1.

Introduction

Allergic contact dermatitis is a significant cause of cutaneous disease affecting many individuals both at home and at work. In Germany, about 7% of the general population suffer from contact allergies,¹ a prevalence similar to that seen in America.² Patch testing is the most useful diagnostic tool for the evaluation of patients with suspected allergic contact dermatitis. Dermatologists routinely evaluate patch tests by means of visual inspection and palpatory findings. Typical clinical endpoints include erythema, infiltration, edema, and vesicle formation, some or all of which are rated on a simple scale^{3, 4} (e.g., 0, +, ++, +++). Such approaches can be criticized as subjective, of poor reproducibility, lacking in sensitivity, and being highly variable between observers and/or institutions. As a result, instrumental methods of assessment have been strongly promoted and do indeed offer several advantages, not least their objectivity.^{5, 6}

Numerous studies have been performed to find more objective methods of evaluating patch test reaction. For example, the following bioengineering methods have been assessed: transepidermal water loss measurement, laser Doppler flowmetry for measurement of cutaneous blood flow, assessment of erythema via colorimetry, IR thermography, and $20\text{-}\mathrm{MHz}$ ultrasound A-scans to measure skin thickness.^{5, 6, 7} However, innovative skin imaging techniques such as confocal laser scanning microscopy and optical coherence tomography (OCT) may be of particular interest because of the high resolution achieved by these techniques, which enable noninvasive visualization of micromorphological structures.^{8, 9} In contrast to ultrasound, OCT uses IR light instead of sound waves. It employs low-coherence interferometry to produce cross-sectional, 2-D or 3-D images of optical scattering from internal tissue microstructures. OCT enables imaging of skin layers up to $1\mathrm{mm}$ . Hence it is particularly suited for presenting morphological features of the epidermis and upper dermis. Briefly, interference fringes are formed when the optical path length of light reflected from the sample matches that given by the reference arm within the coherence length of the light source. The axial depth (A-scan) is obtained by scanning the reference arm length, resulting in localized interference fringes with amplitudes related to sample reflectivity. The fringe intensities in adjacent A-scans are combined to form a 2-D image (B-scan). The source coherence length and the diameter of the beam focus on the sample determine the depth resolution and lateral image resolution, respectively. A lateral resolution of the order of about $10\mu \mathrm{m}$ is typical for conventional OCT scanners. Furthermore, spectroscopic, elastographic, Doppler, and polarization-sensitive functions provide distinct, complementary information to conventional structural OCT. Importantly, there are no contraindications or adverse effects^{9, 10, 11, 12, 13} for OCT.

OCT is increasingly used in clinical and experimental dermatology. Correlation of OCT images with histology recently confirmed observation of anatomic structures such as different skin layers and appendages, including hair follicles and eccrine ducts. Pathological changes have also been described, such as blistering, tumor tissue, and inflammatory conditions including psoriasis and UV-induced dermatitis.^{13, 14, 15, 16, 17, 18, 19} This explorative pilot study was designed to evaluate for the first time acute allergic contact dermatitis (AACD) by means of OCT, and to correlate the clinical grading of patch test reactions with the findings obtained from OCT.

2.

Methods

3.

Results

In the 10 patients selected for further OCT assessments, we observed 20 allergic test reactions, for example, to p-phenylendiamine $(n=3)$ , nickel sulfate $(n=2)$ , colophony $(n=2)$ , and bufexamac $(n=2)$ . One patient had four positive patch test reaction, one patient had three, five patients had two, and the remaining three had one test reaction. All allergic reactions observed were caused by allergens that were prepared in petrolatum as a vehicle. In OCT images of the petrolatum-treated control sites, the skin surface showed a bright homogenous entrance signal without interruptions (first peak of A-scan). The superficial flat layer below the entrance signal corresponded to the epidermis. The border with dermis was usually sharply demarcated showing a highly reflective band of horizontally orientated collagen bundles (second peak of A-scan). Frequently, the dermis was more signal-intense than the epidermis. Some signal-free longish cavities, corresponding to blood vessels, were also observed in the dermis (Table 1). By contrast, OCT of AACD more frequently showed pronounced folding at the skin surface and a thickened and/or disrupted entrance signal mostly accompanied by splitting of the first peak of the A-scan. However, there was no correlation between the clinical scoring of AACD and the degree of thickening and/or splitting of the entrance signal observed in the B-scan ( $r=-0.16$ ; 95% CI: $-0.56$ to 0.31; $P=0.51$ ). Evaluation of the A-scans of AACD revealed in almost all cases the loss of the second peak. Hence ET could be determined only in the B-scan using the SkinDex 300 measure tool. In allergen-treated skin (median ET: $77\mu \mathrm{m}$ ) we observed a significantly ( $P<0.001$ ; difference between medians: $13.6\mu \mathrm{m}$ ; 95.2% CI: $-18$ to 10.5) thicker epidermis as compared to control sites (median ET: $65\mu \mathrm{m}$ ). The increase of ET in allergen-treated skin did not correlate with the clinical scoring, however ( $r=0.29$ ; 95% CI: $-0.22$ to 0.68; $P=0.26$ ). The most striking epidermal alteration found in AACD was the visualization of extensive, clearly demarcated signal-free cavities. The occurrence and size of signal-free cavities strongly correlated with clinical scoring ( $r=0.81$ ; 95% CI: 0.57 to 0.93; $P<0.001$ ). While weakly positive allergic reactions usually showed no epidermal cavities in the OCT image, moderate and strong allergic reactions were accompanied by clearly demarcated cavities within the epidermis, partly even leading to an elevation of the skin surface [Figs. 1(c) and 1(d) ]. In contrast to the control sites, allergic skin reactions displayed a remarkable decrease of reflectivity in the dermis. In extreme cases, there was an almost complete loss of signal intensity even in the upper dermis. The decrease of reflectivity strongly correlated with the clinical patch test scoring ( $r=0.74$ ; 95% CI: 0.44 to 0.89; $P<0.001$ ).

Fig. 1

OCT images of a patient who had several positive patch test reactions. OCT shows a typical presentation of normal skin with a relatively clear demarcation of the epidermis and dermis. The latter displays strongly reflective, partly horizontally arranged structures corresponding to collagen bundles (control, A). The B-scan of the weak patch test reaction shows more pronounced folding at the skin surface (∧). In comparison to the control site, a reduction of reflectivity in the dermis is evident (>). Furthermore the border between the epidermis and dermis appears more diffuse (+, lanolin alcohol, B). The image of the moderate test reaction demonstrates signal-free epidermal cavities (vesicles), thickening and splitting of the entrance signal (focal hyper- and parakeratosis), and considerable reduction of dermal reflectivity (++, ammoniated mercury, C). The OCT scan of the strong test reaction clearly demonstrates small to large signal-free epidermal/dermal cavities (spongiosis, vesicles, blisters). An almost complete loss of dermal reflectivity possibly due to edema is evident as well (+++, bufexamac, D).

Table 1

Data of OCT measurements in correlation to clinical patch test grading.

4.

Discussion

Investigation of AACD is of significant interest in dermatology and occupational medicine. However, most current methods to evaluate AACD are limited by the use of indirect techniques that may be inaccurate or insensitive. Noninvasive imaging techniques for the investigation of AACD enable the skin to be imaged as frequently as required. Standard noninvasive methods such as magnetic resonance imaging and ultrasound, however, provide fairly undifferentiated information.^{1, 2, 3, 4, 5, 6} By contrast, OCT is capable of yielding more micromorphological details than the other methods mentioned. Therefore, it may be a promising tool for the evaluation of inflammatory skin conditions. Pagnoni ¹⁴ previously investigated the effect of dimethyl sulfoxide on human skin using a prototype of the SkinDex 300. They observed an increase in surface folding, a dark vacuolization within the epidermis, and hyporeflectivity and decrease of signal attenuation in the dermis. The latter was also observed following the application of histamine or nictotinic acid.¹⁶ Recently, sodium lauryl sulfate-induced contact dermatitis has been studied by Welzel ⁹ using OCT in vivo. Following skin irritation with sodium lauryl sulfate, they observed larger and more irregular superficial skin folds, a more pronounced entrance signal, increase of ET, and dilated blood vessels in the dermis. Comparison of the aforementioned findings with our results is however difficult, since Welzel ⁹ studied an experimental model of acute irritant contact dermatitis (AICD) and did not report on the clinical grading of skin reactions. Recently, it has been shown in confocal microscopy studies in vivo that superficial epidermal changes, including striking parakeratosis, stratum corneum disruption, and superficial necrosis, are highly indicative of irritant reactions. Though an increase of ET may be observed in AACD, the degree is usually rather mild as compared to AICD. On the other hand, vesicle formation and edema in the upper dermis appear to be more severe in AACD than in AICD. Hence the differentiation between AACD and AICD might be possible by means of confocal microscopy in vivo.^{8, 19} Nevertheless, even the value of histological examination to differentiate between the distinctive forms of contact dermatitis is limited.^{20, 21}

In this study, we found that AACD tended to generate larger folds at the skin surface probably due to edema in the epidermis. Further, we observed thickening and/or splitting of the entrance signal possibly indicating slight to moderate hyper- and parakeratosis caused⁹ by AACD [Fig. 1(c)]. This finding, however, did not correlate with the clinical scoring. The stratum corneum of body skin is normally not visible on conventional OCT. However, in particular pathological circumstances such as hyper- and parakeratosis occurring in psoriasis or sunburn, the stratum corneum may be resolved by means of commercially OCT scanners.^{9, 17} Even though statistically significant, the increase of ET (difference between medians: $13.6\mu \mathrm{m}$ ) observed in AACD was moderate from a clinical point of view and did not correlate with clinical severity. We presume that the increase of ET following AACD may be due to intercellular edema as well as hyperproliferative processes induced by the release of inflammatory cytokines.^{4, 9}

The most striking finding of the present study represents the visualization of well-demarcated signal-free cavities within the epidermis.²² These findings very likely correspond to coalescing spongiosis and vesicle formation which is a typical feature^{4, 20, 21} of AACD. Moreover, we observed a strong correlation between the clinical scoring and decrease of reflectivity in the dermis. Possibly, the latter can be explained by a reduction of backscattering in the upper dermis due to interstitial edema, cellular infiltrates, and dilation of blood vessels. Hence within the patch test reactions, the dermis appeared much darker due to reduced scattering. The latter is mainly influenced by the size, shape, and composition of particles. The regular arrangement of collagen in the upper dermis leads to a strong backscattering.^{18, 19} Any alteration in the orientation of the collagen fibers or the structural composition of the dermis influences its optical properties. As in our data, Raju ⁷ recently observed in a high-frequency ultrasound study on patch test reactions a significant increase in skin thickness and decrease in echogenicity of the upper dermis.

The major limitation of the presented study, however, is that there are no biopsies for histopathologic confirmation of our findings. For example, whether the regions with clear demarcation and reduced reflecting signals are due to vesiculation, or if the increase of ET is due to edema and hyperproliferation is more or less based on speculations only indirectly supported by previous OCT studies including histologic investigations.^{9, 13, 18, 22, 23, 24} Since the clinical patch test readings have major limitations, the study lacks of a “gold standard” by which to compare the accuracy and validity of OCT readings. Therefore, it remains uncertain whether the concordance or discordance between OCT and clinical evaluation adds value to diagnosis of AACD. In conclusion, we demonstrated definite OCT findings in patients with AACD including pronounced skin folding, thickening, or disruption of the entrance signals, and increase in ET. Moreover clearly demarcated signal free cavities and reduction in dermal reflectivity were detected by OCT which significantly correlated with clinical patch test grading. Even though OCT appears to be a useful technique to evaluate AACD, clinical correlations in the future will be needed to determine the true value of OCT in this setting.²⁵