FLUORESCENCE SPECTROSCOPY/BIOMEDICAL IMAGING: Fluorescence 'lifetime' moves toward clinical application

Fluorescence represents a ubiquitous means of achieving label-free optical molecular contrast—and the approach can be applied using a range of instruments, including spectrophotometers, microscopes, and endoscopes. Fluorescence measurements can provide information on the specific molecular features in biological tissues. Various pathological conditions and therapeutic intervention can lead to biochemical, functional, and structural transformations of endogenous fluorescent biomolecular complexes in tissue and cells, including structural proteins, enzyme metabolic co-factors, lipid and lipoproteins constituents, and porphyrins. Thus, such transformations carry diagnostic information. For these reasons, tissue autofluorescence has been increasingly explored over the past two decades as a tool for diagnosis of illnesses ranging from a variety of cancers to cardiovascular diseases.^1,2

Clinical diagnostic applications of autofluorescence have been based mainly on steady-state fluorescence-an approach that uses relatively simple and inexpensive instruments. However, this technique relies on measurement of fluorescence emission intensity and/or spectra that are affected by numerous factors, including irregular tissue surfaces, non-uniform sample illumination, and the presence of endogenous absorbers such as blood in the operative field. An alternative method that addresses such limitations is based on time-resolved (lifetime) measurement of fluorescence emission. This technique can resolve excited-state lifetimes and improve the specificity of fluorescence measurements.³

Despite recognized advantages, the full potential value of time-resolved fluorescence spectroscopy and imaging has been only sparsely evaluated in clinical settings. Barriers to clinical translation include the complexity of the instrumental setups, lengthy data acquisition and analysis, and high instrumentation costs associated with these techniques.¹ Recent advances in multispectral time-resolved fluorescence spectroscopy (ms-TRFS) technologies, however, may overcome such barriers and allow for development of versatile diagnostic systems that can operate as both point-spectroscopy and scanning imaging spectroscopy devices. In addition, development of endoscopic wide-field fluorescence lifetime imaging microscopy (FLIM) systems can add diagnostic value to existing endoscopic systems.

ms-TRFS for in vivo tissue diagnosis

Tissue fluorescence emission represents a superposition of many spectrally overlapping fluorescent molecules. The result is a variation of measured fluorescence lifetime with wavelength. Thus, for a more effective use of fluorescence lifetime(s) information, measurements of spectrally resolved fluorescence decay characteristics are required. This has been typically achieved through the use of either a scanning monochromator⁴ or a set of bandpass filters⁵ that can resolve a sample's emission spectrum sequentially before each spectral intensity is time-resolved. However, recording of fluorescence decays at multiple wavelengths is inherently associated with long data acquisition time-a fact that limits the applicability of this technique to point-static measurements.

New research has produced a solution for rapid and simultaneous recording of fluorescence decays in multiple spectral bands compatible with in vivo applications.⁶ In this configuration, the emission spectrum is resolved using a set of dichroic and bandpass filters, allowing for selection of four channels or wavelength bands (see Fig. 1). Each channel is coupled to optical fibers of distinct lengths that act as optical delay lines. Thus, the spectrally resolved fluorescence decays arrive at the detector separated in time (~50 ns).

This ms-TRFS device is based on pulse sampling and gated detection, and the detector remains on until it reads the entire fluorescence transient pulse. A digitizer with high sampling frequency and analog bandwidth (8 Gsamples/s, 3 GHz) records the pulse. If the sample has high quantum efficiency, a single excitation pulse is enough to simultaneously record a complete transient fluorescent pulse for each channel (wavelength band). This method allows fast recording of fluorescence decays (e.g., <1 μs per data point) in multiple spectral bands generated in response to a single excitation event. It also offers enhanced portability.

Importantly, this versatile system enables acquisition and analysis of fluorescence data in several operation modes. The same instrument can be used for point spectroscopy, for instance, and can also scan in line, planar (XY), and even with pull-back motion-rotational modes. The latter allows for recovery of spectroscopic or fluorescence lifetime imaging (FLIm), even in high spatial resolution. The spectral channels can be adapted to resolve the autofluorescence emission of key biomolecules in tissue, including collagen, elastin, nicotinamide adenine dinucleotide (NADH), lipopigments, flavins, and porphyrins.

FIGURE 2. (a) A tumor (5 × 5 mm area) is clearly evident in this photo of a hamster cheek. (b) Corresponding FLIm images (emission at 390 nm spectral band) were obtained from the measurements of the area pictured in (a). The tumor/lesion area shows a lower lifetime than the surrounding area because fluorescence emission from NADH dominates there, whereas collagen dominates in the normal epithelium. The axial and lateral pixel sizes for FLIM were 0.1 and 0.2 mm, respectively. (c) Fluorescence lifetime histogram of manually segmented tumor area vs. surrounding area.

The feasibility of the ms-TRFS approach has been tested in preclinical settings, including an in vivo animal model of oral carcinoma and explanted human coronary arteries (see Figs. 2 and 3).^7,8 The system has proven suitable for fast, accurate, and precise lifetime measurement of low-quantum-efficiency, sub-nanosecond lifetime fluorophores.

FIGURE 3. (a) Normalized fluorescence intensity and (b) the corresponding lifetime values (452±22 nm emission band) as a fly-through representation in a human coronary artery. (c) Three-dimensional reconstruction of the lumen surface along the entire pull-back sequence. Areas corresponding to collagen-rich plaques were characterized by higher lifetime values when compared with lipid-rich areas.

Wide-field FLIM enables endoscopy

Traditional white-light endoscopes, commonly used in a broad range of clinical applications, enable their operators to see inside the body—but are unable to characterize a tissue's molecular makeup based on its autofluorescence signatures. Adding this ability, however, would improve the diagnostic values of endoscopic procedures.

FIGURE 4. A performance evaluation of a fluorescence lifetime imaging microscopy system using fluorophores and biomolecules shows (a) fluorescence intensity images and (b) average lifetime images of the coumarin 120 (C120) and 9-cyanoanthracene (9CA) dye solutions in capillaries placed on top of a dry collagen bed. The images were acquired using the 460/50 nm filter. (c) Image resolution was evaluated using a USAF test chart at a 0.2 ns gating time. (d) Fluorescence lifetime histograms of the C120, 9CA, and collagen correspond to images in (b).

We have demonstrated that wide-field FLIM can be conducted in vivo in patients. A flexible optical endoscope uses coherent fiber bundles (image guides) to relay imagery from the sample to a time-gated camera. Compared to rigid rod-lens endoscopy, the approach generates lesser image quality and intensity—but rigid systems cannot be applied to intravascular imaging or other evaluations that require flexibility. Figure 4 demonstrates the ability of an endoscopic FLIM system to resolve the distinct fluorescence lifetimes.

FIGURE 5. The imaging bundle probe is depicted as part of (a) the overall FLIM system and (b) in a separate photograph. The probe's distal end includes an optically transparent spacer. The imaging probe has been applied to (c) the oral cavity, and (d) the brain cortex in human subjects.

Recent research has demonstrated the clinical feasibility of a fiber endoscopic FLIM system. The studies were conducted in 1) a neurosurgical setting for intraoperative observation and characterization of brain tissue autofluorescence, and 2) during head and neck surgery for intraoperative delineation of tumor margins (see Fig. 5).^9,10

FIGURE 6. A normal cortex (a) is distinct from tumor tissue (b) in fluorescence lifetime imaging of human patients. Results are for the 460 nm emission band (NADH fluorescence).

The results showed that, for patients undergoing craniotomy and resection of glioblastoma multiforme (GBM), fluorescence lifetime contrast can be achieved between tumor sites and normal cortex (see Fig. 6).

Similarly, in human patients with head and neck cancer, the system demonstrated potential to noninvasively differentiate tumor and normal tissue (see Fig. 7). These studies open new pathways for intraoperative diagnosis of cancer.

FIGURE 7. Representative intensity images generated in vivo in head and neck cancer patients show (a) normal and (b) tumor tissue; corresponding lifetime images are (c) normal and (d) tumor. Histograms compare (e) intensity and (f) lifetime. Results are for 460 nm emission band (NADH fluorescence).

What are the needs?

As discussed, time-resolved fluorescence spectroscopy and imaging systems achieve label-free optical molecular contrast in diseased tissues. And clinical studies conducted with instrumentation based on this principle show that fluorescence lifetime contrast carries diagnostic value. However, the total number of clinical studies using fluorescence lifetime technologies remains sparse.

The availability of a number of affordable components promises to facilitate further development of compact and practical ms-TRFS and FLIM systems. Among these are affordable sub-nanosecond pulsed UV laser light sources, fast-response and sensitive detectors, and new solutions for fast recording of spectrally resolved fluorescence lifetime information. Advancement of analytical methods for fast processing of the fluorescence intensity decay data is another possible enabler. Such solutions may spawn a new generation of practical devices for clinical research studies and systematic testing of their diagnostic potential.

REFERENCES

1. L. Marcu, Annals Biomed. Eng., 40, 2, 304-331 (2012).

2. R. Richards-Kortum and E. Sevick-Muraca, Annu. Rev. Phys. Chem., 47, 555-606 (1996).

3. J. R. Lakowicz, Principles of fluorescence spectroscopy, 3, Kluwer Academic/Plenum, New York (2006).

4. Q. Fang et al., Rev. Sci. Instrum., 75, 151-162 (2004).

5. Y. Sun et al., Opt Lett., 33, 6, 630-632 (2008).

6. D. R. Yankelevich et al., Rev. Sci. Instrum., 85, 3, 034303 (2014); doi:10.1063/1.4869037.

7. H. Fatakdawala et al., Biomed. Opt. Exp., 4, 9, 1724-1741 (2013).

8. D. Gorpas et al., "Bi-modal imaging of atherosclerotic plaques: Automated method for co-registration between fluorescence lifetime imaging and intravascular ultrasound data," Proc. SPIE, 8926, 892638 (2014).

9. Y. Sun et al., Microsc. Microanal., 19, 4, 791-798 (2013).

10. Y. Sun et al., J. Biomed. Opt., 15, 5, 056022 (2010).