Depth-Resolved Macroscopic Fluorescence Lifetime Imaging via High-Spatial-Frequency Structured Illumination

Nanxue Yuan, Saif Ragab, Navid Ibtehaj Nizam, Vikas Pandey, Amit Verma, Tynan Young, John E. Williams, Margarida Barroso, Xavier IntesbioRxiv 2025 · DOI

TL;DR

This paper introduces High Spatial Frequency-Fluorescence Lifetime Imaging (HSF-FLI), a novel technique that uses structured light patterns to precisely distinguish fluorescence signals originating from different depths within scattering tissues. This advancement overcomes a major limitation in macroscopic fluorescence lifetime imaging, enabling more accurate and reliable molecular imaging in living systems.

What Problem Does This Paper Address?

Macroscopic Fluorescence Lifetime Imaging (MFLI) is a powerful tool for non-invasively probing biological processes. It offers contrast that's largely independent of fluorophore concentration and excitation intensity, making it ideal for applications like tracking drug delivery or studying tumor microenvironments. However, a significant challenge in MFLI is determining the depth from which the fluorescence signal originates. In scattering tissues like skin, surface fluorescence can easily overwhelm deeper signals, leading to ambiguous interpretations. Existing methods to mitigate this, like using optical clearing agents, can alter tissue properties and compromise lifetime measurements. This paper aims to solve this depth ambiguity problem by developing a method that can selectively isolate signals from different depths without invasive procedures.

Methods

The core innovation is the High Spatial Frequency-Fluorescence Lifetime Imaging (HSF-FLI) framework. This approach leverages structured illumination, specifically three-phase sinusoidal patterns projected by a digital micromirror device (DMD), to modulate the light's penetration depth. By analyzing how different spatial frequencies of illumination are attenuated within the tissue, the system can effectively deconvolve the total fluorescence signal into surface-scattered (IAC) and subsurface (Isub) components.

A key component of HSF-FLI is the development of a depth-correlated Modulation Transfer Function (MTF). This MTF was derived using Monte Carlo eXtreme (MCX) simulations to model light transport in turbid media and then experimentally validated using agar-based phantoms. The HSF-FLI method was implemented on a time-gated ICCD system. Nonlinear least-squares fitting was then applied to the separated subsurface signals to accurately estimate fluorescence lifetimes.

Datasets

The study utilized two main types of datasets:

  • Simulated Data: Monte Carlo eXtreme (MCX) simulations were used to generate fluorescence data from computationally designed phantoms with embedded fluorescent inclusions at various depths and spatial characteristics. This allowed for the physics-based modeling of light propagation and fluorescence.
  • Experimental Data:
    • Agar-based Capillary Phantoms: These phantoms, filled with fluorescent dyes (AF700), were used to validate the MTF and the HSF-FLI method under controlled conditions. Different spatial frequencies were tested to assess depth resolution.
  • In Vivo Mouse Models: The HSF-FLI method was demonstrated in live mice bearing HER2+ breast tumor xenografts. These experiments involved both non-FRET and FRET probes (MDT-TZM conjugated with donor and acceptor NIR dyes) to assess drug delivery and molecular interactions.
  • Ex Vivo Measurements: Tissue samples from the in vivo experiments were also imaged ex vivo to cross-validate the HSF-FLI results.

Key Insights and Ideas

  • Structured Illumination as a Depth Probe: The paper demonstrates that by carefully controlling the spatial frequency of structured illumination, one can effectively probe different depths within scattering tissues. Higher spatial frequencies are more sensitive to shallower depths, while lower frequencies penetrate deeper.
  • MTF as a Key to Depth Correlation: The development and validation of a depth-correlated MTF is crucial. It provides a quantitative link between spatial frequency and penetration depth, enabling the accurate deconvolution of surface and subsurface signals.
  • Eliminating Surface Bias Without Chemical Agents: HSF-FLI offers a significant advantage by removing surface signal contamination without the need for chemical clearing agents, which can alter biological samples. This preserves the native state of the tissue for more accurate lifetime measurements.
  • Improved Lifetime Precision: By isolating subsurface signals, HSF-FLI significantly reduces lifetime variability and improves the precision of lifetime measurements, especially in deeper tissues where scattering is more pronounced.

Key Results

  • MCX Simulations Validate Depth-Dependent Attenuation: MCX simulations showed strong agreement (R² ≥ 0.87) with experimental phantom data across depths up to 3.4 mm, confirming the model's ability to predict depth-dependent attenuation of structured light.
  • Maximum Usable Spatial Frequency Identified: The study determined that the maximum usable spatial frequency for their ICCD-DMD system is approximately 0.6 mm⁻¹, beyond which pattern distortion becomes significant.
  • Significant Reduction in Lifetime Variability: In phantom experiments, HSF-FLI demonstrated a systematic reduction in lifetime variability as spatial frequency increased. At 0.6 mm⁻¹, the standard deviation of fitted lifetimes decreased by nearly 50%, with mean lifetimes remaining within 5% of scattering-free references.
  • In Vivo Validation in Tumor Models: HSF-FLI successfully delineated tumors and urinary bladders in vivo, suppressing superficial scattering and improving lifetime accuracy.
  • Quantitative FRET Measurements: The method enabled quantitative FRET measurements in mouse models, showing reduced standard deviations in FRET efficiencies compared to wide-field data.

Limitations and Future Work

  • System Constraints: The current ICCD-DMD system has a practical limitation on the maximum spatial frequency (0.6 mm⁻¹), which restricts the achievable depth discrimination.
  • Photon Throughput: Higher spatial frequencies inherently reduce photon throughput, leading to lower signal-to-noise ratios at greater depths.
  • Future Work: The authors suggest several avenues for future research:
  • Hardware Improvements: Utilizing higher-resolution DMD chips, faster gating electronics, or adaptive optics could extend the spatial frequency limit and improve depth discrimination.
  • Optimizing Illumination and Detection: Strategies to mitigate the trade-off between spatial resolution and photon throughput, such as optimizing illumination power or detector settings, are proposed.
  • Advanced Analysis Techniques: Integrating advanced lifetime analysis techniques like phasor plots or global fitting could enhance noise robustness and speed up data processing.
  • SWIR Imaging: Translating HSF-FLI to the short-wave infrared (SWIR) window could offer deeper penetration and lower background autofluorescence, making it more suitable for clinical applications.
  • Multi-Frequency Acquisition: Combining multiple spatial frequencies in a single acquisition could reconstruct depth-resolved lifetime maps more efficiently.

Summary updated

Abstract

Abstract Significance Macroscopic Fluorescence Lifetime Imaging (MFLI) is a powerful, non-invasive imaging modality that offers robust, physiologically relevant contrast largely independent of fluorophore concentration, excitation intensity, and tissue signal attenuation. However, accurately determining the depth of signal origin remains challenging, potentially leading to ambiguity in biological interpretation. Here, we present a novel optical correction method that effectively eliminates surface signal bias, such as that from skin in preclinical imaging, without the need for chemical clearance. This advancement supports the robust applicability of MFLI in translational research. Aim Establishment of a High Spatial Frequency–Fluorescence Lifetime Imaging (HSF-FLI) framework to selectively isolate subsurface fluorescence (deeper signals) from surface fluorescence, while preserving the accuracy of lifetime estimation. Approach A modulation transfer function (MTF) that relates spatial frequency to penetration depth was derived using Monte Carlo eXtreme (MCX) simulations (for physics-based modeling) and validated with agar-based capillary phantoms on a time-gated ICCD–DMD system. Depth-independent fluorescence was decomposed into surface and subsurface components through structured three-phase sinusoidal illumination, and nonlinear least squares fitting was applied to recover lifetime or lifetime based parameters maps. HSF-FLI was demonstrated in vivo in mouse models bearing tumor xenogratfs and was cross validated with ex vivo measurements. Results We extensively characterized the performance of High Spatial Frequency–Fluorescence Lifetime Imaging (HSF-FLI) through simulations and tissue-mimicking phantoms. The approach was further validated in vivo by assessing drug delivery in preclinical models using MFLI-FRET (Förster Resonance Energy Transfer). Conclusion By coupling structured illumination with physics-based depth modeling, HSF-FLI delivers accurate, depth-selective lifetime readouts, setting the stage for robust and fast FLI implementation in translational studies.
BibTeX
@misc{yuan2025depth,
  author = {Nanxue Yuan and Saif Ragab and Navid Ibtehaj Nizam and Vikas Pandey and Amit Verma and Tynan Young and John E. Williams and Margarida Barroso and Xavier Intes},
  title = {{Depth-Resolved Macroscopic Fluorescence Lifetime Imaging via High-Spatial-Frequency Structured Illumination}},
  year = {2025},
  doi = {10.1101/2025.06.10.658928},
  howpublished = {bioRxiv},
}
Figure