Xiangyu Yang

My Story:

Research interests:
Quantitative approaches in MRI

Thesis title:
Quantitative Approaches in MRI with Clinical Applications

Abstract:
Magnetic Resonance Imaging (MRI) has been established as a primary imaging module in clinical practices. However, the mainstream clinical application of MRI is currently dominated by the qualitative reading approach. Advanced quantitative approaches provide more objective information about tissue component and function, which can be essential to improve the accuracy and efficiency of disease diagnosis, treatment plan design, and drug effect monitoring. Several studies had been conducted to explore techniques that can be used to improve various quantitative MRI approaches in clinical-related context. The results were summarized in this text. Pharmacokinetic analysis of Dynamic Contrast Enhanced MRI data is an important method to assess pathophysiological permeability changes in cancerous tissue. A modified two-compartment pharmacokinetic model was proposed, based on more realistic assumptions. Pharmacokinetic parameters derived with this model were demonstrated to have better tissue specificity in distinguishing metastatic tumor from normal tissue in liver. MRI also showed a great potential in quantitative assessment of iron overload diseases. The feasibility of using five relaxometry-based or signal-intensity-ratio-based quantitative MRI approaches as liver iron content (LIC) biomarkers was assessed in a group of 7 patients by comparing MRI LIC measurements with the liver biopsy 'golden standard'. Three relaxometry-based approaches were demonstrated to have insignificant difference with liver biopsy, and identified as feasible LIC biomarkers at 1.5 Tesla. A technical obstacle in applying the MRI tissue iron biomarkers to high and ultrahigh field systems is the accelerated dephasing caused by macroscopic background field (B0) inhomogeneity. A postprocessing strategy that can correct for quadratic B0 inhomogeneity was proposed and applied to phantom and volunteer data to demonstrate its superiority over no correction or linear correction methods. The proposed quadratic correction strategy can also generate an index of reliability of the correction results. Computer simulation is a powerful complementary technique when the experimental design is compromised due to ethic consideration or budget limit. In order to explore subtle details in the tissue-iron contrast mechanism, a high speed Monte Carlo algorithm was proposed to enable simulation on nano scale. The prototype algorithm was implemented. Preliminary testing results showed a promising improvement in computing speed.