Application of Resonance Raman Spectroscopy to the Detection of Carotenoids In Vivo - Carotenoids: Physical, Chemical, and Biological Functions and Properties - page 102

Raman shift (cm -1 )

800 1000 1200 1400 1600 1800

200
160
120

1

C=C

6

C-C

4

C-CH 3

2

2

3

0

510 520 530 540

Wavelength (nm)

(b)

(a)

FIGURE 6.12 (a) Image of clinic- and i eld-usable, computer-interfaced, skin carotenoid RRS instrument,
showing solid state laser, spectrograph, and light delivery/collection module. (b) Typical skin carotenoid
Raman spectra measured in vivo . Spectrum (1) is obtained directly after exposure, and reveals a strong, spec-
trally broad, skin autol uorescence background with superimposed weak, but recognizable Raman peaks
characteristic for carotenoids. Spectrum (2) is obtained after i tting the l uorescence background with a fourth-
order polynomial, subtraction from (1), and scaling of the spectrum. Spectrum (2) is indistinguishable from a
spectrum of a β-carotene solution, shown as (3) for comparison.

absorbing in the visible spectral range. Since all individual C C stretch positions and bandwidths

are indistinguishable at the instrument's spectral resolution, our RRS approach allows us to use the
absolute peak height of the C

=

= C signal at 1524 cm −1 as a measure for the overall carotenoid con-

centration in human skin.
Experiments with varying light excitation intensities showed that the skin carotenoid RRS
response is stable up to the highest intensities permissible for skin applications (Ermakov et al.
2001a). To check the repeatability of the Raman measurements, we compared the RRS measure-
ments of skin with the measurements of a tissue phantom consisting of (a) a mixture of glycerol
and i ne aluminum oxide powder to simulate scattering, (b) -carotene, and (c) an organic dye

(coumarin 540) that simulates the skin autol uorescence background. While the repeatability for
the phantom was excellent, with a standard deviation below 1% for 10 consecutive measurements,
the repeatabilities in living human tissue were signii cantly lower, with standard deviations ranging
between 0.5% and 14% depending on the subject. To further investigate the origin of this effect,
we measured the spatial distribution of a skin tissue sample with a Raman imaging instrument.
The result, shown in Figure 6.13 clearly reveals that the skin carotenoid concentration varies sig-
nii cantly on a microscopic scale. Excitation spot sizes that are too small should be avoided due to
these variations. The relatively large, 2 mm diameter beam spot size used in our skin Raman mea-
surements appear to be an adequate solution to this effect, since it effectively integrates over these
microscopic spatial concentration changes.
To validate the skin carotenoid RRS detection approach, we initially carried out an indirect
validation experiment that compared HPLC derived carotenoid levels of fasting serum with RRS
derived carotenoid levels for inner palm tissue sites. Measuring a large group of 104 healthy male
and female human volunteers, we obtained a signii cant correlation ( p < 0.001) with a correlation
coefi cient of 0.78 (Smidt et al. 2004). Recently, we carried out a direct validation study, in which
we compared in vivo RRS carotenoid skin responses with HPLC-derived results, using the thick

β