Frequency-Domain Fluorescence Lifetime Imaging Systems for Cancer Imaging Applications

Abstract

Current cancer diagnostic protocols rely heavily on subjective visual inspection and invasive biopsies, often failing to detect early-stage malignancy, the phase where the disease is most responsive to treatment. Furthermore, standard screening modalities often have low specificity, leading to a high volume of unnecessary biopsies. Fluorescence Lifetime Imaging (FLIM) offers a quantitative, label-free method for interrogating tissue metabolism, potentially addressing these diagnostic limitations. However, clinical translation has historically been stalled by the bulk, high cost, and latency of benchtop instrumentation. This dissertation details the engineering, characterization, and clinical validation of three compact Frequency-Domain FLIM (FD-FLIM) systems designed for oral and dermatological cancer imaging in clinical settings. First, a Handheld Multispectral FD-FLIM Endoscope was engineered for oral cancer screening, utilizing dual-wavelength excitation to isolate cancer-associated metabolic reprogramming. The probe achieves a spatial resolution of <300 µm across a 10 mm circular field of view. Preliminary tests showed that the system successfully distinguished malignant lesions from healthy tissue using fluorescence lifetime contrast. Following this validation, a total of four imaging systems were built and deployed to partner clinics, establishing a multi-center study. Second, a Wide-Field FD-FLIM Dermascope was developed for macroscopic skin assessment. Providing <200 µm resolution over a 3.5 cm² area, this system enables rapid screening of pigmented lesions. In vivo validation quantified the metabolic homogeneity of benign nevi, statistically differentiating them from surrounding healthy tissue and demonstrating the potential of this technology to prevent unnecessary biopsies. Third, an Intraoperative Open-Top FLIM (IOT-FLIM) system was designed to enable rapid, label-free imaging of fresh resected tissue during Mohs micrographic surgery. Capable of imaging fresh, unsectioned tissue with <60 µm resolution over ~3.6 cm², this platform was utilized to characterize tissue properties. Preliminary results showed the potential to characterize distinct spectral-temporal profiles of normal, cancerous, and adipose tissue, highlighting the system's potential as an auxiliary tool for margin assessment during Mohs surgery. Collectively, these developments demonstrate that compact and high-speed optical instrumentation can successfully integrate quantitative metabolic imaging into clinical settings. This work establishes these three distinct imaging systems as promising auxiliary technologies for improving non-invasive cancer detection and providing complementary metabolic contrast during surgical margin assessment.