Academia.eduAcademia.edu

Figure 9 – uploaded by James McGinty

See full PDFdownloadDownload figure
Figure 22.9. H&E images and FLIM maps of different regions of healthy and diseased human cartilage. The FLIM maps were recorded for the areas highlighted on the H&E images. Regions A and B are from thin and thick regions of healthy cartilage respectively. C contains the frayed cartilage surface and subchondral bone, D contains a fissure with some regions of short lifetime, (See color insert section.)

Figure 22 9. H&E images and FLIM maps of different regions of healthy and diseased human cartilage. The FLIM maps were recorded for the areas highlighted on the H&E images. Regions A and B are from thin and thick regions of healthy cartilage respectively. C contains the frayed cartilage surface and subchondral bone, D contains a fissure with some regions of short lifetime, (See color insert section.)

Related Figures (21)

Figure 22.1. (a) simplified molecular energy level diagram highlighting the spectro-temporal properties of fluorescence and (b) single exponential fluorescence decay plot. The parameter k,, contains contributions from a number of non-radiative processes including quenching, energy transfer, and solvent interactions. To generalise, the radiative decay constant depends mainly on the molecular energy structure and the refractive index of the surrounding medium,’*"° whereas the non-radiative decay constant also has a contribution from the immediate environment. Therefore fluorescence lifetime can report on different molecular species (fluorophores) present in the sample and on variations in the local fluorophore environment.
Figure 22.1. (a) simplified molecular energy level diagram highlighting the spectro-temporal properties of fluorescence and (b) single exponential fluorescence decay plot. The parameter k,, contains contributions from a number of non-radiative processes including quenching, energy transfer, and solvent interactions. To generalise, the radiative decay constant depends mainly on the molecular energy structure and the refractive index of the surrounding medium,’*"° whereas the non-radiative decay constant also has a contribution from the immediate environment. Therefore fluorescence lifetime can report on different molecular species (fluorophores) present in the sample and on variations in the local fluorophore environment.
If the fluorescence decay is not a single exponential, however, the lifetimes calculated from these equations will not be the same and the experiment must be repeated at different excitation modulation frequencies to build up a full (multi-exponential) picture. This is usually calculated by fitting the results to a set of dispersion relationships,***° although the increased acquisition and data processing time can be disadvantageous. For simplicity it is possible to use only one modulation frequency and calculate an average fluorescence lifetime by taking the average of the two lifetimes found from Egs. (4)-(5). 4s 44 4 Tr lhe
If the fluorescence decay is not a single exponential, however, the lifetimes calculated from these equations will not be the same and the experiment must be repeated at different excitation modulation frequencies to build up a full (multi-exponential) picture. This is usually calculated by fitting the results to a set of dispersion relationships,***° although the increased acquisition and data processing time can be disadvantageous. For simplicity it is possible to use only one modulation frequency and calculate an average fluorescence lifetime by taking the average of the two lifetimes found from Egs. (4)-(5). 4s 44 4 Tr lhe
Figure 22.3. Techniques for FLIM implemented in scanning microscopes MULTIDIMENSIONAL FLUORESCENCE IMAGING APPLIED TO BIOLOGICAL TISSUE 22.3.2. Fluorescence Lifetime Imaging (FLIM)
Figure 22.3. Techniques for FLIM implemented in scanning microscopes MULTIDIMENSIONAL FLUORESCENCE IMAGING APPLIED TO BIOLOGICAL TISSUE 22.3.2. Fluorescence Lifetime Imaging (FLIM)
This model has only one additional fitting parameter, h, compared with the single exponential decay and can provide a reasonable representation of real biological decay data.” The physical significance of h is that it corresponds to the width of the lifetime distribution, and may itself be a useful contrast parameter in situations where the local fluorophore environment is complex or changing. See reference for a comparison of the single, double and stretched exponential decay models applied to tissue autofluorescence data. Further examples of the use of this function will be explored in section 22.0 of this chanter. showins nure endosenous fluorescence results imaged in
This model has only one additional fitting parameter, h, compared with the single exponential decay and can provide a reasonable representation of real biological decay data.” The physical significance of h is that it corresponds to the width of the lifetime distribution, and may itself be a useful contrast parameter in situations where the local fluorophore environment is complex or changing. See reference for a comparison of the single, double and stretched exponential decay models applied to tissue autofluorescence data. Further examples of the use of this function will be explored in section 22.0 of this chanter. showins nure endosenous fluorescence results imaged in
Figure 22.7. Endogenous fluorophores and the StrEF. (a) shows the fluorescence lifetime map of the five fluorophores and the lifetime histogram across all pixels is shown in (b). The stretched exponential and single exponential fits and residuals for the collagen decay are compared in (c) and (d). (See color insert section.) sections provide an accessible opportunity to investigate this contrast since problems associated with tissue scattering, background fluorescence and other practical issues are minimized. Furthermore, fixed or frozen sections are the most convenient way to acquire and transport clinical tissue samples. It is well known however, that the fluorescence lifetime can be affected by the local fluorophore environment, and that the fluorescence generated from prepared ex vivo tissue can provide only an approximate picture of that which may be observed in vivo. We have recently used a Krumdieck tissue slicer which has shown greater contrast consistent with living tissue. However, comparison with artificially fixed tissue is a necessary precursor to in vivo imaging and the opportunity to compare high magnification FLIM images with serial sections that have been stained with conventional histopathological preparations (e.g. Haematoxylin and Eosin - H&E) is invaluable to discover the structural and chemical composition.
Figure 22.7. Endogenous fluorophores and the StrEF. (a) shows the fluorescence lifetime map of the five fluorophores and the lifetime histogram across all pixels is shown in (b). The stretched exponential and single exponential fits and residuals for the collagen decay are compared in (c) and (d). (See color insert section.) sections provide an accessible opportunity to investigate this contrast since problems associated with tissue scattering, background fluorescence and other practical issues are minimized. Furthermore, fixed or frozen sections are the most convenient way to acquire and transport clinical tissue samples. It is well known however, that the fluorescence lifetime can be affected by the local fluorophore environment, and that the fluorescence generated from prepared ex vivo tissue can provide only an approximate picture of that which may be observed in vivo. We have recently used a Krumdieck tissue slicer which has shown greater contrast consistent with living tissue. However, comparison with artificially fixed tissue is a necessary precursor to in vivo imaging and the opportunity to compare high magnification FLIM images with serial sections that have been stained with conventional histopathological preparations (e.g. Haematoxylin and Eosin - H&E) is invaluable to discover the structural and chemical composition.
t = 800-2600 ps, formalin-fixed human pancreas, 10 jm section. t= 1000-2500 ps, fresh mouse kidney, 200 11m section. Figure 22.8. Examples of FLIM images of fixed, frozen and fresh tissue. (See color insert section.)
t = 800-2600 ps, formalin-fixed human pancreas, 10 jm section. t= 1000-2500 ps, fresh mouse kidney, 200 11m section. Figure 22.8. Examples of FLIM images of fixed, frozen and fresh tissue. (See color insert section.)
aaiiul VoOVVPY SLIMY OS. The samples were typically atherosclerotic lesions were identified by the thickening and irregu he intima. We observed that it was possible to differentiate between atherosclerotic lesions and artery wall using FLIM when exciting from 355-440 nm. The strongest lifetime contrast be nm the intima was also distinguishab imaged at xl0 magnification and ween the artery wall and a e. For comparison, we a ar surface of herosclerotic esions was obtained for 355 nm excitation, illustrated in figure 22.10, although contrast was still observed for excitation wavelengths up to 440 nm, and at 400 so examined hese samples with conventional spectrally-resolved fluorescence emission imaging. Using multi-spectral analysis we found that the level of contrast obtained using the fluorescence life excitation wavelengths. ime could not be recreated for these from 2 month old male New Zealand White (NZW) rabbits. The rabbi supplemented diet with 1 % (w/w) cholesterol (Sigma) for 8 weeks. Fo s had a lowing sacrifice, the abdominal aorta was formalin fixed in situ, excised, then postfixed also in 10 % buffered formalin and stored at 4°C. Prior to the imaging experiments, the aorta was soaked in saline buffer solution at a neutra pH in order to remove any formaldehyde. 10 ym thick cross-sections of aorta were cut at 30°C using an embedding compound (Cryo-M-Bed) and mounted on microscope slides. "hin entanlas -<axraern txreinalixs duegmarent int wifi erannifinotinx 0 glass ns and
aaiiul VoOVVPY SLIMY OS. The samples were typically atherosclerotic lesions were identified by the thickening and irregu he intima. We observed that it was possible to differentiate between atherosclerotic lesions and artery wall using FLIM when exciting from 355-440 nm. The strongest lifetime contrast be nm the intima was also distinguishab imaged at xl0 magnification and ween the artery wall and a e. For comparison, we a ar surface of herosclerotic esions was obtained for 355 nm excitation, illustrated in figure 22.10, although contrast was still observed for excitation wavelengths up to 440 nm, and at 400 so examined hese samples with conventional spectrally-resolved fluorescence emission imaging. Using multi-spectral analysis we found that the level of contrast obtained using the fluorescence life excitation wavelengths. ime could not be recreated for these from 2 month old male New Zealand White (NZW) rabbits. The rabbi supplemented diet with 1 % (w/w) cholesterol (Sigma) for 8 weeks. Fo s had a lowing sacrifice, the abdominal aorta was formalin fixed in situ, excised, then postfixed also in 10 % buffered formalin and stored at 4°C. Prior to the imaging experiments, the aorta was soaked in saline buffer solution at a neutra pH in order to remove any formaldehyde. 10 ym thick cross-sections of aorta were cut at 30°C using an embedding compound (Cryo-M-Bed) and mounted on microscope slides. "hin entanlas -<axraern txreinalixs duegmarent int wifi erannifinotinx 0 glass ns and
Figure 22.12. Wide-field time-gated FLIM of a formalin-fixed 10 pm-thick tissue section of pancreas showing partial invasion of an Islet of Langerhans (dotted line) by malignant cells (solid line and arrows). (a) Light microscope image of an H+E stained section; (b) False colour FLIM map; (c) Intensity weighted FLIM image. FLIM reveals the uninvaded portion of the islet to be contrasted to both the invaded area and the surrounding malignant parenchyma due to its shorter fluorescence lifetime. There is however no apparent contrast between the malignant parenchyma and the normal pancreatic parenchyma. (Excitation wavelength = 400nm, detection range > 435 nm, microscope objective = x40). (See color insert section.)
Figure 22.12. Wide-field time-gated FLIM of a formalin-fixed 10 pm-thick tissue section of pancreas showing partial invasion of an Islet of Langerhans (dotted line) by malignant cells (solid line and arrows). (a) Light microscope image of an H+E stained section; (b) False colour FLIM map; (c) Intensity weighted FLIM image. FLIM reveals the uninvaded portion of the islet to be contrasted to both the invaded area and the surrounding malignant parenchyma due to its shorter fluorescence lifetime. There is however no apparent contrast between the malignant parenchyma and the normal pancreatic parenchyma. (Excitation wavelength = 400nm, detection range > 435 nm, microscope objective = x40). (See color insert section.)
Our initial work on endoscopic FLIM involved applying the wide-field time-gated FLIM system described in section 22.0 through a flexible fibre bundle endoscope.® While this approach demonstrated hat endoscopic FLIM was possible, the low frame-rate, sensitivity and sub-optimal design made it impractical for clinical experiments. Our recent work has focussed on applying the real-time (sequential) FLIM approach described in t rigid and flexible optical endoscopes.°*'?° Rigid end he previous section to loscopes, such as the arthroscopes used for orthopaedic applications, relay the image through a series of rod lenses and provide relatively high throughou and image quali Flexible optical endoscopes utilize fibre bundles to relay the image, resulting loss of image quality and intensity compared to rigid endoscopes. endoscopes are usually longer (up to several metres) than arthroscopes and t resulting increase in group velocity dispersion (GVD) can temporally stretch fluorescence signals. For both rigid and flexible endoscopes, we general separate low GVD optical fibre to bring the excitation light to the sample at t distal end of the endoscope. The excitation light for our endoscopic FLIM laser (Spectra Physics Inc, Vanguard 350-HMD355), which produces 10 ps pulses at 355 nm at a e output power. This experiments is provided by a frequency tripled Nd: Y VO, repetition rate of 80 MHz , with up to ~350 mW averag laser is relatively compact (60 x 15 x 10 cm) with a portable power sup y: Flexible ne ply that
Our initial work on endoscopic FLIM involved applying the wide-field time-gated FLIM system described in section 22.0 through a flexible fibre bundle endoscope.® While this approach demonstrated hat endoscopic FLIM was possible, the low frame-rate, sensitivity and sub-optimal design made it impractical for clinical experiments. Our recent work has focussed on applying the real-time (sequential) FLIM approach described in t rigid and flexible optical endoscopes.°*'?° Rigid end he previous section to loscopes, such as the arthroscopes used for orthopaedic applications, relay the image through a series of rod lenses and provide relatively high throughou and image quali Flexible optical endoscopes utilize fibre bundles to relay the image, resulting loss of image quality and intensity compared to rigid endoscopes. endoscopes are usually longer (up to several metres) than arthroscopes and t resulting increase in group velocity dispersion (GVD) can temporally stretch fluorescence signals. For both rigid and flexible endoscopes, we general separate low GVD optical fibre to bring the excitation light to the sample at t distal end of the endoscope. The excitation light for our endoscopic FLIM laser (Spectra Physics Inc, Vanguard 350-HMD355), which produces 10 ps pulses at 355 nm at a e output power. This experiments is provided by a frequency tripled Nd: Y VO, repetition rate of 80 MHz , with up to ~350 mW averag laser is relatively compact (60 x 15 x 10 cm) with a portable power sup y: Flexible ne ply that
Figure 22.14. (a) Arthroscope used to obtain endoscopic FLIM images of bisected lambs kidney and FLIM images acquired using (b) conventional WNLLS at 0.1Hz and (c) real time (7.2 Hz) FLIM acquisition and processing. (See color insert section.) does not require water cooling. An optical attenuator, consisting of a half wave plate and polarizing beam splitter, is used to control the excitation power. For our first demonstration of endoscopic FLIM using a rigid endoscope, the sample was directly illuminated, using a holographic diffuser to produce a cone of illumination with a full angle of ~60°. The fluorescence emission was imaged through the conventional (Smith and Nephew), arthroscope pictured in figure 22.14 (a) , passed through a 435 nm long pass filter to block any reflected or scattered excitation light, and imaged onto the photocathode of the GOI A “real-time” FLIM image of the autofluorescence from a sample of lamb’s kidney acquired at an update rate of 7.2 Hz is presented in figure 22.14 (c). For comparison, a conventional wide-field FLIM image of the same tissue (eight 400 ps time gates acquired over 15 seconds and fitted using a weighted non- linear least squares iterative algorithm) is also shown. It will be seen that the “live” real-time image is noisier than that obtained with the conventional acquisition but the fluorescence lifetime contrast, particularly between the calyces and the medulla, is still apparent. We note that the “live” FLIM images appear less nosy that this “snapshot” because the human eye performs a degree of real-time image processing.
Figure 22.14. (a) Arthroscope used to obtain endoscopic FLIM images of bisected lambs kidney and FLIM images acquired using (b) conventional WNLLS at 0.1Hz and (c) real time (7.2 Hz) FLIM acquisition and processing. (See color insert section.) does not require water cooling. An optical attenuator, consisting of a half wave plate and polarizing beam splitter, is used to control the excitation power. For our first demonstration of endoscopic FLIM using a rigid endoscope, the sample was directly illuminated, using a holographic diffuser to produce a cone of illumination with a full angle of ~60°. The fluorescence emission was imaged through the conventional (Smith and Nephew), arthroscope pictured in figure 22.14 (a) , passed through a 435 nm long pass filter to block any reflected or scattered excitation light, and imaged onto the photocathode of the GOI A “real-time” FLIM image of the autofluorescence from a sample of lamb’s kidney acquired at an update rate of 7.2 Hz is presented in figure 22.14 (c). For comparison, a conventional wide-field FLIM image of the same tissue (eight 400 ps time gates acquired over 15 seconds and fitted using a weighted non- linear least squares iterative algorithm) is also shown. It will be seen that the “live” real-time image is noisier than that obtained with the conventional acquisition but the fluorescence lifetime contrast, particularly between the calyces and the medulla, is still apparent. We note that the “live” FLIM images appear less nosy that this “snapshot” because the human eye performs a degree of real-time image processing.
Real-time FLIM has also been realised with the flexible endoscope shown in figure 22.15 (a). Even though the coupling and transmission losses of the fibre imaging bundle are significantly higher than for the rigid endoscope, this system is still capable of producing FLIM images of the endogenous fluorescence of fresh tissue at frame rates of several Hz. For this system the excitation light was coupled into the lower order modes of a 7 m long multimode fibre chosen for low intrinsic fluorescence when transmitting 355 nm light. This excitation fibre was then introduced through the biopsy port of a flexible endoscope and a light shaping diffuser was used at the distal end to produce a cone of illumination with a full angle of ~60°. This was larger than the imaging angle of the endoscope (50°) and completely filled the field of view, providing a reasonably uniform illumination. The imaging channel of the flexible endoscope comprised a graded index objective cemented to a flexible coherent imaging bundle with ~30,000 fibres. The output from the endoscope was imaged directly onto the photocathode of the gated optical imager (GOT) through a 435 nm long pass filter. Figure 22.15 (b) shows a fluorescence lifetime image of the bisected lamb’s kidney acquired at an update rate of 5.5 Tv.
Real-time FLIM has also been realised with the flexible endoscope shown in figure 22.15 (a). Even though the coupling and transmission losses of the fibre imaging bundle are significantly higher than for the rigid endoscope, this system is still capable of producing FLIM images of the endogenous fluorescence of fresh tissue at frame rates of several Hz. For this system the excitation light was coupled into the lower order modes of a 7 m long multimode fibre chosen for low intrinsic fluorescence when transmitting 355 nm light. This excitation fibre was then introduced through the biopsy port of a flexible endoscope and a light shaping diffuser was used at the distal end to produce a cone of illumination with a full angle of ~60°. This was larger than the imaging angle of the endoscope (50°) and completely filled the field of view, providing a reasonably uniform illumination. The imaging channel of the flexible endoscope comprised a graded index objective cemented to a flexible coherent imaging bundle with ~30,000 fibres. The output from the endoscope was imaged directly onto the photocathode of the gated optical imager (GOT) through a 435 nm long pass filter. Figure 22.15 (b) shows a fluorescence lifetime image of the bisected lamb’s kidney acquired at an update rate of 5.5 Tv.
Figure 22.16. Experimental setup. PBS1-3 polarizing beam splitters, FR Faraday rotator, 4/2 half wave plates, L1-5 lenses, M1-4 mirrors, P prism, BD beam dump. Mirror M3 has a triangular-shaped slit attached directly onto the front of the mirror. Inset: output spectrum from fibre plotted on both linear and logarithmic scales. Reproduced from Journal of Physics D: Applied Physics © 2004 IOP Publishing Ltd,
Figure 22.16. Experimental setup. PBS1-3 polarizing beam splitters, FR Faraday rotator, 4/2 half wave plates, L1-5 lenses, M1-4 mirrors, P prism, BD beam dump. Mirror M3 has a triangular-shaped slit attached directly onto the front of the mirror. Inset: output spectrum from fibre plotted on both linear and logarithmic scales. Reproduced from Journal of Physics D: Applied Physics © 2004 IOP Publishing Ltd,
Figure 22.17. (a) Fluorescence intensity (inset: transmitted light) image (26x26 jm) and (b) in situ measurement of the total (relative) fluorescence intensity of GFP in a human B cell (with GFP-labelled GPI) for a small region of the cell membrane as a function of excitation wavelength (lines joining data points are to guide the eye only); (c) intensity merged FLIM image of a GFP labelled cell excited at 480- 500 nm using the TCS. Reproduced from Journal of Physics D: Applied Physics'”° © 2004 10P Publishing Ltd
Figure 22.17. (a) Fluorescence intensity (inset: transmitted light) image (26x26 jm) and (b) in situ measurement of the total (relative) fluorescence intensity of GFP in a human B cell (with GFP-labelled GPI) for a small region of the cell membrane as a function of excitation wavelength (lines joining data points are to guide the eye only); (c) intensity merged FLIM image of a GFP labelled cell excited at 480- 500 nm using the TCS. Reproduced from Journal of Physics D: Applied Physics'”° © 2004 10P Publishing Ltd
Figure 22.18. (a) Fluorescence intensity and (b) fluorescence lifetime images of a fixed unstained 10 pm section of human pancreas obtained in a wide-field time-domain FLIM system with Aex = 440-450 nm (A - small artery, I - islet of Langerhans and C - connective tissue). (c) composite wide-field fluorescence intensity and FLIM images of a multiwell plate array of fluorescent dyes excited with four different wavelength bands of 20 nm width centred on 500, 515, 570 and 610 nm respectively. The upper part shows the first time-gated fluorescence intensity image and the lower part shows the corresponding FLIM images. Reproduced from Journal of Physics D: Applied Physics'*° © 2004 IOP Publishing Ltd. (See color insert section.)
Figure 22.18. (a) Fluorescence intensity and (b) fluorescence lifetime images of a fixed unstained 10 pm section of human pancreas obtained in a wide-field time-domain FLIM system with Aex = 440-450 nm (A - small artery, I - islet of Langerhans and C - connective tissue). (c) composite wide-field fluorescence intensity and FLIM images of a multiwell plate array of fluorescent dyes excited with four different wavelength bands of 20 nm width centred on 500, 515, 570 and 610 nm respectively. The upper part shows the first time-gated fluorescence intensity image and the lower part shows the corresponding FLIM images. Reproduced from Journal of Physics D: Applied Physics'*° © 2004 IOP Publishing Ltd. (See color insert section.)
Figure 22.20. Generalised experimental set-up for hyperspectral FLIM microscopy.LlandL2 are anamorphic optics to produce the slit illumination for single photon excitation; L3 represents the imaging optics; IS the imaging spectrograph and GOI the gated optical image intensifier.
Figure 22.20. Generalised experimental set-up for hyperspectral FLIM microscopy.LlandL2 are anamorphic optics to produce the slit illumination for single photon excitation; L3 represents the imaging optics; IS the imaging spectrograph and GOI the gated optical image intensifier.
Figure 22.21. FLIM images calculated from excitation-emission-lifetime data set for (a) 480 nm excitation, 490-620 nm emission band; (b) 550 nm excitation, 490-620 nm emission band and (c) 480 nm excitation, 620-835 nm emission band. (See color insert section.) For multiphoton microscopy, we use a commercial multi-photon multi-for microscope (TriMScope, LaVision Biotec GmbH, Bielefeld, Germany), which can scan up to 64 excitation spots from a mode-locked Ti:sapphire las to produce a line excitation (as well as being able to rapidly scan the whole fie! of view for wide-field detection). In line-scanning mode, we use a motorise translation stage to scan the sample through the line excitation and image tk resulting fluorescence onto the entrance slit of an imaging spectrograph (SI 300i, Acton Research Corp., Acton MA). At the output of the spectrograph w use our standard wide-field time-gated FLIM system based on a GOI with CC! readout. Using a Ti:Sapphire laser that provides 100 fs pulses at 80 MHz wit ~ 1 W of average output power at 780 nm, we can typically acquire hyperspectral FLIM data set in a few minutes.
Figure 22.21. FLIM images calculated from excitation-emission-lifetime data set for (a) 480 nm excitation, 490-620 nm emission band; (b) 550 nm excitation, 490-620 nm emission band and (c) 480 nm excitation, 620-835 nm emission band. (See color insert section.) For multiphoton microscopy, we use a commercial multi-photon multi-for microscope (TriMScope, LaVision Biotec GmbH, Bielefeld, Germany), which can scan up to 64 excitation spots from a mode-locked Ti:sapphire las to produce a line excitation (as well as being able to rapidly scan the whole fie! of view for wide-field detection). In line-scanning mode, we use a motorise translation stage to scan the sample through the line excitation and image tk resulting fluorescence onto the entrance slit of an imaging spectrograph (SI 300i, Acton Research Corp., Acton MA). At the output of the spectrograph w use our standard wide-field time-gated FLIM system based on a GOI with CC! readout. Using a Ti:Sapphire laser that provides 100 fs pulses at 80 MHz wit ~ 1 W of average output power at 780 nm, we can typically acquire hyperspectral FLIM data set in a few minutes.
Figure 22.22. Fluorescence lifetime images of carotid artery in emission bands (a) 415-455 nm and (b) 570-635 nm. (c) shows the mean wavelength and (d) shows the spectral width as a function of the delay time after excitation for the lipid rich and fibrin rich regions highlighted in (a). (See color insert section.)
Figure 22.22. Fluorescence lifetime images of carotid artery in emission bands (a) 415-455 nm and (b) 570-635 nm. (c) shows the mean wavelength and (d) shows the spectral width as a function of the delay time after excitation for the lipid rich and fibrin rich regions highlighted in (a). (See color insert section.)
Related topics:
GeneticsMolecular BiologyMultiphoton MicroscopyBiomedical ImagingFluorescence LifetimeFluorescent probes
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved