering
local.contributor.corporatename College of Engineering
relation.isOrgUnitOfPublication c01ff908-c25f-439b-bf10-a074ed886bb7
relation.isOrgUnitOfPublication 7c022d60-21d5-497c-b552-95e489a06569
thesis.degree.level Masters
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
JEONG-THESIS-2018.pdf
Size:
1.77 MB
Format:
Adobe Portable Document Format
Description:
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
LICENSE.txt
Size:
3.87 KB
Format:
Plain Text
Description: