2026-09-24T18:56:54Z https://www.ideals.illinois.edu/oai-pmh

oai:www.ideals.illinois.edu:2142/125703 2026-02-03 col_2142_5131 col_2142_5163 com_2142_5130 com_2142_5162 com_2142_234

application/pdf en text Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2026-08-01 The student, Goodluck Okoro, accepted the attached license on 2024-07-09 at 07:46. The student, Goodluck Okoro, submitted this Thesis for approval on 2024-07-09 at 07:56. This Thesis was approved for publication on 2024-07-16 at 15:22. DSpace SAF Submission Ingestion Package generated from Vireo submission #20996 on 2025-02-04 at 21:16:30 Pet-based heterogeneity analysis: Strategies for radiomic feature selection Okoro, Goodluck 2024-07-16 Dobrucki, Wawrzyniec Radiomics Feature Selection Repeatability Heterogeneity Analysis Positron Emission Tomography eng Although several radiomic features have been reported for heterogeneity analysis in Positron Emission Tomography (PET) imaging, there is a lack of standardization and consensus on the most appropriate set of heterogeneity parameters to use in different imaging applications. In this study, with computer-generated heterogeneous images and different heterogeneous phantom patterns, I used 20 widely reported repeatable radiomic features to investigate the impact of critical parameters such as image acquisition time, heterogeneous patterns, image resolution, reconstruction parameters, and choice of filter on repeatability and feature selection. My findings reveal the pivotal role of image acquisition time in influencing the reliability of radiomic features and heterogeneity calculations. Furthermore, by imaging a phantom over a duration of time and analyzing radiomic features at different time points, it is possible to capture the dynamics of feature stability in the context of changing acquisition times. With this methodology, I observed that certain intensity-based radiomic features exhibit greater stability amidst variations in imaging and post-processing parameters. I have also presented suggestions on possible ways to approach heterogeneity analysis for a more robust radiomic feature selection. 2024-08 Thesis https://hdl.handle.net/2142/125703 Copyright 2024 Goodluck Okoro Bioengineering Bioengineering University of Illinois at Urbana-Champaign M.S. Thesis