1.

Introduction

Skin aging is the continuous change of the skin’s structure and functional capacity. In addition to the “normal” course of intrinsic or chronological aging, environmental (e.g., sun light) and life style (e.g., smoking) factors contribute to the so-called extrinsic skin aging. Skin areas that are predominantly exposed to ultra violet radiation (e.g., face, hands, and dorsal aspect of the forearm) are prone to develop various signs of extrinsic aging. These include deep wrinkles, loss of elasticity, and changes in skin pigmentation. Histologically, extrinsically aged skin shows degenerated collagen and elastic fibers in the upper and middermis, which is referred to as solar elastosis.^1–3

Facial appearance contributes to the perceived age of individuals and is an indicator of overall health.^4,5 Thus, there is growing interest in skin aging research and so-called antiaging strategies. The evaluation of cosmetic products and rejuvenation procedures is often based on histological assessments that provide knowledge of the morphological and molecular characteristics of the skin.^6,7 However, the invasive procedure to obtain biopsies contains disadvantages, such as tissue destruction, pain and an increased risk for infections.

Optical coherence tomography (OCT) allows a noninvasive and rapid cross-sectional visualization of the skin. It is a widely applied imaging method in many fields of dermatology,^8,9 for instance, to follow up wound healing, to diagnose skin cancer, or to assess the effectiveness of skin care products.^10–13 To date, only a few studies have used OCT to investigate skin aging. The most often reported parameter in skin aging research measured in OCT images is epidermal thickness (ET). ET has been often observed to decrease with age on intrinsically and extrinsically aged skin areas.^14–16 However, OCT images obviously provide much more information than ET only. Therefore, the idea emerged, whether there are other traits or criteria that might be used to characterize signs of skin aging in OCT images. The aim of this study was to develop and to evaluate possible additional parameters to measure skin aging using OCT.

2.

Methods

3.

Results

3.3.

OCT Item Scores

All seven items were detected and could be quantified in our sample. Figure 1 shows the OCT images representing the lowest (left column) and highest (right column) mean score per item.

Fig. 1

Optical coherence tomography images selected among the 16 participants representing the lowest and highest mean scores per item and image. Mean scores are the averages of the scores of three raters. Scale bar $500\mu \mathrm{m}$.

3.3.1.

Construct validation: skin aging

The grading results of the OCT images are shown in Table 3. Median scores ranged from 0.8 for the item “surface unevenness” to 2.5 for the item “upper dermal reflectivity.” No floor and ceiling effects were observed.

Table 3

Grading results of the developed OCT items.

	Young group				Old group				p valuea
	x˜	Interquartile range (IQR)	min; max	po (%)	x˜	IQR	min; max	po (%)
SC reflectivity
Inner upper arm	1.3	1.0 to 1.9	0.7; 2.3	79.1	1.5	1.3 to 2.0	1.0; 2.0	70.8	0.450
Volar forearm	1.7	1.3 to 1.7	1.0; 1.7	70.8	2.0	1.7 to 2.3	1.3; 2.3	66.6	0.020
Dorsal forearm	1.3	1.0 to 1.7	1.0; 2.0	70.8	1.7	1.7 to 2.2	1.7; 2.3	66.6	0.024
Epidermal reflectivity
Inner upper arm	1.3	1.1 to 1.3	1.0; 1.7	62.5	1.3	1.3 to 1.7	1.0; 2.0	62.5	0.244
Volar forearm	2.0	1.4 to 2.0	1.3; 2.3	45.8	1.3	1.1 to 1.3	1.0; 1.7	75.0	0.007
Dorsal forearm	1.7	1.4 to 2.0	1.0; 2.0	25.0	1.3	0.8 to 1.6	0.7; 2.0	58.3	0.083
Upper dermal reflectivity
Inner upper arm	2.0	1.7 to 2.3	1.0; 2.7	70.8	1.7	1.2 to 2.0	1.0; 2.0	62.5	0.206
Volar forearm	2.5	2.1 to 2.7	1.3; 3.0	75.0	1.7	1.1 to 2.0	0.7; 2.3	79.1	0.013
Dorsal forearm	2.2	2.0 to 2.3	1.3; 2.7	79.1	1.7	1.3 to 1.9	1.0; 2.0	54.1	0.013
Lower dermal reflectivity
Inner upper arm	1.7	1.4 to 1.7	1.3; 1.7	66.6	1.3	1.1 to 1.6	1.0; 2.0	79.1	0.090
Volar forearm	1.7	1.3 to 2.0	1.0; 2.3	50.0	1.5	1.0 to 1.7	0.7; 2.3	62.5	0.308
Dorsal forearm	1.5	1.3 to 1.7	1.3; 2.0	62.5	1.3	1.3 to 1.9	1.3; 2.3	70.8	0.907
Dermoepidermal contrast
Inner upper arm	2.0	1.8 to 2.6	1.7; 2.7	79.1	1.8	1.4 to 2.0	0.7; 2.3	79.1	0.128
Volar forearm	1.3	0.7 to 1.9	0.7; 2.0	79.1	2.2	1.5 to 2.3	0.7; 2.7	75.0	0.037
Dorsal forearm	1.0	0.7 to 1.3	0.3; 3.0	83.3	1.5	1.1 to 1.9	1.0; 3.0	70.8	0.047
Vessel density
Inner upper arm	1.7	1.1 to 2.0	0.7; 2.3	79.1	1.7	1.3 to 2.5	1.0; 3.0	66.6	0.632
Volar forearm	1.0	0.5 to 1.3	0.0; 1.7	58.3	2.0	0.8 to 2.3	0.3; 2.3	83.3	0.111
Dorsal forearm	0.7	0.3 to 0.9	0.0; 1.3	62.5	1.2	0.5 to 1.6	0.3; 2.3	75.0	0.088
Surface unevenness
Inner upper arm	0.8	0.1 to 1.6	0.0; 2.0	83.3	1.8	1.2 to 2.3	0.7; 3.0	79.1	0.030
Volar forearm	1.2	0.8 to 1.3	0.3; 2.0	66.6	1.8	1.0 to 2.6	1.0; 2.7	83.3	0.097
Dorsal forearm	1.5	0.3 to 1.7	0.0; 1.7	70.8	1.2	0.8 to 1.3	0.0; 2.0	83.3	0.669

Note: x˜, median; IQR, interquartile range; min, minimum; max, maximum; po, interrater agreement; and SC, stratum corneum.

^aMann–Whitney U test.

Median “SC reflectivity” was higher in older subjects on all three skin areas with differences being statistically significant for the forearm sites. “Upper dermal reflectivity” was significantly lower in the old group on both forearm areas. “Dermoepidermal contrast” and “vessel density” were higher in the old group on the forearm and comparable on the inner upper arm. “Epidermal reflectivity” and “lower dermal reflectivity” were reduced in the old group without reaching statistical significance. “Surface unevenness” increased with age on the inner upper arm and volar forearm and decreased on the dorsal forearm.

Differences between the skin areas were observed for “dermoepidermal contrast,” “vessel density,” and “surface unevenness.” Other item scores like “lower dermal reflectivity” and “upper dermal reflectivity” showed minor differences between the skin areas.

3.3.2.

Interrater agreement

Interrater agreement is presented in Table 3. The highest proportions of agreement were obtained for the items “SC reflectivity,” “upper dermal reflectivity,” “dermoepidermal contrast,” and “surface unevenness” ($p_{\mathrm{o}}=83.3\%$). The lowest proportion of agreement were observed for the item “epidermal reflectivity” ($p_{\mathrm{o}}=25.0\%$).

3.3.3.

Criterion validation: skin microtopography

The mean surface roughness parameters of the two age groups are presented in Table 4. $R_z$ ranged from $38.2\mu \mathrm{m}$ on the volar forearm in the young group to $55.0\mu \mathrm{m}$ on the dorsal forearm in the old group. $R_{\max }$ ranged from $48.3\mu \mathrm{m}$ on the volar forearm of the young group to $74.1\mu \mathrm{m}$ on the dorsal forearm of the old group. SDs were always higher in the aged group. Differences between the age groups were statistically significant on the inner upper arm and volar forearm. In both age groups, surface roughness was highest on the dorsal forearm. Spearman correlation coefficients of the item scores of “surface unevenness” were highest with $R_{\max }$ on the inner upper arm and on the volar forearm ($r_{\mathrm{S}}=0.442$ and $r_{\mathrm{S}}=0.434$). The lowest correlation coefficients were observed for $R_z$ and $R_{\max }$ on the dorsal forearm ($r_{\mathrm{S}}=0.101$ and $r_{\mathrm{S}}=0.151$).

Table 4

Skin surface roughness parameters Rz (μm) and Rmax (μm) measured using Visioscan® VC98.

	Young	Old		Correlation with “surface unevenness”
	Mean (SD)	Mean (SD)	p valuea	rS
$R_z$
Inner upper arm	39.1 (6.2)	53.3 (9.5)	0.003	0.412
Volar forearm	38.2 (4.3)	46.5 (7.1)	0.013	0.351
Dorsal forearm	48.6 (3.3)	55.0 (11.2)	0.148	0.101
$R_{\max }$
Inner upper arm	50.0 (7.6)	69.0 (11.1)	0.001	0.442
Volar forearm	48.3 (5.0)	60.5 (9.0)	0.005	0.434
Dorsal forearm	65.0 (5.6)	74.1 (14.5)	0.122	0.151

Note: SD, standard deviation and rS, Spearman’s correlation coefficient.

^aStudent’s t test.

4.

Discussion

This study introduces additional characteristics besides ET to measure skin aging in OCT images. In a sample of young and aged women, different sun-exposed and sun-protected skin areas were measured. Based on the literature and our own experience with OCT imaging, we inductively developed seven potentially observable skin characteristics.

The most common parameter measured in OCT images is ET, for which good correspondence with confocal laser scanning microscopy has been reported.¹⁵ The ET estimates obtained in our study were comparable with previously reported results.^14,16 We were able to detect a slight thinning of the epidermis with age on the inner upper arm that confirms the results of previous studies.^14,25 Nevertheless, ET thinning tends to occur considerably only up to the 30s¹⁶ and seems to be attenuated by the effects of photoaging.^14,15,19

All newly proposed characteristics were detected and could be quantified in our sample. The items had an overall good distribution of the scores and showed no floor and ceiling effects. This indicates that the item scores were suitable to discriminate between different skin characteristics.

Interrater agreement was high for the items “SC reflectivity,” “upper dermal reflectivity,” “dermoepidermal contrast,” and “surface unevenness,” indicating that these items were easier to grade leading to similar rating results. By contrast, agreement among the items “epidermal reflectivity” and “lower dermal reflectivity” was lower indicating that these items might be inadequate for assessing skin aging.

Study results demonstrated that there are age- and site-dependent differences regarding the item scores. We have chosen the construct “skin aging” to evaluate the validity of the rating results. In intrinsically aged skin, the SC turnover rate is much slower²⁶ and corneocytes are larger.^27,28 Moreover, older subjects have a reduced amount of intercellular lipids.²⁹ These age-dependent changes may lead to alterations in the optical properties of the SC by affecting the penetration of light through the SC,³⁰ which might be a possible explanation for the higher “SC reflectivity” on both forearm sites of the old group. Whether the intensive surface reflection is really caused by the structure of the SC or by slightly different refractive indices between air and skin is ambiguous.³¹ However, we observed an obvious difference between young and older skin.

Significant differences between the age groups were also observed on both forearm sites for “upper dermal reflectivity” and “dermoepidermal contrast,” which assumes that these parameters in OCT images do most clearly change with age. The decreased “upper dermal reflectivity” which was pronounced on both forearm sites likely reflects the deposition of disorganized and highly backscattering elastin from the papillary dermis to the reticular dermis and the loss of collagen content resulting in a signal attenuation of the upper dermis.^32,33 In close relation to less “upper dermal reflectivity” on both forearm sites is the substantial increase in “dermoepidermal contrast.” The more the reflectivity of the upper dermis decreases the higher is the ability to distinguish between epidermis and dermis.

“Lower dermal reflectivity” seemed to be neither related to age nor to sun exposure. This finding is likely to be explained by the limited penetration depth and lateral resolution of the OCT signal leading to an increased signal noise and a diffuse appearance of the deeper dermis.

We expected to observe age- and site-dependent differences in “vessel density” because horizontal cutaneous vessels have been shown to be increased in the elderly.³⁴ In addition, there is evidence that photodamaged skin exhibits reduced numbers of dermal vessels compared to intrinsically aged skin.³⁵ On the other hand, an increased scattering coefficient of the dermis can lead to the reduction of dermal reflectivity thus affecting the ability to visualize blood vessels.³⁶ However, an age- and site-dependent change in “vessel density” could be shown in our sample supporting the construct validity of this item.

We observed an increase in “surface unevenness” in older subjects indicating a rougher and more folded skin surface in this age group. Epidermal changes and dermal degeneration processes during aging are considered to promote increased skin roughness.^37,38 We could confirm this finding by Visioscan® measurements. $R_z$ and $R_{\max }$ are regarded as the most reliable roughness parameters for quantification of the skin surface topography²² and were, therefore, chosen as reference standards for criterion validation in this study. Mean $R_z$ and $R_{\max }$ estimates were similar to previous findings in young and old women.^22,23 All measured roughness estimates increased with age in this study, which is in line with previously reported studies.^39,40 Validity of the item “surface unevenness” was demonstrated by correlation with both roughness estimates. The item scores were associated with $R_z$ and $R_{\max }$ on the inner upper arm and volar forearm, supporting that these estimates measure a similar concept on these sun-protected skin areas. In conjunction with a slightly lower surface unevenness score on the photoaged dorsal forearm skin in the old group, our results support the hypothesis that chronic UV exposure reduces the effect of intrinsic aging on this skin area,⁴¹ indicating that sun damage counteracts increasing roughness during aging. Furthermore, the limited range of $R_z$ und $R_{\max }$ values at the dorsal forearm attenuates observed validity coefficients thus explaining the low correlation.