3.1 STWRFS system verification on fluorescent biomolecules
Endogenous fluorophores (collagen and elastin) were measured to test the system STWRFS performance and compared with the results obtained with a scanning TR-LIFS system available in the lab previously reported [16
]. Fluorescence impulse response profiles were recorded from collagen and elastin powders at the three wavelength sub-bands as shown in . Average lifetimes of collagen were retrieved as 2.94 ± 0.12 and 2.85 ± 0.17 ns after deconvolution for filters F1 and F2, respectively. The signal from filter F3 was too weak for precise lifetime evaluation. Fluorescence lifetimes of elastin for F1, F2, and F3 were 1.72 ± 0.02, 1.81 ± 0.01, and 1.83 ± 0.03 ns, respectively. These values were in good agreement with those reported in the literature [1
3.2 STWRFS scanning performance – validation on ex-vivo tissue samples
Two representative results from scanning measurements made on ex vivo human atherosclerotic aorta using the STWRFS system with the fiber-optic catheter and the motorized stage are summarized in the following.
demonstrates the scanning procedure and results from one sample of atherosclerotic aorta characterized by heterogeneous composition and morphology. During experiments, aorta tissue was opened along the artery axial direction and scanned in the perpendicular direction with a scan length of around 30 mm. The color photograph in shows a human artery tissue with the scanned line used for the measurement marked. This tissue was determined to have three representative regions labeled in the picture as fibrotic plaque, early-stage plaque, and the surrounding normal tissue. This identification was based on visual inspection first and was further tested by histopathology analysis conducted by an independent pathologist according to a trichrome staining. Fibrotic plaque is usually considered to be a late stage lesion that is rich in collagen fibers, while early-stage plaques exhibit developing lipid and collagen accumulation [22
]. The intensities of the three sub-bands are plotted . During the scanning measurements, the fluorescence intensity fluctuated along the tissue surface. Although the intensity in the F1 channel is expected to vary depending on the tissue collagen content, the signal was also affected by the irregular morphology. Thus, the value of the intensity alone was not found reliable for tissue characterization.
Fig. 3 Example of dynamic scanning of aorta tissue: (a) The aorta was opened and scanned perpendicular to the axial direction. The different regions of tissue were marked with a color bar seen in (a) and (d). (b) STWRFS profiles recorded from a collagen-rich (more ...)
Further analysis of intensity ratios for distinct wavelength sub-bands and fluorescence decay characteristics within this sub-bands demonstrated that two sets of parameters show a clear association with tissue types and chemical compositions without being affected by the scanning and irregular morphology. These were the intensity ratio of F1 to F2 (IF1/IF2) and the fluorescence lifetime of F1 (τF1, 1/e decay time of fluorescence impulse response). demonstrates the IF1/IF2 and τF1 along the scanning line shown in . From left to right this scan traversed a fibrotic plaque, a normal area, and then another fibrotic plaque, which were independently identified based on histopathology analysis. The mean of average lifetimes (τF1) of the three regions were 2.77 ± 0.18, 1.84 ± 0.08, and 2.33 ± 0.04 ns, respectively. In addition, the shoulder in this data next to the second peak was consistent with an early-stage plaque (intermediate lesion), with a lifetime of 2.57 ± 0.05 ns. It was found that the fluorescence lifetime increased by 51% for atherosclerotic plaque compared to the surrounding normal tissue. The average intensity ratios (IF1/IF2) in the three regions are 1.26 ± 0.15, 0.67 ± 0.05, and 1.3 ± 0.14, respectively. The intensity ratio rose by 94% in late-stage plaque in contrast to the normal tissue. The early-stage plaque has the intensity ratio at 1.08 ± 0.03. Both IF1/IF2 and τF1 allowed clear visualization of the three regions consistent with the tissue types.
A second scanning measurement of a different human aorta sample is summarized in
. This sample included three regions, classified as (from left to right): surrounding normal, early-stage fibrolipidic plaque, and fibrotic plaque with calcified necrotic core, as identified by histopathology results independently from the fluorescence measurement. The three regions were clearly discriminated by IF1/
as seen in . The average lifetimes (τF1
) were also distinctive: 1.88 ± 0.11, 2.37 ± 0.07 and 2.74 ± 0.08 ns, respectively. Here the lifetime increased 46% for fibrotic plaque compared with the normal tissue. The average intensity ratio (IF1/
) also increased in the three regions from 0.75 ± 0.08, to 1.44 ± 0.08, and then to 1.55 ± 0.32 respectively. The intensity ratio rose over 107% in the third stage. These trends are in agreement with our earlier publication [11
] reporting on the time-resolved fluorescence characteristics of arterial beds including atherosclerotic aorta.
Fig. 4 Comparison of STWRFS results with pathology analysis and ultrasound imaging: (a) Close-up photograph of the tissue scanned through three regions: surrounding normal, early-stage plaque, and fibrotic plaque. The arrow indicates the scanning direction. (more ...)
This second aorta sample was also used to demonstrate the corroboration of fluorescence data with the UBM image (obtained with an unfocused single element 45 MHz transducer) used to visualize the tissue structural features (axial and lateral resolutions 38 and 200 μm, respectively) in the cross section under the scanned line, as shown in . The image shows the thickened intimal layer at the top of the tissue cross section that corresponds to the early stage and fibrotic plaques areas. The region corresponding to the advanced fibrotic plaque also exhibited an increased hyperechoic region in the top layers corresponding to the dense collagen structure and ultrasound reflection (acoustic shadow) indicating the presence of a calcified necrotic core underneath the fibrotic cap. These features were also validated by histopathological analysis of the sample.
displays the representative images of three types of tissue from trichrome staining histopathology slides used to validate the STWRFS and UBM measurements: A - normal, B - early-stage plaque with intracellular lipid accumulation collagen fiber formation, C - fibrotic plaque with lipid core and fibrotic/calcific regions. Trichrome staining is a common method for differentiating structural proteins (collagen, elastin fibers), smooth muscle cells, and to distinguish these from the other lipidic and calcified structures in plaques. It can be observed that the main component in normal tissue (A-area) is elastin as depicted by the dark coloration lines in the media of the healthy aorta. Within the 250 μm arterial penetration depth estimated at 337 nm [12
], the thickened intima of early stage plaque (B-area) contain a combination of collagen (green) and lipid-rich areas including macrophages and lipid components (white areas). In fibrotic plaque (C-area) the whole cross section is rich in collagen fibers. Part of a large necrotic cores is also visible on the button left of the tissue histology section.
The fluorescence results are seen to be consistent with tissue composition because both intensity ratios and fluorescence lifetimes vary depending on the ratio of collagen (F1) to elastin (F2). The elastin was the main endogenous fluorophore in healthy artery wall while the composition of collagen gradually increased with the progression of atherosclerosis [11
]. Plaque tissue with more collagen showed a higher IF1/F2
value compared with the normal artery. Furthermore, τF1
was dominated by the collagen lifetime component (2.5 to 5 ns), which was found longer than that of elastin (~1.5 ns). Therefore, the plaque showed a relatively long lifetime compared with normal tissue. The shorter fluorescence lifetime of early stage plaque (rich in collagen and lipids) relative to fibrotic plaques (primarily rich in collagen) denotes the contribution of short-lived fluorescent lipid components (~1 ns) to the overall fluorescence decay dynamics.
3.3 STWRFS scanning performance – validation in vivo
The scanning (catheter pullback sequence) experiments conducted intravascularly in vivo in swine arteries under IVUS guidance were used to evaluate the ability of the STWRFS apparatus to record robust fluorescence lifetime data in conditions that mimic conventional catheterization procedures including IVUS. The in vivo intravascular STWRFS application faced several challenges including the limitation on the diameter of optical probe, attenuation of fluorescence intensity by blood absorption, operation of the catheter in pulsatile blood flow, and the constraint for a rapid dynamic measurement. Therefore the requirement for light collection efficiency and probe positioning were much stricter than other types of application on open tissue surfaces.
Representative results from the still-mode (control) measurements are displayed in
, , and showing the intensity, intensity ratio, and average lifetime at different times. Each data set was recorded over a one second time interval at the same location. The fluorescence intensity showed weak variation between different pulses but the intensity ratio and lifetime were found to be constant, as shown in , and . The intensity ratios of IF1/IF2, IF1/IF3, and IF2/IF3 were 0.42 ± 0.01, 1.49 ± 0.02, 3.57 ± 0.05, respectively. The average fluorescence lifetimes were 1.60 ± 0.01 ns and 1.86 ± 0.05 ns for F1 and F2, respectively. Due to the low intensity values with F3 it was difficult to achieve a robust deconvolution of the fluorescence IRF, thus lifetime values for F3 are not presented.
Fig. 5 In vivo test of STWRFS in pig femoral artery: (a), (b), and (c) are the data in still mode − the probe was held at the same location (~1 second) for multiple measurements. (d), (e), and (f) are the results from a pull-back measurement when the (more ...)
Here both florescence intensity ratio and lifetime values were in a good agreement with fluorescence of the pig normal arterial wall that is rich in elastin [21
]. Blood absorption had previously been investigated since it was a major concern in the optical design for the illumination and fluorescence collection. 337 nm light has very limited penetration depth in blood, which also has strong absorption at the fluorescence wavelengths under investigation. It is possible for these factors to cause the loss or distortion of autofluorescence signal of artery when measured in vivo
in pulsatile blood flow. We determined that in still-mode all spectroscopic variables investigated here (intensities ratios and lifetimes) remained constant during the measurement. These results suggest that when the catheter is in contact with tissue the blood flow in the artery did not have an obvious effect on the autofluorescence measurement. It also indicates that the MMC catheter and STWRFS provide robust results at a very low flushing rate of 0.07-0.1 ml/s.
Representative results from the scanning-mode experiments are depicted in , , and for intensity, intensity ratio, and average lifetime at different positions along the artery wall. It was observed that the fluorescence intensity fluctuated significantly when the catheter moved along the lumen (). This effect is most likely a result of (1) changes in light excitation-emission geometry due to fluctuations of the distance between the probe and sidewall and changes in the lumen diameter; and (2) attenuation of light intensity due to diffusion of blood between fiber optic tip and arterial wall. Although more robust when compared to intensity values, the intensity ratio values for the three filters also fluctuated especially for IF1/IF3 (). Findings suggest that a ratiometric intensity measurement does not fully compensate for factors that affect the fluorescence intensity. The average intensity ratios of IF1/IF2, IF1/IF3, and IF2/IF3 are 0.49 ± 0.02, 1.76 ± 0.08, 3.58 ± 0.05, respectively.
The fluorescence lifetimes values for F1 and F2 were constant during pullback measurements () with values of 1.60 ± 0.03 ns and 1.82 ± 0.04 ns, respectively. The constant fluorescence lifetime indicated the composition of the artery wall along the lumen is similar, which was consistent with our expectation because the fluorescence of healthy young pig artery is primarily due to elastin emission as previously shown and confirmed by histology. Similar average fluorescence lifetime results depicting small standard deviation values (<0.05 ns) or intra-animal variability were obtained for each of the eight arterial segments linearly scanned in vivo. Moreover, the average fluorescence value (mean ± SD) for all independent measurements (total of 8 arterial segments) conducted in vivo in all animals was determined as 1.70 ± 0.16 for F1 and 1.8 ± 0.06 for F2. The larger SD value in this latter case is due to inter-animal mean lifetime value variability rather than intra-animal mean lifetime value changes. These findings also indicated that the MMC catheter and STWRFS when deployed in the close proximity of the vessel wall and pulled back provide robust results at a very low flushing rate of 0.07-0.1 ml/s during dynamic fluorescence recording.