Abstract

Cancer intervention increasingly relies on technologies that visualize disease, guide its removal, and treat residual tumor burden with high spatial precision. Optical imaging is especially attractive for this purpose because it can provide real-time visualization of tumor margins, molecular contrast, and treatment response. However, current optical approaches are constrained by practical and biological limitations. Fluorescence-guided surgery can improve tumor localization, but many clinical systems are bulky, operator-dependent, and difficult to standardize across users, institutions, and imaging environments. Photodynamic therapy can eradicate residual disease, but its efficacy is limited by photosensitizer localization, tissue autofluorescence, and the shallow penetration of visible light. Thermal imaging offers a complementary, label-free approach to disease detection, but passive thermography lacks specificity and provides limited information about deeper tissue structure. These limitations underscore the need for multidimensional theranostic platforms that integrate imaging, guidance, and therapy for precision cancer intervention. This dissertation addresses this need by developing and evaluating complementary optical and hybrid imaging strategies for cancer detection, surgical guidance, and therapy. First, we developed a single-operator Cancer Vision Goggles (CVG), a wearable near-infrared fluorescence imaging platform designed to standardize fluorescence-guided surgery and reduce operator-dependent variability. Unlike conventional fluorescence imaging systems that often require multiple operators and variable acquisition geometries, the CVG enables single-operator, hands-free imaging via a head-mounted binocular system integrating visible and near-infrared cameras, real-time image co-registration, an optical see-through display, dual green alignment lasers, and a posture-dependent laser safety interlock. The alignment lasers converge at a 50-cm preset working distance, enforcing reproducible imaging geometry while preserving the natural surgical workflow. A Bluetooth foot pedal and a graphical user interface further enable hands-free acquisition of paired laser-on and laser-off images for quantitative analysis. The CVG was developed and characterized as an instrumentation platform for quantitative near-infrared fluorescence-guided oncologic surgery. Phantom and resolution studies demonstrated that the device achieved 281 µm spatial resolution at the standardized 50-cm working distance and provided a broad, reproducible excitation field. Sensitivity testing with ICG-Intralipid phantoms showed detectable fluorescence at low picomolar concentrations, with linearity preserved across low-to-moderate signal ranges. In murine tumor models using LS301-HSA and in ex vivo human head and neck cancer specimens from patients injected with a near-infrared molecular probe, the CVG detected tumor-associated fluorescence and enabled real-time visualization of tumor contrast. Dynamic thresholding further enhanced tumor-background delineation and provided quantitative fluorescence metrics directly on the display device. These studies established the CVG as a wearable imaging system capable of combining practical surgical guidance with radiometrically faithful quantitative data capture. Second, we extended this platform from device validation to translational benchmarking across preclinical and clinical settings. The CVG was evaluated against established fluorescence imaging systems using quantitative endpoints, including tumor-to-nontumor ratio, normalized intensity maps, and Dice spatial overlap. In preclinical tumor models, the CVG produced significant tumor contrast and high spatial agreement with reference imaging systems. Importantly, the CVG maintained stable fluorescence detection over a wide working-distance range, from 10 to 60 cm, whereas handheld systems are more susceptible to changes in working distance and operator positioning. This stability supports more reproducible fluorescence measurements across users and experimental settings. The clinical component further demonstrated that the CVG could image human tumor specimens from patients administered a near-infrared molecular probe, with performance comparable to an established clinical fluorescence-guided surgery system, while offering advantages in portability, single-operator use, and workflow integration. This work validated the CVG as a standardized bridge between preclinical fluorescence imaging and clinically relevant surgical guidance. Building on this foundation, the dissertation next addresses a central challenge in fluorescence-guided cancer surgery. Even when image-guided resection improves tumor removal, microscopic residual disease can persist and drive recurrence. This challenge is particularly significant in malignant glioma, where 5-aminolevulinic acid-induced protoporphyrin IX (5-ALA-PpIX) is used clinically for fluorescence-guided resection. Unfortunately, PpIX’s visible fluorescence is weak, tissue autofluorescence is pronounced, and light penetration in tissue is shallow. Here, we introduce LS301, a tumor-targeted near-infrared fluorescent probe, as a surrogate biomarker for PpIX-enriched glioma tissue. Because LS301 emits in the near-infrared (NIR) range, whereas PpIX emits in the visible range, the combination provides complementary optical contrast. In glioma cells and mouse tumor models, LS301 co-localized with PpIX with high spatial correlation, enabling LS301 to identify PpIX-enriched tumor regions even when the PpIX signal was weak or difficult to distinguish from background autofluorescence. Fluorescence lifetime imaging further separated LS301, PpIX, and endogenous autofluorescence signals, improving the specificity of tumor mapping. In immunocompetent glioma models and simulated fluorescence-guided resection settings, LS301 detected residual disease not readily visualized by PpIX and enabled targeted PDT of tumor cavity margins. We then pivoted from image-guided surgery to image-guided phototherapy. Beyond its role as a near-infrared surrogate marker, LS301 was shown to function as a photosensitizer under 793-nm irradiation, generating reactive oxygen species and producing measurable photodynamic cytotoxicity. This enabled a dual-wavelength photodynamic therapy strategy in which visible-light activation of endogenous PpIX was paired with near-infrared excitation of exogenous LS301 to treat both superficial and deeper residual tumor burdens. Dual-wavelength therapy delayed tumor relapse compared with 5-ALA-PpIX photodynamic therapy alone. We discovered that LS301-mediated PDT produced antitumor effects without the skin scarring observed with 5-ALA-PpIX treatment, suggesting a potentially improved safety profile. This work amplifies the therapeutic effect of surgical resection by demonstrating how fluorescence guidance can be extended beyond tumor visualization to spatially targeted eradication of residual disease. Finally, we broadened the concept of precision intervention beyond fluorescence alone by introducing ultrasound-stimulated thermal imaging (USTI), a hybrid strategy for dynamic mapping of tissue structure and disease using infrared cameras. Conventional thermal imaging is widely used to detect infrared emissions associated with metabolism, inflammation, and perfusion. However, passive surface thermography is limited by environmental sensitivity, low specificity, and shallow sampling depth. USTI addresses these limitations by using ultrasound to perturb local thermal equilibrium and infrared imaging to capture the resulting thermal stimulation and recovery patterns at the tissue surface. Because different tissues vary in acoustic absorption, bioheat transfer, and thermal relaxation, their responses to ultrasound stimulation encode information about tissue architecture and pathology. Incorporating a pinhole stimulation geometry improved spatial resolution and enhanced tissue classification. In mouse tumor models, USTI localized tumors and delineated tumor-normal boundaries based on delayed thermal re-equilibration in malignant tissue, likely reflecting disrupted cellular, stromal, and vascular architecture. Although focused ultrasound thermal ablation was not implemented in this dissertation, the platform establishes a foundation for future theranostic applications in which ultrasound-compatible therapeutic interventions could be integrated for deeper disease detection, treatment, and response monitoring. Collectively, this dissertation presents a multidimensional theranostic framework for precision cancer intervention. The CVG chapters establish standardized, quantitative, single-operator near-infrared fluorescence imaging for surgical guidance and translational benchmarking. The LS301-PpIX chapter extends optical guidance into therapy by using near-infrared fluorescence as both a surrogate for visible PpIX fluorescence and a therapeutic photosensitizer for dual-wavelength photodynamic treatment of residual glioma. The USTI chapter expands the platform toward label-free, dynamic, noninvasive mapping of deeper tissue structure and disease. Together, these studies evolved from visualizing tumors more reproducibly, to guiding their removal more precisely, to treating residual disease more effectively, and finally to probing deeper tissue features beyond the reach of conventional optical imaging. This integrated body of work supports the central premise that multidimensional imaging and therapy platforms can improve the precision, reproducibility, and impact of cancer interventions.

Committee Chair

Samuel Achilefu

Committee Members

Lori Setton, Christine O'Brien; Michael Rettig; Song Hu

Degree

Doctor of Philosophy (PhD)

Author's Department

Biomedical Engineering

Author's School

McKelvey School of Engineering

Document Type

Dissertation

Date of Award

8-17-2026

Language

English (en)

Download

Available for download on Monday, August 14, 2028

Share

COinS