Innovations in imaging sequences and reconstruction pipelines for ultrasound localization microscopy

Shin, YiRang

Permalink

https://hdl.handle.net/2142/130033 Copy

Description

Title

Innovations in imaging sequences and reconstruction pipelines for ultrasound localization microscopy

Author(s)

Shin, YiRang

Issue Date

2025-07-17

Director of Research (if dissertation) or Advisor (if thesis)

Song, Pengfei

Doctoral Committee Chair(s)

Oelze, Michael

Committee Member(s)

• Anastasio, Mark
• Lam, Fan
• Yu, Cunjiang

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Ultrasound imaging
• Super-resolution ultrasound
• 3D ultrasound imaging
• Functional ultrasound imaging

Abstract

The microcirculation network, consisting of arterioles, capillaries, and venules, is the primary site for exchanging oxygen, nutrients, and waste products. This network is crucial for maintaining tissue viability and function. Imaging the microvasculature is vital because it reveals subtle changes in tissue perfusion and vessel integrity -often precursors to overt organ dysfunction. Such imaging enables earlier disease detection, targeted therapies, and monitoring treatment outcomes across various pathologies. In the past decade, Ultrasound localization microscopy (ULM) has emerged as a super-resolution imaging technique capable of imaging microvascular networks in deep tissue. It achieves this by localizing and tracking intravascular microbubbles at sub-wavelength precision, bypassing the resolution-penetration trade-off of ultrasound. ULM provides micron-scale spatial resolution at a depth unattainable by optical modalities and offers finer in vivo vascular detail than that of MRI and CT. ULM uses widely accessible ultrasound systems and FDA-approved microbubbles, giving it great promise for clinical use. It can also complement conventional vascular imaging by offering a unique balance of resolution, depth, and functional insight. However, translating ULM into clinical practice faces several challenges. A primary limitation is the prolonged data acquisition time, as it requires accumulating thousands of frames to capture a sufficient microbubble localization event. The technique also relies on a serial post-processing pipeline (beamforming, clutter filtering, localization, tracking) that limits its real-time application. Furthermore, ULM has inherent limitations stemming from 2D planar imaging, which cannot accurately capture complex 3D vascular architectures. To address these current ULM limitations, this thesis introduces a comprehensive framework incorporating innovative imaging sequences and advanced reconstruction pipelines for ULM. Specifically, we present LOCA-ULM, a deep learning-based simulation and localization pipeline that enables robust microbubble localization even at high concentrations, thereby substantially reducing acquisition time. Moreover, we adopt Trackformer to replace the computationally expensive microbubble detection and tracking stages with an end-to-end tracking-by-attention framework. Finally, we introduce Fast3D-AM, a high-volume acquisition rate 3D nonlinear imaging method using multiplexed 2D matrix arrays to reduce tissue clutter signals and facilitate robust, high-quality 3D ULM. We validate these frameworks with structural, functional, and volumetric imaging of diverse organs in small animal models and demonstrate its capability for comprehensive microvascular visualization and hemodynamic mapping. These technological advances represent a significant step towards the successful translation of ULM into both preclinical research and clinical settings.

Graduation Semester

2025-08

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/130033

Copyright and License Information

Copyright 2025 YiRang Shin

Owning Collections