In vitro hyperspectral analysis of tattoo dyes

There is no method that can guarantee effective, quick, and noninvasive removal of tattoo dyes. Laser methods are considered to be the method of choice. In this study, an attempt was made to determine the in vitro spectral characteristics of selected dyes used in permanent makeup and tattoos and to analyze the obtained parameters in terms of laser treatments optimization.


INTRODUCTION
A tattoo is an increasingly popular form of body decoration, which bases on implementing dyes to the skin. Originally, the tattoo had a ritual meaning, but now its role is mainly comes down to a decorative function. It is estimated that about 40% of Europeans have at least one tattoo. [1][2][3] The popularity of tattoos is followed by the need to develop effective methods of removing them. Currently, there is no method that can guarantee effective, quick, and noninvasive removal of tattoo dyes. Nevertheless, laser methods are considered to be the method of choice. The mechanism of interaction of laser radiation with the tissue, which is the basis for the use of lasers in medicine, is based on the theory of selective photothermolysis. 4,5 Identification of spectral parameters of chromophores (responsible for the absorption of laser radiation) and their distribution is a key factor determining the optimal course of the laser treatment. The currently used methodology for performing laser treatments provides for a priori determination of which chromophore will absorb laser radiation, whereas the type, number, and distribution of chromophores, and even more so their spectral characteristics, are not subject to any objective measurement. As a result, a number of laser procedures do not show the expected effectiveness and/or lead to the occurrence of side effects. [6][7] Therefore, in this study, an attempt was made to determine the in vitro spectral characteristics of selected dyes used in permanent makeup and tattoos and to analyze the obtained parameters in terms of laser treatments optimization.

Histopathological aspects of dyes implantation into the skin
After implantation, tattoo dyes are deposited primarily in fibroblasts and macrophages of the dermis. Dye-containing cells are often accompanied by fibrosis. Small amounts of dye are also found in the connective tissue in the form of extracellular aggregates. 8 The size of the dye particles after implantation into the skin varies between 0.5 and 4 μm. Turquoise and red dyes can create aggregates even twice as large. 9 The location of the dye varies significantly along with the time passing since implantation. The most important ingredient related to the function of the dyes is the color-imparting pigment. Both organic and inorganic components are used as pigments. Taking into account the chemical composition, the main elements that build dyes used in tattoos are as follows: aluminum (87%), oxygen (73%), titanium (67%), and carbon (67%). The composition is important for the tattoo removal procedures. 10

Tattoo removal
Along with the growing tendency of desire to have a tattoo, there is a need to develop effective methods of removing it. It is estimated that about 5% of tattooed people regret their decision and wish to have their tattoos removed. 11 Among the methods of tattoo removal, some can be mentioned. 12

Laser tattoo removal
The mechanism of laser tattoo removal is based on the phenomenon of selective photothermolysis. The theory of selective photothermolysis was developed by Anderson and Parish, and it is the base of all currently used laser treatments in aesthetic medicine and dermatology. [15][16][17][18] According to the selective photothermolysis theory, it is assumed that the tattoo dye is an egosogenic chromophore. Therefore, the laser radiation length used should be selectively absorbed by the tattoo dye. The selectivity of laser radiation absorption is a necessary factor for the laser tattoo removal treatment to be both safe and effective. It should be taken into account that, apart from the tattoo dye, laser radiation can also be absorbed by other naturally occurring chromophores in the skin, mainly water, hemoglobin, and melanin. 17,18 Absorption of laser radiation by any of the endogenous chromophores (water, melanin, and hemoglobin) warms the skin and consequently causes its thermal damage. Therefore, the length of the laser radiation should be selected in a way, that is, efficiently absorbed by the dye and, as little as possible, by the skin. 15,16 When optimizing the parameters of the laser radiation in relation to the dye, other parameters should also be taken into account. A very important factor is the depth of laser radiation penetration into the skin. It can be written, to some extent, that the depth of penetration of the laser radiation into the skin is directly proportional to the wavelength of the laser radiation: the longer the wavelength, the deeper the radiation penetrates the skin. Due to the fact that the tattoo dyes are located relatively deep, at the dermo-epidermal junction, lasers with a wavelength shorter than 500 nm are not useful to remove them. Another key laser performance parameter that should be considered in laser tattoo removal procedures is the pulse duration. 15 are becoming more and more popular. In this case, the pulse duration is 10 −12 s. 17,18 The mechanism of interaction of laser radiation with dye particles has not been well understood so far. Under the influence of the laser, the optical properties of the dye change. As a result, the pigment discoloration occurs-the spectral range of light absorption in the range of visible radiation changes. The degradation of the dye is done probably by thermal, photochemical, and photoacoustic action, and its fragmentation helps in its removal by the skin's immune system. 15,16 The effectiveness of laser tattoo removal mainly depends on the following factors: the color of the dye, the chemical properties of the dye, the phototype of the patient's skin, the laser used, the ability to optimize the treatment parameters in relation to the patient, and the ability to perform the procedure correctly. [19][20][21] At the same time, different dye colors have different absorption maxima, which results in the necessity to use lasers with several wavelengths. Table 1 shows the optimal length of laser radiation (optimal laser) in relation to different dye colors. 22 Effective tattoo removal usually requires from 8 to 16 treatments.
In the case of multicolored tattoos, it is necessary to perform more treatments-even 20. Considering that the treatments are performed with a 4-6-week interval, laser removal of the tattoo may last several months. At the same time, when removing color tattoos, several different types of lasers should be used: for example, the Nd:YAG KTP laser for the removal of red dyes, the alexandrite laser for green dyes, and the Nd:YAG laser for black and navy blue dyes. 21,24 Moreover, dyes are often a mixture of various organic and inorganic pigments, which also makes it difficult to determine their spectral parameters (absorption maxima). 25

Aim of the study
The aim of this study was an attempt to determine the spectral properties of dyes to maximize the effect of laser radiation absorption by pigments.

MATERIALS AND METHODS
Seven dyes used in permanent makeup and tattoos were analyzed in vitro ( Figure 1). The names of the dyes together with the manufacturer are presented in In the first stage of the research, two drops of the dye were placed on the glass slide. Then, using a dry brush, the dye was gently spread over the entire surface of the slide to cover its entire surface. So, prepared slide was left to dry.
In the next stage, the hyperspectral analysis and the analysis in visible light using a classic camera were carried out.
Hyperspectral analysis was performed to determine the spectral characteristics of the dye on the entire surface of the slide. The maximum reflectance and the wavelength for a given dye were determined for the maximum reflectance in the studied wavelength range: 400-1000 nm. The choice of such a wavelength range was dictated by the fact that this range coincides with the wavelengths of lasers used in laser tattoo removal treatments.
In the next stage, the optical properties of the dyes were determined based on visible light imaging using camera.

Hyperspectral parameters analysis
The hyperspectral analysis of the dyes was performed with a hyperspectral camera using image analysis and processing methods.
Computer-assisted optical imaging is increasingly used in medicine.
It allows us not only to compare the features of objects with the pattern but also allows for the biometric assessment of tissue parameters, which is not possible with a standard physical examination.

Analysis of visible light
The analysis of visible light was aimed to identify parameters such as homogeneity and contrast for the entire surface of the tested dyes distributed on the surface of the slide. Due to the fact that after drying the dyes did not have a homogeneous color, it was decided to quantify the color changes for the entire surface of the tested dyes.
For this purpose, the dyes after drying on the slides were photographed using a Canon 5D camera with a Tamron 28-75 mm, f/2.8 USM lens. The tested dyes were illuminated with an incandescent light source with a high color rendering index. The photos were acquired in the * .RAW format.
Then, on each photo of the dyes, a region of interest (ROI) was identified that corresponded to the surface of the dried dye. The prepared images were imported to the MATLAB environment.
The following image parameters were determined: 1. homogeneity,

GLCM contrast (Gray Level Co-occurrence Matrix).
In order to determine and compare homogeneity and contrast among dyes, all images were normalized. Due to the fact that the acquired images were colored, the first step was to transform the color images into gray images. To do this, first the image pixel of a specific color was read with the appropriate share of each of the basic components, and the R, G, and B components were separated. Then the brightness of each of the primary colors was summed, and this sum was divided by 3, without the rest. Then the images were normalized by expanding the range of colors to the full range of grays from white to black (0-255). Contrast and homogeneity were identified for prepared images (normalized in gray scale).
The GLCM method was used to determine the difference among adjacent temperature fields. The application of the GLCM method is designed to determine the homogeneity of the distribution of temperature fields on the thighs of patients before and after the therapy. As a measure of homogeneity, the contrast calculated based on the GLCM matrix was adopted.
The idea of the GLCM method is to calculate in the ROI indicated by the operator or for the whole image, the number of neighborhoods of individual pixels. These neighborhoods can be analyzed from different directions: horizontal, vertical, and diagonal. The closest possible neighborhoods are analyzed most often-for example, horizontally these are two pixels adjacent to each other. The number of neighborhoods between each pixel brightness is recorded in the GLCM by specifying the number of neighborhoods between pixels "1" and "0." The contrast is then calculated based on the GLCM matrix.
In the example described above, a two-dimensional model (0,1) was used, whereas in the analyzed case, an n-dimensional model was used, where n is the brightness difference between neighborhood image pixels. The GLCM analysis in this case included pixels adjacent to each other on the right side in the lines of the rows.
The homogeneity test was aimed at determining the brightness uniformity across the entire ROI. In the adopted research model, homogeneity is understood as where i is the brightness of the tested pixel; j is the brightness of the adjacent pixel.

F I G U R E 5
Reflectance spectra of all registered dyes respectively: red (A), yellow (B), dark pink (C), black (D), light pink (E), brown (F), and white (G)

Hyperspectral analysis
The proposed method of hyperspectral analysis allowed us to obtain a series of images recorded at successive wavelengths. A total of 128 images were recorded.   pigment it is almost 38 times higher and amounts to 3222 relative units.
The relatively large difference in reflectance between similar colors is also noteworthy. And so, for the dye no. 1 (red), reflectance is 751 relative units, and for dye no. 3 (dark pink), it is almost twice as high: 1304 relative units.
It should be noted that reflectance, in addition to the optimal wavelength of laser radiation, is a factor determining the effectiveness of laser tattoo removal treatments. According to the assumption that reflectance is inversely proportional to absorbance (with little influence of transmittance), it can be assumed that high reflectance will reduce the effectiveness of laser tattoo removal treatments. Thus, based on the results presented in Figure 7, it can be concluded that black dye will be the easiest to remove, and then in descending order: brown, red, dark pink, yellow, light pink, and white. In the case of the white dye, almost all of the incident energy is reflected and/or dissipated. The dye will therefore not absorb the laser radiation and will be extremely difficult to remove.

Analysis in visible light
In addition to the hyperspectral analysis, a visible light analysis was also performed. This analysis was aimed to determine the homogeneity of the distribution of chromophores inside the dyes. For this purpose, the GLCM contrast of the tested dyes and their homogeneity were determined.
Images of the analyzed dyes in visible light after performing normalization procedure are shown in Figure 8. The purpose of the normalization was to make it possible to compare the brightness of individual pixels between the tested dyes. The normalization was carried out for each of the images recorded in the RGB format (Figure 1), and the brightness was extended to the full range from 0 to 255. Thus, in the image of each dye, the brightest pixel was identified and was given the brightness of 255 (white) and the darkest pixel the brightness of 0 (black). The brightness of the remaining pixels was then extended proportionally over the full range of gray levels.

DISCUSSION
The method of choice for tattoo removal is laser techniques. During the procedure, laser radiation should be selectively absorbed by the dye. Under the influence of the absorbed laser radiation, the dye is fragmented, which facilitates its removal through macrophages and conditions changes in the optical parameters of the dye, which results in a reduction of radiation absorption in the range of visible radiation. In this study, an attempt was made to determine the spectral parameters of dyes, which are particularly resistant to laser removal treatments. The wavelength of which the maximum reflectance of the laser radiation occurs was determined using hyperspectral imaging.
The tested radiation ranged from 400 to 1000 nm, which covers almost 100% of currently used non-ablative lasers areas.
The obtained results show that for the whole group of dyes tested, lasers with emission ranging from 634 to 732 nm should not be used.
A hyperspectral profile was determined for each dye, which allowed us to optimize the laser radiation length. It is worth emphasizing that potentially similar dyes, such as light pink and dark pink, have significantly different spectral profiles. Thus, it is not possible to determine the optimal wavelength of the laser radiation used for its removal solely based on the color of the dye. These significant differences in spectral profiles probably results from the mixture of numerous compounds in a dye, causing a specific shade of color. Even a small (quantitatively) admixture of another pigment can significantly affect the spectral parameters.
Moreover, the maximum reflectance was determined for all analyzed dyes. It should be noted here that reflectance, next to the optimal wavelength, will be a key parameter responsible for the effectiveness of laser tattoo removal treatments. It should be expected that for high reflectance pigments-such as white and yellow, their removal from the skin will be extremely difficult and will require at least a dozen or so treatments. On the contrary, for dyes with low reflectance (high absorbance), laser tattoo removal will be much easier.
It is not possible to estimate the reflectance from the color of the dye alone same as with the optimal wavelength. And so for dyes 1 and 3 (red and dark pink, respectively), the brightness is similar, whereas the reflectance is almost twice as high in the case of the dark pink dye.
In addition to the hyperspectral analysis, the GLCM contrast and homogeneity for images recorded in visible light were also determined.
The aim of the analysis was to determine the homogeneity of the distribution of chromophore molecules within the dyes. The white dye is the most homogeneous and the yellow dye the least homogeneous. In the case of white dye, its chemical composition is probably responsible for its high homogeneity. White dyes are made of micronized titanium dioxide. The use of only one pigment allows the creation of a highly homogeneous suspension.
In yellow dyes, mixtures of many different pigments, including organic and inorganic, are used, which may result in lower homogeneity.
Summing up, it should be stated that the proposed method can Moreover, it should be emphasized that the conducted analyzes were performed in the form of in vitro tests. Thus, the spectral parameters of the skin, where the dyes are implanted, were not considered.
The optimal solution, but practically very difficult, would be to verify the hyperspectral parameters of the tattoo in vivo, which creates opportunities for further research in this area.

CONCLUSIONS
The conducted research allows us to present the following conclusions: 1. The maximum radiation reflectance ranges from 634 to 732 nm for the tested dyes.
2. Visually very similar colors (e.g., red and dark pink) may differ significantly in the wavelength for which the maximum absorption of the radiation occurs.
3. White and yellow dyes are characterized by the highest reflectance value. This determines difficulties in elimination them from the skin with laser methods.
4. The black dye is characterized by the lowest reflectance coefficient-almost 40 times lower than the white dye.
5. Low reflectance of black dye results in more safe and effective removal treatments.
6. The homogeneity of radiation absorption can be identified using methods of analysis and processing of images in visible light.
7. Optimization of the wavelength of which the maximum absorption/reflectance of radiation occurs may allow us to increase the effectiveness of laser treatments for removing permanent makeup and tattoos.