Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Computational Pathology for Quantifying Spatial Heterogeneity in Digital Images of Tissue Sections from Solid Tumors

Nguyen, Luong (2017) Computational Pathology for Quantifying Spatial Heterogeneity in Digital Images of Tissue Sections from Solid Tumors. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

[img]
Preview
PDF
Download (84MB) | Preview

Abstract

Tumor heterogeneities have been linked to cancer patient outcomes and responses to therapies. With the increasing popularity of whole slide imaging systems and the development of hyperplexed immunofluorescence platforms,
the volume of digital images of tissue sections from solid tumors is growing with an unprecedented rate. The era of big data poses a challenge to extracting useful information about tumor heterogeneity. This thesis presents computational pathology algorithms for modeling tumor heterogeneity in transmitted light and immunofluorescence digital images of breast and colon tissues. In transmitted light images of hematoxylin and eosin stained tissue sections, we characterize tumor heterogeneity by the relative spatial arrangement of different histological structures in the tissues. To identify these structures, we developed two automated segmentation methods based on color statistics and internuclear distance distributions. After segmenting
histological structures, we assign clinically relevant labels (e.g. invasive carcinoma, blood vessels, etc.) to them based on their cytological and architectural
features. Finally, we quantify the spatial distributions of these histological structures, classify the whole image using the spatial characteristics, and validate our classification against pathologists' annotations.
In immunofluorescence images, tumor spatial heterogeneity can be quantified using correlations between biomarker expressions of different cell types in a neighborhood of various sizes. We fi�nd that the biomarker correlations are superior to clinical histopathological covariates
in terms of predicting recurrence status of colorectal cancer patients.


Share

Citation/Export:
Social Networking:
Share |

Details

Item Type: University of Pittsburgh ETD
Status: Unpublished
Creators/Authors:
CreatorsEmailPitt UsernameORCID
Nguyen, Luonglun5@pitt.edulun5
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairTaylor, D. Lansingdltaylor@pitt.edudltaylor
Thesis AdvisorChennubhotla, S. Chakrachakracs@pitt.educhakracs
Committee MemberYang, Gegeyang@cmu.edugeyang
Committee MemberLee, Adrian V.avl10@pitt.eduAVL10
Committee MemberFine, Jeffreyfinejl@upmc.edufinejl
Date: 10 August 2017
Date Type: Publication
Defense Date: 4 August 2017
Approval Date: 10 August 2017
Submission Date: 7 August 2017
Access Restriction: 5 year -- Restrict access to University of Pittsburgh for a period of 5 years.
Number of Pages: 139
Institution: University of Pittsburgh
Schools and Programs: School of Medicine > Computational Biology
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: computational pathology
Date Deposited: 10 Aug 2017 13:22
Last Modified: 10 Aug 2022 05:15
URI: http://d-scholarship.pitt.edu/id/eprint/33022

Metrics

Monthly Views for the past 3 years

Plum Analytics


Actions (login required)

View Item View Item