2.1.

Experiment Set Up

As shown in Fig. 1, the experiment setup consists of a tungsten halogen lamp with 20-W power (HL2000; Ocean Optics Inc., Dunedin, Florida) and spectrometer (USB200, Ocean Optics) with a spectral range of 350 to 1100 nm and an optical fiber probe (R200-REF, Ocean Optics, Dunedin). This probe consists of a bundle of seven optical fibers—six illumination fibers around one read fiber—each of which is $200\mu \mathrm{m}$ in diameter ($\mathrm{NA}=0.22$, $\text{length}=2\mathrm{m}$) and a distance of source–detector separation of $200\mu \mathrm{m}$. A $3^{″}\times {0.25}^{″}$ stainless steel ferrule houses the fiber bundle. We also use a force gauge to measure the pressure.

Fig. 1

The scheme of experiment setup.

2.2.

Phantom Preparation

We prepared a two-layer phantom (Fig. 2). The underside layer consisted of Intralipid^® 20% (3.3 ml), distilled water (13 ml), agar (0.7 g), and beta carotene [$\ge 97.0\%$ (UV) Sigma-Aldrich] in different concentrations of 0.2, 0.3, 0.4, and $0.6\mu \mathrm{M}$. The upper layer is a mouse ear (2-month old, 20 g) with an average thickness of $250\mu \mathrm{m}$ (one-half of its full thickness is composed of two layers: the epidermis and the dermis, and below the skin layers, cartilage forms the structural support for the mouse ear).²⁵ Eight ear specimens were obtained from Shahid Beheshti Medical Science University and stored at 4°C until experiments were performed (15 min). All animals participating in this experiment were cared for according to the ethic committee of Shahid Beheshti University, Iran.

Fig. 2

A schematic of phantom. Underside layer consists of Intralipid, agar- and beta-carotene, and ear of mouse applied as an upper layer.

2.3.

Optical Clearing Methods

Here, we used two different OC methods to detect absorption spectra of beta-carotene from reflectance spectra. First, the method of experimental OC (compression-immersion OC) is explained, and then the method of analytical method is stated.

2.3.2.

Analytical method

As mentioned in Ref. 23, the analytical form of diffuse reflectance $R(\lambda )$ from semi-infinite phantom to determine optical properties can be stated as follows:

Eq. (1)

$$R(\lambda )=\frac{{\mu }_s^{\prime }(\lambda )}{k_1+k_2{\mu }_a(\lambda )},$$

where $k_1$ and $k_2$ are the parameters depended on the probe geometry and refractive indices of the studied sample and were determined to be $0.025{\mathrm{mm}}^{-1}$ and 0.075, respectively. ${\mu }_{\mathrm{s}}^{\prime }(\lambda )$ is the wavelength-dependent reduced scattering coefficient, and ${\mu }_{\mathrm{a}}(\lambda )$ is the wavelength-dependent absorption coefficient of tissue-like phantom. Assume $R_0$ indicates the diffuse reflectance for ${\mu }_{\mathrm{a}}(\lambda )=0$, so ${\mu }_{\mathrm{s}}^{\prime }(\lambda )=k_1{R}_0$. This diffuse reflectance $R_0$ can be estimated from the wavelength region of 600 to 900 nm using the scattering power law, as mentioned in Ref. 26 (see Fig. 3). Finally, the absorption coefficient is obtained as follows:

Eq. (2)

$${\mu }_{\mathrm{a}}(\lambda )=\frac{k_1}{k_2}.[\frac{R_0}{R(\lambda )}-1].$$

Fig. 3

Estimation of $R_0$ (dotted line) for measured results (solid line) using scattering power law. For the range $\lambda >600\mathrm{nm}$, the spectrum graph (solid line) can be fitted by the scattering power law $a{\lambda }^{-b}$ (dotted line), where $a=26.65$ and $b=0.63$ (with 95% confidence bounds).

In the visible spectrum, blood and melanin are the main chromophores that affect the absorption coefficient of our phantom, while variation of water and lipid does not have significant effects in this spectrum interval. Hence, the absorption coefficient of phantom can be expressed as follows:²⁷

Eq. (3)

$${\mu }_{\mathrm{a}}{(\lambda )}_{\text{phantom}}=M{\mu }_{\mathrm{a}}{(\lambda )}_{\mathrm{mel}}+B{\mu }_{\mathrm{a}}{(\lambda )}_{\text{Beta}}+Bl{\mu }_{\mathrm{a}}{(\lambda )}_{\text{blood}},$$

where

Eq. (4)

$${\mu }_{\mathrm{a}}{(\lambda )}_{\text{blood}}=\alpha Bl{\mu }_{\mathrm{a}}{(\lambda )}_{\mathrm{HBO}2}+(1-\alpha )Bl{\mu }_{\mathrm{a}}{(\lambda )}_{\mathrm{HB}},$$

where $M$, $B$, and $Bl$ denote melanin, beta-carotene, and blood volume fractions, respectively, and $\alpha$ is the degree of oxygenation of hemoglobin. The absorption spectrum of the mentioned chromophores is selected from the results presented by the “natural phenomena simulation group in University of Waterloo” and data presented by Jacques in 2013.^28,29

3.1.

Experimental Clearing

Figures 4(a) and 4(b) demonstrate the effects of experimental clearing methods on diffuse reflectance and absorbance [the absorbance $A(\lambda )$ is calculated by $A(\lambda )=-\log R(\lambda )]$. First, we apply compression OC and study the variation of diffuse reflectance (absorbance). The results including error bars show that the removal of blood dip between 500 and 570 nm depends on the amount of mechanical compression. This is because the transport of blood is a function of the amount of compression and Young’s module of tissue. One can see in Fig. 4(a) that, after the application of pressure and when blood is removed, the reflection increases due to less blood absorption. Indeed, scattering is also changed a bit in this range (scattering by RBC), but, in this wavelength range, blood absorption prevails.

Fig. 4

The influence of compression and immersion OC on the spectrum of reflectance and absorbance. The black line shows the (a) reflectance and (b) absorbance before applying experimental clearing. The red line depicts the variations of reflectance (absorption) due to compression OC ($300\mathrm{N}/{\mathrm{m}}^2$). The blue line presents the effects of the combination of immersion and compression OC on (a) reflectance and (b) absorbance. Eight ears are used while doing the experiment for investigating the repeatability of this method. Arrows point out absorption peak of beta-carotene.

Then, we simultaneously studied the influence of both compression and a chemical clearing agent on tissue phantoms. Blue lines including error bars in Figs. 4(a) and 4(b) reveal a noticeable change in absorbance due to utilizing DMSO, as the chemical OC agent, in OC of a mouse ear. Immersion of ears in DMSO leads to a decrease in optical scattering of them, so one can see an increase in optical depth (diffusion of DMSO in phantom causes a reduction in the scattering coefficient by refractive index matching).³⁰ Results depict that the combined OC (before 600 nm) can enhance the absorption spectra, whereas these two methods have similar effects in an optical window (a wavelength longer than 600 nm).

As shown in Fig. 4 and during application of OC, beta-carotene absorption spectrum appears in its specified wavelength. We define an indicator $A_{\text{Factor}}=A-B$, as shown in Fig. 5(a), to clarify the effect of OC on absorbance spectrum ($A$ is the absorbance at $\lambda =490\mathrm{nm}$ and $B$ is the absorbance at $\lambda =505\mathrm{nm}$). Figure 5(b) depicts the amount of $A_{\text{Factor}}$ for different values of beta-carotene concentration. The results show a linear relation between the concentration of beta-carotene and $A_{\text{Factor}}$.

Fig. 5

(a) Illustration of $A_{\text{Factor}}$ obtained from absorbance spectra of optically cleared phantoms, where $A$ is the absorbance at the wavelength 490 nm and $B$ is the absorbance at the wavelength 505 nm. The black solid line depicts the spectrum before applying compression OC, and the red dotted line illustrates the absorbance spectra of optically cleared phantom after applying compression. (b) Relation between $A_{\text{Factor}}$ and concentration of beta-carotene for different phantoms. The black line shows the influence of compression OC, and the red line presents the effects of the combination of compression and immersion OC on $A_{\text{Factor}}$.

3.2.

Analytical Clearing

Figure 6 compares the modeled absorption coefficient spectra and actual absorption coefficient obtained from diffuse reflectance spectra, as explained in Sec. 2.3.2. As shown in Fig. 6, these two curves are correlated with a high correlation coefficient of $R^2=0.94$, so the fitting parameters indicating the volume fraction of chromophores can be obtained from modeled spectra.

Fig. 6

Modeling measured absorption coefficient (black solid line) using diffuse reflectance presented by the red dotted line.

As shown in Fig. 7, analytical OC could extract and quantify the absorption coefficient of each chromophore within tissue-like optical phantom without experimental clearing methods for detection of beta-carotene.

Fig. 7

Extracted absorption coefficient based on the optimum volume fraction of main chromophores.

4.1.

Experimental Clearing (Compression-Immersion OC)

First, we used mechanical compression to weaken the effect of high blood absorption and obtain the amount of beta-carotene in our phantoms. Ermakov and Gellermann⁸ investigated the correlation between results obtained by compression mediated reflection spectroscopy and Raman spectroscopy. They illustrated the feasibility of this OC to extract the absorbance spectrum of beta-carotene. Because the blood distribution inside biological tissue depends on the amount of mechanical pressure and the stiffness of tissue, the accuracy of compression mediated reflection spectroscopy can be a key limitation for detecting chromophore at deep depths.

Hence, we used DMSO as a chemical clearing agent during compression-immersion clearing to obtain beta-carotene absorption because DMSO is a well-known skin penetration enhancer and its permeability rate $(176\pm 42\mathrm{g}/{\mathrm{m}}^2\mathrm{h})$ is $\sim 10$ times greater than that of water.^17,31 In addition, DMSO induces a reversible conformational change of protein when it substitutes for water.¹⁷ By contrast, for example, glycerol does not have a high enough penetration rate since it can hardly penetrate into dermis to achieve OC by topical application because of the barrier of stratum corneum.¹⁹ As one can see in Fig. 4(b), applying DMSO as the OC agent improves the detection of the absorption spectrum of beta-carotene.

In this experiment, we use a phantom consisting of a mouse ear and beta-carotene at a depth of $\sim 0.2\mathrm{mm}$ from the surface of phantom. (In fact, we apply a bilirubin-free phantom having a low amount of melanin that influences the results.) Using DMSO and compression helps to decrease the influence of blood and mismatched refractive index to better detect beta-carotene absorption spectra. The efficacy of our method to measure the beta-carotene content in human skin depends on the scattering coefficient of skin. The light scattering in skin depends on different elements, such as melanin and other tissue structures. The presence of melanin ($n=1.7$) on the background of low refractive index of cell cytoplasm and interstitial fluid ($n=1.34$) causes high scattering in the epidermis layer of the skin.³² However, very white skin having a low melanin content has a high scattering coefficient.^33–35 Hence, in the range of 400 to 700 nm, high absorption and scattering limits the ability of experimental and analytical OC to efficiently detect the desired chromophores.

4.2.

Analytical Clearing (Extracting the Absorption Coefficient of Beta-Carotene)

Johansson³⁶ measured the reflectance spectra of a thin cerebral phantom to estimate its absorption coefficient. Using linear regression, the absorption spectra were fitted to extract the main chromophore spectrum inside cerebral phantom. We applied this idea to perform analytical OC. That means, the blood and melanin spectra were digitally removed from the measured absorption spectra, then the absorption spectra of the desired chromophore should be apparent. The spectral elimination of dominant chromophores enhanced the capability of this method to detect the low value of concentration of chromophore at deep tissues without applying any experimental clearing. For example, Fig. 8 shows the spectral absorption coefficient for different values of concentration of beta-carotene inside the phantom. In general, a combination of all three considered modalities should be helpful for improving capabilities of the OC approach because of the reduction of tissue scattering and blood absorption impact on analytical reconstruction of chromophore concentration with the masked spectrum. To prove this claim, we also combine analytical clearing with experimental clearing methods. As shown in Fig. 9, the value of the extracted absorption coefficient of beta carotene is improved. Figure 10 also compares the value of the extracted absorption coefficient at a wavelength of 490 nm for these two methods. It is clearly specified that the estimated value of the absorption coefficient using this hybrid method (the combination of analytical and experimental clearing) is larger than experimental or analytical clearing alone. However, compression and immersion OC should be applied when analytical OC is not enough to get reliable data.

Fig. 8

Extracted spectral absorption coefficient of different values of concentration of beta-carotene using analytical OC.

Fig. 9

Extracted spectral absorption coefficient of different values of concentration of beta-carotene using combination of analytical and experimental OC.

Fig. 10

Relation between extracted absorption coefficient of beta-carotene at a wavelength of 490 nm with a concentration of beta-carotene inside phantoms for analytical clearing method and combination of analytical and experimental clearing method.