David Qin
BME PhD Proposal Presentation

Date: 2025-12-09
Time: 2:00 - 4:00 pm EST
Location / Meeting Link: MoSE 3201A / https://gatech.zoom.us/j/97740646500?pwd=dGoRParjtfFak2K0RWP9sCOnbaIlsK.1

Committee Members:
Stanislav Emelianov, PhD (Advisor); Costas Arvanitis, PhD; Richard Bouchard, PhD; Erin Buckley, PhD; Brooks Lindsey, PhD


Title: Development of Techniques to Improve Accuracy of Quantitative Photoacoustic Imaging

Abstract:
Photoacoustic (PA) imaging integrates the high contrast of optical methods with the spatial resolution of ultrasound, making it a powerful modality for noninvasive assessment of tissue physiology and pathology. It has broad potential applications in cancer diagnosis, treatment monitoring, and quantification of various biomarkers such as tissue oxygen saturation (SO2). However, accurate quantitative PA imaging remains limited by various technical challenges, among which are: (1) depth-dependent attenuation of laser fluence, which confounds estimation of local chromophore concentrations, and (2) noise corruption in low-signal regions, which inhibits accurate spectroscopic PA (sPA)-based quantification of biomarkers. The objective of this thesis is to develop two complementary methods to address these challenges. In Aim 1, we will establish an experimental approach for determining the effective optical attenuation coefficient of heterogeneous tissue without prior knowledge of tissue composition. Using combined ultrasound/PA imaging during controlled mechanical displacement of tissue, we will measure changes in PA amplitude with respect to optical path length and estimate effective attenuation coefficients in tissue-mimicking phantoms and ex vivo tissue. In Aim 2, we will develop a method for improving the accuracy of biomarker quantification in sPA imaging and demonstrate it in SO2. We show that noise biases SO2 estimation toward physiologically plausible but inaccurate values, and that implementation of a residual-based thresholding method can reject noise-corrupted pixels and restore accuracy. We will validate this method in vitro using tube phantoms of blood and in vivo in rats during controlled oxygen challenges. Successful completion of these aims will yield experimentally-grounded strategies for both fluence compensation and noise rejection that can be easily integrated into the conventional PA imaging workflow. The proposed work aims to improve the accuracy and robustness of quantitative PA imaging, thereby facilitating translation into preclinical and clinical studies.