Archives of Dermatological Research

, 297:218

Acute skin alterations following ultraviolet radiation investigated by optical coherence tomography and histology

Authors

    • Department of DermatologyRuhr-University Bochum
  • Stefanie Boms
    • Department of DermatologyRuhr-University Bochum
  • Markus Stücker
    • Department of DermatologyRuhr-University Bochum
  • Georg Moussa
    • Department of DermatologyRuhr-University Bochum
  • Alexander Kreuter
    • Department of DermatologyRuhr-University Bochum
  • Michael Sand
    • Department of DermatologyRuhr-University Bochum
  • Daniel Sand
    • Department of Physiological ScienceUniversity of California
  • Peter Altmeyer
    • Department of DermatologyRuhr-University Bochum
  • Klaus Hoffmann
    • Department of DermatologyRuhr-University Bochum
Original Paper

DOI: 10.1007/s00403-005-0604-6

Cite this article as:
Gambichler, T., Boms, S., Stücker, M. et al. Arch Dermatol Res (2005) 297: 218. doi:10.1007/s00403-005-0604-6

Abstract

Optical coherence tomography (OCT) appears to be a promising technique to study skin in vivo. As part of an exploratory study to investigate UV induced effects non-invasively we aimed to evaluate the kinetics of acute UVB- as well as UVA1 induced skin alterations by means of OCT, and to correlate the results obtained with routine histology. Twelve healthy subjects received daily 60 J/cm2 of UVA1 and 1.5 minimal erythema doses of UVB on their upper back over three consecutive days. One day (24 h) after the last UV exposure, OCT measurements and skin biopsies were performed in four subjects (day 1) on the centre of the irradiated sites and an adjacent non-irradiated control site. The same procedure was performed in four subjects 3 days and 6 days after irradiation, respectively. Prior to OCT assessment two waterproof marks were drawn on the centre of UVB and UVA1 exposed sites and the control site. The OCT scanner, SkinDex 300, was used in the RI1D measurement modus in order to investigate morphological features, epidermal thickness, and scattering coefficients. Immediately after OCT assessment, 4 mm punch biopsies were taken from the previously marked sites. OCT as well as histological examinations performed on day 1, 3, and 6, revealed markedly higher values for epidermal thickness on UVB exposed skin sites, and slightly increased epidermal thickening in UVA1 exposed sites. UVB exposed sites showed disruption of the entrance signal in the B-scan of OCT resulting in a thickened layer with a signal-poor centre corresponding to hyperkeratosis and parakeratosis as confirmed by routine histology. Surprisingly, the mean scattering coefficients of the epidermis were slightly lower on UVA1 exposed sites, as compared to non-irradiated skin. By contrast, the scattering coefficient of the upper dermis of UVA1 irradiated skin was hardly altered. Moreover, the scattering coefficient of the upper dermis assessed on UVB exposed skin on day 1 was clearly smaller than the scattering coefficient observed on non-irradiated and UVA1 exposed skin. Conclusively, it was possible to demonstrate by means of OCT differences of epidermal thickness and pathological features of the stratum corneum following UV exposure. UVA1 induced epidermal pigmentation as well as UVB induced dermal inflammation may affect the light attenuation in the tissue indicated by a decrease of the scattering coefficient. OCT seems to be a useful tool to monitor UV induced effects in vivo.

Keywords

InterferometrySkin morphologyPhotoprovocationPhotoadaptionUltraviolet radiationPhotodermatology

Introduction

Acute effects of ultraviolet (UV) radiation on skin morphology include erythema as a result of vasodilation, intercellular oedema because of increased vasopermeability, pigmentation, and skin thickening. UV induced epidermal changes such as acanthosis and stratum corneum thickening have intensively been investigated by means of routine histology [16]. Since skin biopsy alters the original skin morphology, it cannot be performed repeatedly on the same site. Besides, it always requires an iatrogenic trauma, hence non-invasive methods may be of great advantage that allow one to study the skin in vivo. In the past decade, advances have been made in imaging human skin in vivo by high-frequency ultrasound, confocal laser scanning microscopy, and optical coherence tomography (OCT) [7].

Optical coherence tomography works analogously to an ultrasound scanner—the major difference is, however, that ultrasound pulses are replaced by a fiber-optic Michelson interferometer with a low-coherence-length broadband light source. The corresponding short coherence length permits a spatial resolution in depth direction of less than 10 μm. The lateral resolution is given by the numerical aperture of the employed objective as long as single scattering prevails. Depending on the scattering properties of tissue and some accepted loss in resolution, a penetration depth of about 1 mm can be achieved. The source coherence length and the spot size of the beam focus on the sample determine the depth resolution and lateral image resolution, respectively. New systems with ultra-high resolution of about 1 μm have recently been developed, however lateral resolution of approximately 10 μm is more typical. OCT is particularly capable of presenting morphological features of the epidermis and upper dermis. It can provide cross-sectional high resolution images of structures below the tissue surface in analogy to histology [810]. In previous studies, different skin layers such as the stratum corneum, the epidermis and the dermis including skin appendages, e.g., hair follicles and eccrine ducts, were identified by means of OCT. Correlation of OCT images with histology confirmed observation of morphological changes such as blistering, tumour formation, and inflammatory conditions including psoriasis and contact dermatitis [1014]. Apart from skin morphology OCT is capable of probing other interesting tissue parameters. For example, birefringence for the characterization of ordered structures from collagen and elastin fibres has recently been described in polarization-sensitive OCT studies [15]. Further possible dermatological applications of OCT elastography include differentiation of hard and soft masses and evaluation of wound healing [16]. Doppler flow imaging in vessels becomes increasingly popular because of the inherent sensitivity to frequency shifts in the detected OCT signal [17]. Moreover tissue characterization by local biophysical parameters such as refraction and scattering can be desired for dermatological research. Alterations of skin hydration, for example, can be evaluated by means of the refraction index and scattering coefficient [18, 19].

Non-invasive imaging techniques employed to study UV induced effects on human skin could present a major progress in the field of photobiology and general dermatology. As part of an exploratory study to investigate UV induced effects non-invasively we aimed to evaluate the kinetics of acute UVB as well as UVA1 induced skin alterations by means of OCT in vivo, and to correlate the results obtained with routine histology.

Subjects and methods

Subjects

Twelve subjects (median age: 56.4; range: 35–83 years) with skin type II (Fitzpatrick classification) were enrolled the study after giving their informed consent [20]. The study was approved by the Ethical Review Board of the Ruhr-University Bochum, Germany. Exclusion criteria included history of photosensitive disorders, relevant drug use, pregnancy and UV exposure for at least 2 months prior to the study. With respect to age and gender of the subjects, we sought to form three comparable groups, each including four subjects.

UV irradiation

UV irradiation was performed in accordance with a commonly employed phototesting and photoprovocation protocol, respectively [21]. Prior to the study, UV source analysis was carried out with the spectral radiometer MSS 2040 (MSS Elektronik GmbH, Fröndenberg, Germany). The minimal erythema dose (MED) for broadband UVB (Fig. 1a) and broadband UVA was determined on the subject’s buttock with the Saalmann Multitester SBB LT 400 (Saalmann GmbH, Herford, Germany). UVB intensity measured with the RM-11 radiometer (Dr. Gröbel, Ettlingen, Germany) was 4.2 mW/cm2. UVA intensity was 23 mW/cm2 (UV-Meter, Waldmann, Willingen-Schwenningen, Germany). The UVB doses for MED determination ranged 0.032–0.08 J/cm2, and the UVA doses ranged 7.2–18 J/cm2. The MEDs were read after 24 h.
https://static-content.springer.com/image/art%3A10.1007%2Fs00403-005-0604-6/MediaObjects/403_2005_604_Fig1_HTML.gif
Fig. 1

Spectral irradiance of the UV sources used for the application of 4.5 MED-UVB (erythemally-weighted broadband UVB, a) and 180 J/cm2 UVA1 (b)

On the following 3 days, the subjects received daily 1.5 MED-UVB (Saalmann Multitester SBB LT 400, Herford, Germany) as well as 60 J/cm2 UVA1 (Fig. 1b) with the partial body light source Sellamed 2000 System Dr Sellmeier (Sellas, Gevelsberg, Germany) on two opposite sites (7×7 cm) on the right and left medial scapular region, respectively. The Sellamed 2000 light source is a high-pressure metal halide lamp filtered with a UVA1 filter (Sellas) and an infrared absorbing filter UG1 (Schott, Mainz, Germany). Spectral output consisted exclusively of UV wavelength greater than 340 nm and smaller than 450 nm. UVA1 intensity at body distance, measured with MP-100 OPTICAL Radiometer (UVA1-MED, Wennigsen, Germany), was 31.4 mW/cm2. The resulting cumulative UV doses were 4.5 MED for UVB and 180 J/cm2 for UVA1.

OCT and histology

One day (24 h) after the last UV exposure, OCT measurements and skin biopsies were performed in four subjects (day 1) on the centre of irradiated (UVB, UVA1) sites on the medial scapular region and on a laterally adjacent (2.5 cm) non-irradiated site. The same procedure was performed in four subjects 3 days and 6 days after irradiation, respectively. Prior to OCT assessment two waterproof marks 4 mm in distance were drawn on the centre of UVB and UVA1 exposed skin and the control site. A commercial OCT scanner (SkinDex 300, ISIS optronics GmbH, Mannheim, Germany) was used in this study. The performance of this system regarding spatial resolution and field-of-view is as follows. A bandwidth Δ λ=70 nm and a central wavelength of λ0=1,300 nm is utilized. Under the assumption of an average refractive index of the sample medium nmed=nobj=1.43, the coherence length for depth resolution A-FWHMInt=7.4 μm. The numerical aperture of the focusing lens is NA=0.19. Thus the diffraction limited lateral resolution yields A-FWHMFoc=4.5 μm. The architecture of the system with eight parallel scanning channels allows fast scans. Within 2 s a number of 512 scans are acquired over the length of 1 mm in lateral a direction and an axial range of 0.9 mm. Echo signals are digitised with 14 bits amplitude resolution. An integrated CCD camera with a field-of-view of 4.5 mm2 delivers optical images of the skin surface. With the aid of these images it was possible to perform OCT assessments exactly on the skin site previously marked. Pre-treatment of each measurement site for index matching was performed using conventional ultrasound gel. The RI1D measurement modus of the SkinDex 300 provides algorithms for the determination of epidermal thickness as well as the scattering coefficient (μs). In order to determine epidermal thickness (ET), we semi-automatically performed distance measurements between the entrance peak (skin surface) and the valley prior to the second intensity peak (dermo-epidermal junction) of the averaged A-scans. This was done by using the integrated OCT software including the cursor and mouse click function [22]. Since the absorption coefficient (μa) is much smaller than μs for tissues in the near-infrared, absorption could have been neglected throughout the paper. When a single beam of light with starting intensity (I0) hits a given material, μs contains the information, how much light is scattered over depth (z) and, how much light can pass through it respectively. The remaining intensity over depth, I(z), can be calculated. Looking at logarithmic intensity curves, one can acquire local μs with linear fits which an integrated SkinDex 300 algorithm provides. Determination of μs of the regions of interest (viable epidermis, upper dermis) was based on the following formula: I(z) = I0 × exp(−μs × z). Qualitative OCT image analysis of irradiated and non-irradiated skin was performed on the averaged B-scan by viewing the images of interest on the screen side by side. In all scans, we used the same image modalities (two-sided threshold operation: 50 dB/5 dB). To minimize inter-observer variability all OCT measurements were performed by the same investigator (T. G.).

Immediately after the OCT assessment, 4 mm punch biopsies were cautiously taken from the previously marked sites under local anaesthesia (1% lidocaine subcutaneously). The excised tissue was fixed in formalin solution and embedded in paraffin. Histological slices of 5 μm were performed for routine haematoxylin and eosin staining. For further evaluation, we selected one histology slice of each patient showing the best quality with regard to preservation of the horny layer and absence of artefacts. The thickness of the maximum epidermis defined by the valleys of the papillae was measured by the same investigator (S. B.) on five random chosen places in the histological preparation at magnification × 40. The mean value of the five measurements was calculated for each subject. In addition, histopathologic alterations such as hyperkeratosis, parakeratosis, acanthosis, basal vacuols, dyskeratosis, and perivascular infiltrates with oedema were graded by means of a simple score as follows: − = none, (+) = slight, + = moderate, ++ = strong. In view of the small sample size per group we refrained from the use of analytical statistics, however, data of ET and μs measurements have been expressed as mean ± SD.

Results

In all subjects, well-demarcated pigmentation was visible on the UVA1 exposed sites at all points of time. Moderate to strong erythema was seen on UVB exposed sites, in particular on day 1. On day 1, 3, and 6, OCT and histological examinations revealed markedly higher values for ET of the UVB exposed skin sites (Table 1). However, the UVA1 exposed sites also showed a slight increase of ET at all times of examination. The marked increase of ET of the UVB exposed sites, particularly found on day 1 and 3, was accompanied by hyperkeratosis, parakeratosis, and acanthosis as confirmed by histology. Histologic features such as dyskeratosis, basal vacuolisation, acanthosis, and/or mild intercellular oedema were also most prominent in UVB exposed skin sites investigated on day 1 and 3 (Table 1). Interestingly, UVB exposed sites showed disruption of the entrance signal in the B-scan of OCT resulting in a 15–30 μm thick layer with a signal-poor centre (Fig. 2). Stratum corneum thickening was confirmed by histology. The same phenomenon, less marked, was also observed in the UVA1 exposed skin sites of three subjects (day 1 and 3). The disruption of the entrance signal was also indicated by splitting of the first peak of the A-scan. In all cases, this effect was most prominent on day 1 and 3.
Table 1

Data of OCT measurements and histology following UVA1 (180 J/cm2) and UVB (4.5 MED) irradiation on the back of 12 subjects

Day

EP

Epidermal thickness (μm)

Disruption of entrance signal

Scattering coefficient

Hyperkeratosis

Parakeratosis

Acanthosis

Dyskeratosis

Perivascular infiltrates and oedema

OCT

Histology

OCT

OCT viable epidermis

OCT upper dermis

Histology

Histology

Histology

Histology

Histology

mean ± SD

mean ± SD

n.o.s./Score

s) (mm−1)

s) (mm−1)

n.o.s./Score

n.o.s./Score

n.o.s./Score

n.o.s./Score

n.o.s./Score

1 (n=4)

None

91.6±16.9

87.8±15

1/(+)

13.3±4.9

4.8±1.8

1/(+)

UVA1

103.5±12.9

94.5±33.4

1/(+)

10.6±3.7

4.7±2.2

1/(+)

2/(+)

1/+

2/(+)

1/+

UVB

129.3±33.8

117.5±33

1/(+), 2/+, 1/++

12.9± 33.6

2.1±1.3

3/+, 1/(+)

1/++, 2/+, 1/(+)

2/+, 2/(+)

1/++, 2/+

1/++, 3/+

3 (n=4)

None

73.3±12.9

68.3±26.8

12.6±4.5

3.4±1.4

UVA1

85.5±14.4

75.3±19

2/(+)

9.2±3.8

3.9±2.1

1/(+)

1/(+), 2/(+)

2/(+)

UVB

105.5±10.3

101.5±11.4

1/(+), 1/+, 2/++

11.8±3.5

2.6±.2.9

2/++, 2/+

3/+

4/+

3/+

2/+

6 (n=4)

None

77.1±16.4

87.1±18.6

14.9±3.9

4.7±2.5

UVA1

94.3±6.6

89.7±15.7

10.8±5.3

5.9±2.6

1/+

1/+

1/(+)

UVB

101.3±19.4

92.5±9.2

2/+, 1/++

14.4±4.9

5.1±3.3

4/+

2/(+)

2/++

1/+

2/(+)

EP exposure, n.o.s. number of subjects, score: − = none, (+) = slight, + = moderate, ++ = strong

https://static-content.springer.com/image/art%3A10.1007%2Fs00403-005-0604-6/MediaObjects/403_2005_604_Fig2_HTML.jpg
Fig. 2

OCT of non-irradiated skin (a) and OCT 3 days after 4.5 MED-UVB showing remarkable disruption of the entrance signal with a signal-poor centre (b, arrow heads). Corresponding histology (c) displaying marked parakeratosis and hyperkeratosis

For non-irradiated skin, the mean μs of the epidermis was 13.6 mm−1 and for the upper dermis, 4.3 mm−1. The values of μs of the viable epidermis were slightly lower on UVA1 exposed sites, as compared to UVB exposed and non-irradiated skin. By contrast, the scattering coefficient of the upper dermis of UVA1 irradiated skin was hardly altered (Fig. 3, Table 1). Overall, UVA1 exposed skin showed only minor histologic changes. On day 1, the μs of the upper dermis assessed on UVB exposed skin (2.1±1.3 mm−1) was clearly smaller than μs observed on non-irradiated (4.8±1.8) and UVA1 exposed skin (4.7±2.2). Accordingly, histologic examinations of skin specimens obtained on day 1, revealed slight to moderate perivascular inflammatory infiltrates and mild oedema in the upper dermis of UVB exposed skin sites (Table 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00403-005-0604-6/MediaObjects/403_2005_604_Fig3_HTML.jpg
Fig. 3

OCT image displaying measurements of μs of the viable epidermis (evaluated from dermo-epidermal junction up to 50 μm above) and the upper dermis (evaluated from the dermo-epidermal junction up to 50 μm below) investigated on non-irradiated (a) and UVA1 exposed (b) skin. For non-irradiated skin, μs of the viable epidermis was 10.1 mm−1 and μs of the upper dermis was 3.7 mm−1. UVA1 exposed skin shows a marked drop of μs (7.4 mm−1) in the viable epidermis, whereas μs (3.8 mm−1) of the upper dermis was hardly altered

Discussion

Epidermal hyperplasia and thickening of the horny layer is one of the most important photoadaptive effects occurring after UV exposure [16]. Both in OCT and histology we observed thicker epidermis on UV exposed skin than on control sites. In accordance with a recent OCT study, the effect was however more prominent on UVB exposed sites [14]. Nevertheless, our in vivo measurements using OCT as well as confocal laser scanning microscopy confirm that not only slightly erythemogenic UVB doses, but also suberythemogenic UVA doses may induce thickening of the epidermis [14, 23]. Previous studies have clearly shown that UV induced thickening of the epidermis is predominantly due to an increase of stratum corneum thickness that is also accompanied by parakeratosis [16]. Normally the stratum corneum of the skin on the back is not thick enough to be visible at the resolution used in conventional OCT technique. However, this appears to be different in epidermal alterations such as hyperkeratosis and parakeratosis. Interestingly we have observed more or less marked disruption of the entrance signal in almost all images taken from UVB exposed skin sites. Welzel et al. [10, 12] have previously shown in psoriasis lesions that the entrance signal was more pronounced with parallel layers corresponding to the scaling in psoriasis. Similar to our findings, they observed that the irregularity of the stratum corneum with sheets of parakeratosis led to strong reflectivity zones in the OCT images. Moreover, Del Bino et al. [24] performed immunolabelling for filaggrin, which is a major terminal differentiation marker that is normally located in the granular layers. They found an altered pattern of staining in UVB exposed skin leading to a band-like negative area between the superficial part of the stratum corneum and the granular layer. Accordingly, we assume that the disruption of the entrance signal observed in OCT imaging following UVB exposure is due to a transitory band-like accumulation of parakeratotic cells near the skin surface, which have scattering properties that differ from the full cornified cells of the stratum corneum.

According to the Beer-Lambert law, light attenuation inside tissues is exponential. The slope of this exponential attenuation is proportional to the total attenuation coefficient of ballistic photons (μt = μs + μa, where μs is the scattering and μa is the absorption coefficient, respectively). Because μa ≪ μs for tissues in the near-infrared, the exponential attenuation is proportional to μs, also defined as the product of the number density of the scattering particles and the scattering cross section of the particles, which is in other words the probability per unit length of a photon being scattered. Only ballistic photons backscattered to the OCT system contribute to the OCT image; therefore, by analysing the exponential profile of light attenuation detected by the OCT system, one can obtain information on tissue scattering properties. For OCT, the system’s wavelength has to be such that absorption in the sample is minimised to allow maximal imaging depth. In the tissue, absorption is minimised between 700 and 1,300 nm, in what is known as the window of transmission [25, 26]. Tissue scattering properties are highly dependent on the ratio of the refractive index of scattering centres (e.g., cell membranes, cellular components, and protein aggregates), ns, to the refractive index of the interstitial fluid (ISF), nISF : Δnns/nISF. For example, raising intercellular oedema increases nISF, decreases μs of tissues, and decreases the refractive index mismatch. As clinically confirmed, UVA1 exposures induced an increase of skin pigmentation that was accompanied by a decrease of the mean μs value of the viable epidermis at all points of time evaluated. Welzel et al. [11] did not observe significant changes of the signal attenuation coefficient of the epidermis following a single exposure to 120 J/cm2 UVA that induced an insignificant increase of the melanin index as assessed spectrophotometrically. However, they found a significantly enhanced mean attenuation coefficient in the upper dermis, while we observed hardly any change in μs values for this skin layer. The drop of μs following the increase of pigmentation as observed in the present study is difficult to interpret. Indeed one would rather expect an increase of μs in the epidermis and downstream signal attenuation following significant skin pigmentation, since melanin is known to be a strong scatterer of light. The role of scatter in the attenuation of the downstream signal is likely to have two components, each resulting in a reduction in the light returned to the detector. The first is the deviation of light away from deeper tissues so that less is incident on them. Second, any light that reaches deeper tissues and is returned toward the detector may be further scattered by the melanin and therefore deviated away from the detector [27]. With regard to the impact on scattering, alterations of distribution, size, density and orientation of particles are of higher importance than the concentration of pigment, for example. Hence, one may speculate that the UVA induced redistribution of melanin from the malphigian layers into the stratum corneum may have decreased the light attenuation in the epidermis [28, 29].

A further remarkable finding of our study is the significant reduction of μs of the upper dermis that was observed on day 1 in UVB exposed sites. The decrease of μs was very likely due to the inflammatory alterations, which were histopathologically confirmed by cellular infiltrates and mild oedema in the upper dermis. The latter changes may have influenced the density and distribution of dermal collagen fibres whose regular arrangement normally leads to strong scattering. Any change in orientation of the fibres or the structural composition of the dermis alters the optical properties. As already mentioned, an increase of intercellular water content may have contributed to a decrease of scattering [11, 12, 18]. However, acute UV induced inflammatory reactions usually disappear after about 3 days [1, 4]. Hence, μs of the upper dermis observed on day 3 and 6 were comparable to those of non-irradiated skin.

In the present paper we have demonstrated that OCT is a very promising technique to investigate the influence of UV radiation on the epidermis and upper dermis in vivo. Apart from the characterization of the different skin layers, structural alterations and inflammatory reactions can be evaluated by means of OCT in vivo. OCT may therefore be an interesting non-invasive technique for monitoring over time photoadaptive processes, UV induced inflammation, and phototherapeutic effects. However, data reported here is only of preliminary character and not definitive because of the limited study sample. Hence, further controlled studies on larger study populations are warranted to fully elucidate the validity and reproducibility of OCT in skin research, in particular in photobiological studies.

Acknowledgments

We would like to thank Martin R. Hofmann (PhD) and Tuyen N. Le (graduated engineer) from the Institut für Werkstoffe und Nanoelektronik (Ruhr-University Bochum) for the critical review of our manuscript. This study was performed in cooperation with the Ruhr Centre of Competence for Medical Engineering (KMR, Bochum, Germany) supported by the Federal Ministry of Education and Research (BMBF), grant no. 13N8079.

Copyright information

© Springer-Verlag 2005