3.1.

AF Spectra Obtained using a He–Cd Laser and UV LD

Figure 2 compares an AF image of the normal bronchus excited by the He–Cd laser (through the bronchoscope of the LIFE system) and the UV LD (direct irradiation). The image was obtained using the visualization system of the LIFE system for both excitation light sources through the bronchoscope. Images obtained using the two different excitation sources were very similar. Green bright AF was observed from the normal areas and a reddish color was observed from the tumor areas. This preliminary result could imply that the origin of the AF signal obtained by those sources is probably identical. Figure 3 shows a spectrum of a normal tissue sample excited by the He–Cd laser (442 nm) and the UV LD. The spectrum obtained using the UV LD was essentially the same as that obtained using the He–Cd laser. Much intense green and blue luminescence was observed when the He–Cd laser was used. The spectrum, however, a mix of the excitation beam (the multimode He–Cd laser beam) and the AF signal in the blue region (around 460–480 nm) was difficult to resolve. The intensity of the He–Cd laser of the LIFE system at the end of the bronchoscope was around 20 mW, and the intensity of the UV LD was about a few mW. The difference in excitation source power resulted in a difference in AF intensity. On the other hand, a well resolved peak at around 483 nm (blue) was clearly observed without the necessity of using any optical filters when the UV LD was used. The excitation light of the UV LD influenced only the region whose wavelength was shorter than 420 nm. Both samples showed a side peak at around 560 nm. The spectrum obtained from the tumor tissue by UV LD excitation (Figure 4) was also similar to the spectrum of tumor cells reported using other conventional methods.^{18 19 20} The intensity of the green (blue) fluorescence mostly disappeared when excited by the either He–Cd laser or the UV LD. The AF intensity was affected by the excitation source intensity, but the peak position of the green AF spectrum was not affected by the wavelength or the intensity of the excitation source used for this study.

Figure 2

Autofluorescence image of the human bronchus using (a) the He–Cd laser and (b) the UV LD.

Figure 3

Autofluorescence spectrum of normal tissue excited by (a) the He–Cd laser and (b) the UV LD.

Figure 4

Autofluorescence spectrum of tumor tissue excited by the UV LD.

3.2.

AF Signal Intensity versus Histopathological Findings

The AF signal intensity was measured for various regions of the human bronchus. Figure 5 shows an AF image of the bronchial bifurcation region. As is shown the wide area of the sample was irradiated by the UV LD beam. A green band pass filter was used to obtain the green AF image. An AF image obtained using the blue band pass filter was also taken and the image was similar to that obtained using the green band pass filter. As can be seen from Figure 5, the intensity of the green fluorescence showed spatial distributions; bright fluorescence was obtained from the region of bronchial bifurcation as well as from part of the bronchial wall along the airway where the elastin fiber bundle would exist. The signal intensity variation for various areas of the normal bronchial wall was measured and is given in Figure 6. As is shown, the AF intensity was affected by the density of the elastin fiber bundle, but the AF spectrum shape was not affected; a strong AF signal was observed in the sample which included a high density of elastin fiber bundles whereas the AF signal intensity was weak for areas where the bundle density was less.

Figure 5

AF image of the bronchus bifurcation.

Figure 6

Variation in signal intensity of the AF for various regions: (a) the region with a high density of elastic fiber bundles and (b) the region with less elastic fiber bundles.

Histopathological analysis was carried out along with spectrum analysis. Figure 7 shows histopathological findings of a saddle point region for the bronchial bifurcation. An adjacent region of the sample was sliced and stained by HE and EL, respectively. As can be seen from the cross-sectional view of the bifurcation, the existence of dense elastin in the saddle point region below the basement emblem was confirmed. Elastin was not clearly confirmed in the epithelium region. Figure 8 shows EL staining of the bifurcation for two different regions. The sample was sliced along with two different contours. The density of the elastin was high near the saddle point of the bifurcation, and the density became weak as the region became farther from the saddle point. The variation of the elastin density observed from these histopathological findings agreed well with the variation in intensity of the green AF shown in Figure 5; the region that has a high density of elastin seemed to show strong AF compared to regions that showed less density of elastin. There are several articles which describe the AF characteristics of elastin.^{5 11 15} The reported spectrum data associated with elastin were similar to the data obtained using the UV LD, but the peak position of the AF reported was slightly different, probably because of variation of the excitation sources.

Figure 7

Histopathological findings of normal bronchial bifurcation. Cross-sectional views of (a) HE staining, and (b) EL staining.

Figure 8

EL staining of the bronchus bifurcation. The sample was sliced along the contour. (a) Near the saddle point of the bifurcation and (b) far from the saddle point.

The depth profile of the elastin was characterized, and the EL staining along with the HE staining (shown in Figures 7 and 8) confirmed that elastin was observed in the submucosal area of the bronchial bifurcation, and was hardly observed in the epithelium region. This observation was consistent with previous reports that AF could be observed in the upper region of the submucosa, but not in the epithelium.²¹ The tumor sample which did not include elastin (evaluated by EL staining) showed a very weak AF; blue and green AF peaks were barely observed from spectrum analysis.

3.3.

AF Spectra from Biopsied Samples

Several human bronchus samples after resection were analyzed in conjunction with standard biopsy procedures in order to obtain a depth distribution profile of the fluorescent materials. The histopathological analysis indicated that the tissues used for this experiment were normal. As illustrated in Figure 9, the spectroscopic analysis was performed for samples prior to biopsy, immediately after the biopsy, as well as for the biopsied sample. An ex vivo biopsy was performed with standard biopsy tools used for routine diagnosis in hospitals. The thickness of the biopsied sample was about 0.4 mm. Figure 10 compares spectra of the three samples. A spectrum obtained from the sample prior to the biopsy was a typical spectrum of normal tissue. The biopsied sample showed a fairly weak but dominant broad peak at around 560 nm. There is an additional feature in the shorter wavelength region. A certain intensity of the AF was observed in the wavelength region between 450 and 500 nm; the peak position was not clearly identified. EL staining of the biopsied sample showed that the sample included the epithelium and a very thin layer of submucosa which included a very small amount of elastin. The spectrum shape obtained from the sample after biopsy was similar to that of the sample prior to biopsy. The AF spectrum shape and intensity were quite comparable. These phenomena would suggest that the AF peak observed at around 483 nm in the normal tissue probably mainly originated from tissue beneath the epithelium, and that the AF material whose spectrum peak was around 560 nm would be located in a much wider region in the depth direction.

Figure 9

Schematic of the ex vivo biopsy procedure and the AF analysis: (a) prior to the biopsy, (b) the biopsied sample, and (c) after the biopsy.

Figure 10

AF spectrum of the ex vivo biopsy samples: (a) prior to the biopsy, (b) the biopsied sample, and (c) after the biopsy.

A typical AF spectrum of dysplasia samples which were biopsied and classified through pathological analysis is shown in Figure 11. The spectrum shape was similar to that of the biopsied normal tissue, and a relatively weak signal was observed in the blue region (around 480 nm) as well as a clear peak in the green region (around 560 nm). Further study is required to understand the change in AF intensity associated with the development of precancerous tissues since those changes would take place in epithelium tissues, not in submucosa tissues where elastin is located.

Figure 11

AF spectrum of the biopsied sample. The sample was classified as dysplasic by pathological analysis.

AF spectra of elastin (soluble, of bovine origin), NADH, and FAD solved in saline were also measured and compared with those from bronchial tissues. Measurement of the liquid was performed using the same setup as that for measuring the bronchial tissues. The liquid solution was excited directly by the UV LD. As shown in Figure 12, the AF spectrum obtained from the elastin solution was similar to that of the normal tissues obtained from the human bronchus, and an intense peak was observed at around 483 nm with a weaker peak at around 560 nm. The spectrum obtained from NADH is similar to that of elastin, but the spectrum from the FAD was quite different. These facts imply that the AF peak located at around 483 nm through UV LD excitation should be related to elastin and/or NADH. Monitoring and precise analysis of the condition of elastin (which could easily be performed using the UV LD whose excitation wavelength is about 400 nm) would provide further information for detecting the early stage of lung cancer and of other diseases associated with the biochemical condition of elastin.²² Monitoring of AF associated with elastin would also provide useful information for laser welding techniques of tissues.¹⁵

Figure 12

AF spectra of various fluorescent materials.