Quantitative image feature extraction from volume data : methods and applications to cancer imaging
Abstract/Contents
- Abstract
- This dissertation aims to increase the stability and reduce the execution time for computing radiomics features using three different approaches: (1) developing an open source, modular, locally-run or server-based system for 3D radiomics feature computation that can be used on any computer system and included in existing workflows for understanding associations and building predictive models between image features and clinical data, such as survival, (2) investigating the utility of obtaining "core samples" of regions in CT volume scans for extraction of radiomic features, and (3) proposing a new paradigm, called "digital biopsy, " that allows for the collection of intensity- and texture-based features from sub-regions of each volume of interest at least 1 order of magnitude faster than the current manual or semi-automated segmentation methods. For (1) I developed a software package that exploits various levels of parallelization for use on multiprocessor systems. It consists of a managing framework and 4 stages: input, pre-processing, feature-computation, and output. Each stage contains one or more swappable components, allowing run-time customization. Two versions of the QIFE have been released: 1) the open-source MATLAB code posted to Github, 2) a compiled version loaded in a Docker container, posted to DockerHub, which can be easily deployed on any computer. Computational efficiency can be improved by parallelizing execution at the cost of memory usage. Different parallelization levels provide different trade-offs, and the optimal setting will depend on the size and composition of the dataset to be processed. I benchmarked the engine using various levels of parallelization on a cohort of CT scans presenting 108 lung tumors. The QIFE processed all 108 tumors in 2:12 (h:mm) using 1 core, and 1:04 (h:mm) hours using 4 cores, with object-level parallelization. For (2) I asked four readers to outline tumors in three representative slices from each phase of multiphasic liver CT images taken from 29 patients (1128 segmentations) with hepatocellular carcinoma. Core samples were obtained by automatically tracing the maximal circle inscribed in the outlines. Image features describing the intensity, texture, shape, and margin were used to describe the segmented lesion. The intra-class correlation between the features extracted from the readers' segmentations and their core samples was used to characterize robustness to segmentation between readers, and between human-based segmentation and core sampling. I concluded that despite the high inter-reader variability in manually delineating the tumor (average overlap of 43% across all readers), certain features such as intensity and texture are robust to segmentation. More importantly, this same subset of features can be obtained from the core samples, providing as much information as detailed segmentations while being simpler and faster to obtain. For (3) a radiologist reviewed automated segmentations of lung nodules from 100 preoperative volume computed tomography scans of patients with non--small cell lung cancer, and manually adjusted the nodule boundaries in each section, to be used as a reference standard, requiring up to 45 minutes per nodule. A different expert generated a digital biopsy for each patient using a paintbrush tool on multiple contiguous cross-sections, a procedure that required an average of 3 minutes per nodule. I simulated additional digital biopsies using morphological operations. Finally, I compared the features extracted from these digital biopsies with our reference standard using the intra-class correlation coefficient (ICC) to characterize robustness. Comparing the reference standard segmentations to our digital biopsies, I found that 84/94 features had an ICC 0.7. Comparing erosions and dilations of our digital biopsies to the reference standard segmentations resulted in 41/94 and 53/94 features, respectively, with ICCs 0.7. I concluded that many intensity- and texture-based features can be reliably obtained via digital biopsy while substantially reducing the amount of operator time required. I conclude that these three methods, when implemented, reduce the time and increase the stability in the first two stages of the radiomics workflow: segmentation and feature extraction.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2017 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Echegaray, Sebastian |
|---|---|
| Associated with | Stanford University, Department of Electrical Engineering. |
| Primary advisor | Napel, Sandy |
| Thesis advisor | Napel, Sandy |
| Thesis advisor | Nishimura, Dwight George |
| Thesis advisor | Pauly, John (John M.) |
| Advisor | Nishimura, Dwight George |
| Advisor | Pauly, John (John M.) |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Sebastian Echegaray. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2017. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2017 by Sebastian Echegaray
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...