ESTIMATING PERFUSION AND VASCULAR PROPERTIES FROM MEDICAL IMAGES: QUANTITATIVE TRANSPORT MAPPING (QTM)

Perfusion refers to the passage of fluid through the circulatory system to tissue and is defined as delivery of blood to the capillary bed in tissue. Perfusion as well as other vascular properties, including vessel permeability, vascular space volume, and extravascular space volume, directly reflect whether the organ is under normal physical condition. Direct estimation of perfusion parameters is not feasible in most situations, as capillaries are deeply buried in human body, and invasive operations will disrupt blood flow and affect organ function. Under this circumstance, contrast enhanced medical imaging techniques were developed, where contrast agents are injected into human body (alternatively, spin labeled water molecule can also work as contrast agent), and then images are acquired while contrast agent travels through the region of interest. By analyzing images before and after contrast agent injection, contrast agent propagation and underlying blood perfusion parameters can be estimated.This thesis developed a fast and accurate tracer propagation simulation method based on underlying fluid mechanics, as well as an inversion method that can recover perfusion parameters from contrast enhanced medical images (Quantitative Transport Mapping, QTM). The tracer propagation simulation method provides a rigorous way to simulate the traveling of contrast agent in organs. QTM overcame the disadvantages of traditional perfusion parameter estimation method (kinetic modeling method), and makes perfusion parameter estimation from medical images accurate, fast and automatic.

Name thesis:degree_name

Ph. D., Biomedical Engineering

Level thesis:degree_level

Doctor of Philosophy

Discipline thesis:degree_discipline

Biomedical Engineering

Grantor

Cornell University

Year dc:date.issued

2023

Author dc:creator

• Zhang, Qihao

Committee members dc:contributor.committeemember

• Datta, Ashim
• Kuceyeski, Amy

Language dc:language.iso

en

DOI

https://doi.org/10.7298/9m5p-xt38

Dc Identifier Other

ProQuest Submission ID: 13475
ProQuest Publication ID: 30249594

OAI identifier oai:identifier

oai:ecommons.cornell.edu:1813/114180