Chromophore Mapping: a New Technique to Characterize Aging Human Skin, In Vivo Paul J. Matts, Ph.D 1; Jeremy Carey, MA MEng2; Symon D. Cotton, Ph.D 2; 1The Procter & Gamble Company, Rusham Park Technical Centre, Egham, Surrey, UK; 2Astron Clinica Ltd, Cambridge, UK Visible light images and corresponding melanin, haemoglobin Introduction Methods Discussion Visible light images and corresponding melanin, haemoglobin and collagen SIAgraphs taken from random subjects in the and collagen SIAgraphs taken from random subjects in the “young” 20-25 year age group “aged” 60-65 year age group Visible light images Whilst the SIA technique has proven highly valuable in the (actual imaged field Visual inspection of the chromophore SIAgraphs of young and aged It is simple fact that we judge one other, consciously or unconsciously, diameter 12mm) subjects reveal a dramatic contrast, particularly for the melanin and based in part on the aesthetic appearance of our outer integument, measurement of diseased skin, namely the early diagnosis of haemoglobin chromophores. In short, in young skin, these chromophores “skin”. As the human retina responds to a narrow bandwidth of the malignant melanoma, it was not known whether it could be used are distributed in an extremely homogeneous manner, consistent with the electromagnetic spectrum (so-called “visible light”, a nominal 400- to successfully map chromophores in normal skin in order to explain delicate, even palette of youth. With age, a significant increase in both changes in human cutaneous colouration with age. This present Melanin SIAgraphs total concentration and inhomogeneity is evident for both chromophores. 700nm), the interaction of these wavelengths with skin, therefore, is of (corresponding to visible poster, therefore, describes a study to evaluate the utility of the SIA This is wholly consistent with the known changes in these chromophores as primary importance in our understanding of the way we perceive others images above – note technique for this purpose: darker pixel = greater a function of actinic ageing – namely the accumulation of damaged and are ourselves perceived. A modern understanding of cutaneous melanin concentration) melanogenic units (producing hyper-melanotic lesions such as solar optics is based largely on the development of increasingly-sophisticated Subjects: 400 Caucasian female subjects (Fitzpatrick skin types I-III) lentigos, diffuse hyperpigmentation and hypo-melanotic lesions such as mathematical models to explain the interaction of light with skin. This aged 10-70 with normal, un-diseased skin were recruited in Haemoglobin SIAgraphs idiopathic guttate hypomelanosis) and damaged vasculature (including understanding of cutaneous light transport highlights the surprising fact Reading, UK. Subjects were recruited so that they fell equally (corresponding to visible telangiectasia and low-grade purpura)7,8. These visual observations were across the 10-70 age range into 12 cohorts of 5 years each, images above – note darker confirmed emphatically by objective analysis of the SIAgraphs by custom that normal human skin appearance is driven primarily by surface containing approximately 33 subjects per cohort (i.e., Group 1, 10- pixel = greater haemoglobin algorithms. Results confirmed an apparent “lag” phase for total skin topography and only three internal chromophore components, that is, concentration) 15 years; Group 2, 15-20 years, etc.). concentrations of both chromophores (lasting until approx age 30 for the concentration and spatial distribution of melanin, haemoglobin and melanin and 40 for haemoglobin), after which there was a steady increase collagen. Whereas there are a plethora of methods to characterise and Measurement Site: the skin between the thumb and the first Collagen SIAgraphs in each chromophore total concentration. There was also a concurrent quantify skin surface topography, chromophore mapping remains a finger of each subject’s right hand was used as the measurement (corresponding to visible progressive increase in inhomogeneity of each pigment. More detailed remarkably un-researched area. site for this study. The dorsal hand was chosen as it has been shown images above – note lighter analysis of the melanin SIAgraphs revealed a progressive accumulation of to receive approximately one half of the dose of erythemally- pixel = greater dermal melanin “spots” across a lifetime, with a linear increase in total spot area collagen density) effective solar UV radiation relative to vertex7 (the face receives and an apparent merging of spots at approximately age 50-55 (continued A new technique “Spectrophotometric Intracutaneous Analysis” only one third). The dorsal hand, therefore, represented a site area increase with parallel drop in spot count). (“SIA”) has been developed, commercialised and shown to have where we had confidence that both chronological effects and excellent sensitivity and specificity in the early identification of malignant Visual inspection of the collagen SIAgraphs reveal an apparent overall loss actinic damage to the chromophores of interest in this study should melanoma in human skin1,2. The technique is based upon a unique with age both of density and the fine “egg-box” lattice attributed to be well-expressed. Close-up views of SIAgraphs from young and dermal papillae. Once again, these subjective observations are supported combination of dermatoscopy and contact remittance Measurements: Subjects were equilibrated for 20 minutes in a old subjects by analysis of the SIAgraphs. A progressive decrease in collagen density spectrophotometry. The SIA hardware comprises a hand-held imaging was noted across the entire age span, accompanied by a progressive controlled atmosphere (20+1oC; 50+10%RH). The measurement site probe, attached by umbilical to a laptop computer, that is placed in decrease in homogeneity. These trends fit extremely well with the known on the subject’s right dorsal hand was then lightly sprayed with a 20% contact with the skin surface. High-intensity LEDs illuminate the skin progressive atrophy of dermal papillae and total dermal collagen bulk, (v/v) aqueous ethanol solution (to act as an optical “matching surface with four consecutive discreet wavebands between 400nm and consistent with intrinsic and extrinsic ageing 9. fluid”). SIA measurements were then taken; each lasted approx 6 1000nm, spanning the visible spectrum and a small range of near seconds. Melanin in young subject: evenly distributed; virtually no Haemoglobin in young subject: delicate vascular network; discreet capillary Collagen in young subject: even arrangement of collagen; “egg-box” lattice Melanin in elderly subject: very uneven distribution of melanin; formation of Haemoglobin in elderly subject: disruption of vasculature evident – Collagen in elderly subject: noticeably more inhomogeneity; apparent reduction The observations of this current study were predicted and modeled by one infrared radiation. A digital image is captured for each waveband. inhomogeneity apparent bundles present (small dots as viewed from above) of dermal papillae just visible melanin “plaques” as a result of melanocyte photodamage agglomeration, dilation, tortuosity and formation of talengiectasia in homogeneity of egg-box papillae of the present authors 10. A model was proposed that described human skin Custom SIA algorithms then solve the complex relationship that exists Image analysis: The SIA parametric maps (“SIAgraphs”) texture and color changes with age in the unorthodox terms of “ amplitude” between R-G-B-NiR “colour-space” and melanin-haemoglobin-collagen constituted 8-bit 1024x1024 256 level greyscale maps of and “frequency”. The principle components of human skin color, namely Mean melanin spot total area (pixels x1000) 146 42 Age Groups 34 150 chromophore concentration in PNG image file format and, as such, melanin and haemoglobin, both display overall increases in total “histology-space”, using a sophisticated model of cutaneous light- 1 = 10-15 37 2 = 15-20 31 were readily amenable to sophisticated image analysis techniques concentration (increased “amplitude”) and heterogeneity (decreased Mean haemoglobin index 136 Mean collagen index Mean melanin index 3 = 20-25 140 transport3,4. Since there is a proven one-to-one mapping between 32 4 = 25-30 for calculation of a variety of relevant endpoints. The following 126 5 = 30-35 28 “frequency”) as a function of chronological / actinic ageing. Effective colour-space and chromophore parameters5,6, individual chromophore 6 = 35-40 analyses were undertaken: (a) mean global grey-scale 27 7 = 40-45 8 = 45-50 130 treatments to improve appearance, therefore, need to reverse this effect. 116 25 parameter values can be retrieved from the model, given the colour (corresponding to mean chromophore concentration in the 22 9 = 50-55 Overall, therefore, these data appear to confirm chromophore mapping 10 = 55-60 11 = 60-65 120 vector obtained from each point in a colour skin image. The magnitude imaged skin field) (b) internal image standard deviation 106 17 12 = 65-70 22 using the SIA technique as a remarkable new tool to characterize and of each chromophore parameter is displayed at each pixel location in a (corresponding to an index of chromophore homogeneity) and (c) 96 1 2 3 4 5 6 7 8 9 10 11 12 12 1 2 3 4 5 6 7 8 9 10 11 12 19 1 2 3 4 5 6 7 8 9 10 11 12 110 1 2 3 4 5 6 7 8 9 10 11 12 explain the visual appearance of ageing skin. separate image, giving three parametric maps: epidermal melanin, for melanin SIAgraphs, contiguous cluster analysis (to identify Figure 1: Mean melanin index as a function of age (equal 5-year Figure 3: Mean melanin spot total area (within field of view) as a Figure 5: Mean haemoglobin index as a function of age (equal 5- Figure 7: Mean collagen index as a function of age (equal 5-year cohorts, age 10-70) function of age (equal 5-year cohorts, age 10-70) year cohorts, age 10-70) cohorts, age 10-70) dermal haemoglobin and collagen (a dermal melanin map is also pigmented spots) and subsequent feature (spot) count and mean 16.4 12.9 1 = 10-15 spot area calculations. 1. Moncrieff M, Cotton S, Hall PN & Claridge E, Extracting the diagnostic features if 6. Cotton S, Claridge E, Hall PN, A skin imaging method based on a colour 7.1 17.0 provided as a diagnostic criterion for melanoma). In short, the SIA 2 = 15-20 Mean internal melanin image SD Mean internal haemoglobin image SD skin surface microscopy for pigmented skin lesions using spectrophotometric formation model and its application to the diagnosis of pigmented skin lesions. Mean internal collagen image SD 3 = 20-25 intracutaneous analysis. Medical Image Understanding and Analysis, 49-52, 2000 Proceedings of Medical Image Understanding and Analysis 99, 49-52, 1999 11.9 Mean melanin spot count 4 = 25-30 6.6 16.5 technique is able to obtain a high-resolution white-light image of the skin Image analysis interval data were treated using multi-factor 14.4 10.9 5 = 30-35 6 = 35-40 2. Moncrieff M, Cotton S, Claridge E & Hall P, Spectrophotometric intracutaneous analysis - a new technique for imaging pigmented skin lesions. Brit. J. Derm. 146(3), 7. Griffiths, C.E.M., The clinical identification and quantification of photodamage, Brit. J. Dermatol ., 127 (Suppl . 41), 37-42, 1992 6.1 16.0 over a 12x12mm area and four additional maps that display the ANOVA techniques with age as the main factor (grouped per 5- 7 = 40-45 448-457, 2002 8 = 45-50 8. Gilchrest, B.A., Blog, F.B. and Szabo, G. Effects of ageing and chronic sun 12.4 9.9 9 = 50-55 5.6 15.5 3. Claridge E, Cotton S, Hall P & Moncrieff M, From colour to tissue histology: Physics exposure on melanocytes in human skin. J. Invest. Dermatol . 73, 141-143, 1979 concentration of epidermal melanin and haemoglobin, collagen and year cohort). Results constituted the SIAgraphs themselves and 10 = 55-60 11 = 60-65 based interpretation of images of pigmented skin lesions. Medical Image Computing and Computer-Assisted Intervention - MICCAI'2002. LNCS 2488, vol I, 730-738, 2002 9. Sauermann, K, Clemann , S, Jaspers, S, Gambichler , T, Altmeyer , P, Hoffmann, K, 8.9 15.0 Ennen, J, Age related changes in human skin investigated with histometric melanin in the papillary dermis, pixel by pixel. plots of image analysis LSD means and errors vs respective age 10.4 12 = 65-70 5.1 4. Claridge E, Preece , S, An inverse method for the recovery of tissue parameters from colour images. Information Processing in Medical Imaging, LNCS 2732, 306-317, 2003 measurements by confocal laser scanning microscopy in vivo, Skin Res. Tech ., 8, 52- 56, 2002 7.9 14.5 group, shown in the adjacent panel. 4.6 5. Cotton, S.D., Claridge, E. (1996) Developing a predictive model of human skin 10. Matts, PJ, Understanding and measuring the optics that drive visual perception colouring. Proceedings of the SPIE Medical Imaging, vol 2708, 814-825, 1996 of skin appearance, in The Essential Stratum Corneum , Dunitz , 213-222, 2002 8.4 6.9 4.1 14.0 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Figure 2: Mean internal melanin SIAgraph standard deviation as a Figure 4: Mean melanin total spot count (within field of view) as a Figure 6: Mean internal haemoglobin SIAgraph standard deviation as Figure 8: Mean internal collagen SIAgraph standard deviation as a function of age (equal 5-year cohorts, age 10-70) function of age (equal 5-year cohorts, age 10-70) a function of age (equal 5-year cohorts, age 10-70) function of age (equal 5-year cohorts, age 10-70) This research funded by P&G Beauty.