Transcript et al.
SPECTROPHOTOMETRIC DIFFERENTIATION OF HUMAN SKIN MELANOMA. I. DIFFUSE REFLECTANCE OF LIGHT V.G. PETRUCK1, A.P. IVANOV2, S.M. KVATERNYUK1, V.V. BARUN2 1Vinnica National Technical University, Vinnica, Ukraine 2B.I. Stepanov Institute of Physics, Belarus National Academy of Sciences, Minsk, Belarus [email protected] [email protected] • Our goal is to investigate and to propose an objective, noninvasive optical tool for discriminating melanoma skin of a person from intact (healthy) or benign nevus (nondangerous) skin. As a first step, this paper studies spectral diffuse reflectance of skin tissues. • ABSTRACT. Measurement results of spectral reflection characteristics under multiple light scattering by healthy skin and by skin regions with melanoma or nevus are given. Experimental setup that operates on the base of the Taylor method is described. It is shown that the diffuse reflectance R of melanoma skin is lower at all the studied wavelengths from the range about 450 to 1000 nm as compared with that of healthy and benign nevus skin. This conclusion is confirmed by a large number of measurements of big groups of different persons. The gathered data demonstrated an opportunity to differ malignant and healthy or benign formations, while operating in the visible to near IR range. Examples of such a differentiation at several wavelengths are given. Content • • • • • • 1. Introduction 2. Experimental setup 3. The persons 4. Reflection characteristics 5. Discrimination of skin pigment formations 6. Conclusions 1. Introduction The traditional means for non-invasive investigation of human skin melanoma is visual dermatoscopy. One also applies digital devices that enable a magnified tumor image to be get on a computer monitor. However, even by using modern devices equipped by modern optics and digital CCD cameras, the diagnostic methods of oncologic tumors based on dermatoscopic index according to the ABCD rules are generally subjective. Besides, the accuracy of the differentiation of melanomas and nevuses is low and depends essentially on the experience of oncologic physicians. There are known various publications, where melanomas were studied by optical means in the visible, near-IR, and IR [1 – 23]. This work is devoted to complex investigation of spectral light reflection by healthy skin and by pigment formations on skin, starting from measurements themselves and their probabilistic treatment up to assessing operational and diagnostic characteristics of a method for differentiating melanoma. 2. Experimental setup (Vinnica) The main components are monochromator 1 (wavelength = 450 – 1000 nm) and two integrating spheres 2 and 3. One of them (2) is used as a standard and another (3) as a measuring sensor. The latter is in close contact with biological object. Experimental setup (Minsk) а 2 б 1 3 2 4 6 5 2 4 1 3 4 It uses a white light source (1), a single integrating sphere (3), and a spectrometer (5). On the right panel, there are shown various accessories to measure optical characteristics of biological objects and humoral media in vivo and in vitro. 3. The persons We investigated spectra healthy tissues of 552 persons, who do not have any pathology of skins, bones, or internal organs. There were among them 237 men (43 %) and 315 women (57 %) of ages 15 to 45 years. The spectra were measure too for 31 patients with melanoma. There were among them 9 men (29 %) and 22 women (71 %) of ages 21 to 89 years. The melanoma diagnosis in different histological forms was confirmed for all the sick patients according to the classification adopted by the World Health Organization. The spectrophotometric studies of nevus were made for 20 patients with suspicion of maligned nevus. Among the investigated persons of this group, there were 8 men (40 %) and 12 women (60 %) of preadult and mature ages. All the nevus carriers were operated, and the diagnosis was confirmed by the histological studies. 3. REFLECTION CHARACTERISTICS Fig. 3. Diffuse reflectance of healthy skin as a function of wavelength, nm. Here are shown the median (point in a box), interquartile range containing 50 % of sample observations between 25th and 75th procentiles (box), maximal and minimal values (ends of vertical bars) Melanoma skin Fig. 4. The same as in Fig. 3, but for melanoma skin Benign nevus skin Fig. 5. The same as in Fig. 3, but for benign nevus skin 0.08 ã 0.5 0.06 1 3 0.04 0.4 2 0.02 0.3 600 800 0 1000 , í ì Fig. 6. Mean values of spectral reflectance (solid lines, left ordinates) and its variances (dashed, right ordinates) for healthy skin (1), benign nevus (2), and melanoma (3) 4. DISCRIMINATION OF SKIN PIGMENT FORMATIONS AT VARIOUS WAVELENGTHS (N – nevus, Int – healthy skin, M – melanoma) 570 nm 820 nm 600 nm 700 nm 830 nm 920 nm Probability density (AU) of R values for intact skin (Int), skin near pigment formation (CSM), skin with melanoma (M) and benign nevus (N) at = 880 nm 5. Assessing some tissue parameters We used a procedure [24]. It fits experimental reflectance values at several wavelengths to calculated ones. The procedure as applied to mean reflectances (not to individual ones) enabled us to observe a tendency of changes in blood volume fraction Cv and in the product of melanin volume fraction by epidermis thickness fd for healthy skin, skin with benign nevus and melanoma. The solution to the inverse problem gave Cv = 0.01 – 0.02 and fd = 4 – 6 mm (healthy); 0.01 – 0.02 and 6 – 7 mm (nevus); 0.024 – 0.045 and 7 – 9 mm (melanoma). We could not retrieve the blood oxygenation degree S owing to several reasons. They are (1) low sensitivity of the reflectance to S, (2) the use of mean reflectances, and (3) large dispersion of the measurements at wavelength 600 nm. It follows for the data obtained that melanoma is featured in noticeable increase of blood concentration Cv and of absorptive epidermis thickness. The parameters of healthy skin and skin with nevus are close to each other. CONCLUSION Measurements of diffuse reflectance have showed that there are several wavelengths in the visible and near IR, at which it is possible to discriminate melanoma from healthy and benign nevus skin. At these , R values for melanoma skin are lower that for other studied pigment formations. This conclusion is confirmed by the experiments to be statistically valid. Unfortunately, there are a lot of tissue structural, biophysical and optical parameters that really vary for skin pigment formations, such as epidermis thickness, melanin concentration in epidermis, blood volume faction, blood oxygenation degree, and others. This prevents from strict theoretical simulations of skin diffuse reflectance. However, it is our future objective to use known theoretical methods for computing R spectra and comparing them with experimental results. This will enable one to propose practical methods for noninvasive optical diagnostics of melanoma cancer at its early stages. References • • • • • • • • • • • • • • • • • • • • • • • • ([1 – 5] – diffuse reflectance, [3, 6 – 10] – spatially resolved spectroscopy, [10 – 14] – multispectral images, [18 – 23] – dynamic thermal images) [1] R.Marchesini, et al. Photochem. Photobiol., 53 (1991) 77–84 [2] R.Marchesini, et al. Photochem. Photobiol., 55 (1992) 515 – 522 [3] G.Zonios, et al. J. Biomed. Opt., 13 (2008) Paper 014017 [4] G.Zonios, A.Dimou. Opt. Express, 14 (2006) 8661–8674 [5] V.G.Petruk, et al. Proc. SPIE, 8698 (2013) Paper 86980F [6] V.P.Wallace, et al. Phys. Med. Biol., 45 (2000) 735–751 [7] B.W.Murphy, et al. J. Biomed. Opt., 10 (2005) Paper 064020 [8] V.P. Zakharov, et al. Phys. of Wave Proc. and Radiotech. Syst., 11 (2008) 89–97 [9] I.Diebele, et al. Biomed. Opt. Express, 3 (2012) 467 – 472 [10] V.P. Zakharov, et al. Izv. Samar. Sci. Center RAN, 15, No.6 (2013) 126–130 [11] C.-C.Yu, et al. Opt. Express, 16 (2008) 16227–16239 [12] R.Marchesini, A.Bono, M.Carrara. J. Biomed. Opt., 14 (2009) Paper 014027 [13] E.V.Filonenko, et al. Sibir. Oncolog. J., No. 3 (51) (2012) 50–53 [14] G.Lu, L.Halig, et al. Proc. SPIE, 9034 (2014) Paper 903413 [15] K.G.Phillips, et al S.L.Jacques. J. Biomed. Opt., 15 (2010) Paper 041514 [16] S.L.Jacques, J.C.Ramella-Roman, K.Lee. J. Biomed. Opt., 7 (2002) 329–340 [17] N.Kollias, G.N. Stamatas. Opt. Diagnostics in Dermatol., 7 (2002) 64–75 [18] J.Cary, L.Kalisher, N.Sadowsky, B.Mikic. Radiology, 115 (1975) 73–77 [19] M.P.Cetingul, C.Herman. Int. J. Therm. Sci., 50 (2011) 421–431 [20] C.Herman, M.P.Cetingul. J. Vis. Exp., 51 (2011) 2679–2681 [21] T.-Y.Cheng, C.Herman. Int. J. Therm. Sci., 86 (2014) 175–188 [22] M.P.Cetingul, C.Herman. Phys. Med. Biol., 55 (2010) 5933–5951 [23] C.Song, et al. IEEE Trans. Biomed. Eng., 31 (1984) 9–16 [24] A.P. Ivanov, V.V. Barun. Opt. Spectrosc., 104 (2008) 300 - 307 THANK YOU FOR THE ATTENTION!