活體倍頻顯微術：皮膚內在老化於活性表皮與真皮乳突之定量分析

(1)

୯ҥᆵ᡼εᏢႝᐒၗૻᏢଣӀႝπำᏢࣴز܌!

ᅺγፕЎ!

Graduate Institute of Photonics and Optoelectronics College of Electrical Engineering and Computer Science

National Taiwan University Master Thesis

ࢲᡏ७ᓎᡉ༾ೌǺ!

ҜጥϣӧԴϯܭࢲ܄߄Ҝᆶ੿Ҝ٢ँϐۓໆϩ݋!

Quantitative Analysis of Intrinsic Skin Aging in Viable Epidermis and Dermal Papillae by In Vivo Harmonic

Generation Microscopy ڬߞՙ

Sin-Yo Chou

ࡰᏤ௲௤Ǻ৊௴Ӏ റγ Advisor: Chi-Kuang Sun, Ph.D.

ύ๮҇୯102ԃ11Д November 2013

(2)

(3)

Abstract

Intrinsic aging of skin is a natural process genetically determined and inalterable.

The aging-related changes increase the susceptibility of skin and reduced the skin barrier

function, which increases incidence of inflammatory or infectious skin disorders in the

elderly people. To reveal the changes of intrinsic skin aging, a 1230 nm femtosecond

Cr:forsterite laser was used as the excitation, and the non-invasive harmonic generation

biopsy (HGB) system collecting second-harmonic generation and third-harmonic

generation signals was applied to acquire in vivo images with high spatial resolution from

ventral forearms of 48 Asians with Fitzpatrick skin phototype III or IV. Using the HGB

system, the XYZ image stacks of equally-spaced parallel horizontal sections were

acquired with step size 2 ȝm or 5 ȝm beginning from the stratum corneum into the upper

reticular dermis of the skin. Using XYZ image stacks, 14 parameters were defined to

inspect morphological changes of viable epidermis and mainly dermal papillae in a

three-dimensional point of view.

In the in vivo analysis of viable epidermis thickness, viable epidermis thickness

excluding rete ridge was not found related to neither age nor gender, but viable

epidermis thickness including rete ridge decreased with age. Results together implied

that the dermal-epidermal junction flattened in the aged skin. In the in vivo analysis of

(4)

isolated dermal papilla (IDP), the average, intra-subject SD, and maximum of IDP height

and volume all decreased with age. Only the average of IDP height was found related to

gender. The results indicated that intrinsic aging resulted in the shrinkage of the IDP in

size, and there were only small IDPs remained in the elder skin. In the in vivo analysis of

dermal papillae within dermal papilla zone (DPZ), both depth of DPZ and 3D

interdigitation index decreased with age, but dermal papillae volume ratio within DPZ

increased with age. However, number density of dermal papillae, dermal papillae volume

per unit area, and average volume per dermal papillae were not found related to age. The

results indicated that dermal-epidermal junction became flatter with age, but the dermal

papillae within DPZ were not reduced in volume. The independent dermal papillae

connected to each other quicker with age, and the whole dermal papillae within DPZ

extended in width. Moreover, the analysis of 3D interdigitation index showed the

decreasing interface area of dermal-epidermal junction with age, which might related to

the fragility of the aged skin due to the weakening of epidermal-dermal adherence.

Keywords: harmonic generation microscopy, in vivo imaging, intrinsic aging, viable

epidermis, dermal papillae

(5)

ᄔा ᄔा ᄔा

ᄔा!!!!

ҜጥޑϣӧԴϯࢂ΋ঁҗӃϺ୷Ӣ܌،ۓЪคݤׯᡂޑԾฅၸำǶԴϯ܌೷

ԋޑׯᡂቚуΑҜጥޑ௵ག܄ǴΨ෧১ΑҜጥޑٛៈфૈǴ೭ኬޑ௃ݩቚуΑԴ

ԃΓளډҜጥݹੱ(inflammation)ϷҜጥ໺ࢉ੯ੰޑวғ౗ǶࣁΑඟ៛Ҝጥޑϣӧ

ԴϯǴךॺஒ΋Ѡݢߏࣁ 1230 ڼԯޑ०ࣾሐǺᗔᐌ᡻ҡႜ৔բࣁӀྍǴ٠ஒх֖

Β७ᓎϷΟ७ᓎૻဦޑߚߟΕԄ७ᓎϪТس಍ၮҔܭڗளଯှ݋ࡋޑࢲᡏΓᡏҜ

ጥቹႽǶךॺ΋ӅڗளΑ 48 Ӝ٥ࢪڙ၂ޣޑ߻ᖉϣୁҜጥޑࢲᡏቹႽǴڙ၂ޣޑ

ጥՅϩᜪ(Fitzpatrick skin phototype)ࣣࣁಃΟ܈ಃѤᜪǶךॺவ؂Տڙ၂ޣޑҜጥ

ڗளΑ໔ຯ࣬ӕޑΟᆢНѳϪय़ቹႽׇӈǴቹႽޑڗளҗف፦ቫ(stratum corneum)

΋ޔుΕډ΢ቫᆛރ੿Ҝቫ(reticular dermis)ǴԶቹႽ໔ຯ߾೏೛ۓࣁ 2 ༾ԯ܈ 5

༾ԯǶճҔΟᆢНѳϪय़ቹႽׇӈǴךॺۓကΑ 14 ঁୖኧǴаΟᆢޑᢀᗺٰᔠຎ

ࢲ܄߄Ҝ(viable epidermis)ᆶ੿Ҝ٢ँ(dermal papillae)ޑࠠᄊׯᡂǶ

ӧࢲ܄߄Ҝࠆࡋޑࢲᡏϩ݋ύǴךॺ٠҂ว౜όх֖ᆛ஠(rete ridge)ޑࢲ܄߄

Ҝࠆࡋᆶԃស܈܄ձ࣬ᜢǴԶх֖ᆛ஠ޑࢲ܄߄Ҝࠆࡋ߾ᒿ๱ԃសቚуԶΠफ़Ƕ

ϩ݋่݀སҢ๱੿Ҝᆶ߄Ҝௗӝೀ(dermal-epidermal junction)ӧԴϯޑҜጥύᡂள

ѳڶǶӧېҥ੿Ҝ٢ँ(isolated dermal papilla)ޑࢲᡏϩ݋ύǴېҥ੿Ҝ٢ँଯࡋک

ᡏᑈޑѳ֡ǵԾي኱ྗৡᆶനεॶࣣᒿ๱ԃសቚуԶΠफ़ǶځύѝԖېҥ੿Ҝ٢

ँଯࡋޑѳ֡ᆶ܄ձ࣬ᜢǶԜϩ݋่݀ࡰрǴϣӧԴϯ೷ԋΑېҥ੿Ҝ٢ँޑᡏ

(6)

ࠠᕭ෧ǴԶӧԴϯޑҜጥύѝഭΠၨλޑېҥ੿Ҝ٢ँǶӧ੿Ҝ٢ँ୔ୱ(dermal

papilla zone)ϣ ޑ ੿ Ҝ ٢ ँ ࢲ ᡏ ϩ ݋ ύ Ǵ ੿ Ҝ ٢ ँ ୔ ୱ ϐ ࠆ ࡋ ᆶ Ο ᆢ ΰ ӝ

(interdigitation)ࡰ኱ࣣᒿ๱ԃសቚуԶΠफ़ǴԶ੿Ҝ٢ँ୔ୱϣ੿Ҝ٢ँᡏᑈКٯ

߾ᒿ๱ԃសቚуԶ΢ϲǶόၸǴ੿Ҝ٢ँኧໆஏࡋǵൂՏय़ᑈ܌֖੿Ҝ٢ँᡏᑈ

ک੿Ҝ٢ँѳ֡ᡏᑈ߾҂ว౜ᆶԃសϐ࣬ᜢ܄Ƕ၀ϩ݋ޑ่݀ࡰр੿Ҝᆶ߄Ҝௗ

ӝೀᒿ๱ԃសቚуԶᡂளѳڶǴՠ੿Ҝ٢ँ୔ୱϣ੿Ҝ٢ँޑᡏᑈ٠҂ᒿ๱ԃស

෧ϿǶᒿ๱ԃសቚуǴᐱҥޑ੿Ҝ٢ँᡂளၨזᆶځд٢ँ࣬ೱௗǴԶӧ੿Ҝ٢

ँ୔ୱϣ᏾ᡏޑ੿Ҝ٢ँᡂளၨቨǶԜѦǴΟᆢΰӝࡰ኱ޑϩ݋ᡉҢǴ੿Ҝᆶ߄

Ҝௗӝೀޑय़ᑈᒿ๱ԴϯԶ෧ϿǴ೭ёૈᆶԴϯҜጥޑ੿Ҝᆶ߄ҜᗹߕΚ෧১Ԗ

ᜢǴ٠Ꮴठૄ১ޑԴϯҜጥǶ

ᜢᗖຒ ᜢᗖຒ ᜢᗖຒ

ᜢᗖຒǺǺǺǺ!७ᓎᡉ༾ೌǵࢲᡏԋႽǵϣӧԴϯǵࢲ܄߄Ҝǵ੿Ҝ٢ँ!

(7)

Figure Contents

Fig. 1. A schema of the 1230 nm-based harmonic generation biopsy system; PMT:

photomultiplier tube; DBS: dichroic beam splitter. ... 7 Fig. 2. A Part of a representative XYZ image stack combing SHG and THG signals from a 24-year-old female subject. The image stack recorded different skin layers including stratum corneum (SC), stratum granulosum (SG), stratum spinosum (SS), stratum basale (SB), papillary dermis (PD), and upper reticular dermis (RD). The depths of images were labeled, and the depth of the first frame of the stack that showed the stratum corneum was set as 0 ȝm. SHG and THG signals were represented by pseudo-colors green and purple. Image dimensions were 240 ȝm × 240 ȝm. ... 8 Fig. 3. A chart showed the rectangular estimation, which approximated an object by a pile of cylinders. The volume and surface area of the object could be estimated by measuring the section areas, their circumferences, and step sizes between sections. ... 11 Fig. 4. Four consecutive frames of an image stack from the ventral forearm of a 24-year-old female subject. The cross-sections of isolated dermal papillae were observed as independent round areas (arrows), and connected with each other in the following frame (arrow heads). Image dimensions were 240 ȝm × 240 ȝm. The distances between images were 5 ȝm. SHG and THG signals were represented by pseudo-colors green and purple. ... 14 Fig. 5. Diagrams of (a) isolated dermal papillae and (b) dermal papillae within dermal papilla zone (DPZ), which were denoted by the green areas. Dermal papilla zone was the region between two dotted lines in (b). ... 15 Fig. 6. Diagrams showed the required measurements for calculating the parameters defined in the analysis of dermal papillae within DPZ. The total number, the total interface area, the max occupied section area of dermal papillae within DPZ were specified in (a); the depth of DPZ and the total dermal papillae volume within DPZ were specified in (b). ... 18 Fig. 7. Average thickness of viable epidermis (excluding rete ridge) versus age from the ventral forearms of 47 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.3614 (not statistically significant); P-value for gender = 0.5475 (not statistically significant). In ANOVA, p-value for age group = 0.0107 (statistically significant); p-value for gender = 0.8920 (not statistically significant). .... 26

(10)

ventral forearms of 47 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0487 (statistically significant); P-value for gender = 0.3396 (not statistically significant). Correlation coefficient R to age = -0.293. In ANOVA, p-value for age group = 0.1592 (not statistically significant); p-value for gender = 0.4282 (not statistically significant). NS: not significant. ... 26 Fig. 9. Average of isolated dermal papilla height versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age

= 0.0003 (statistically significant); P-value for gender = 0.0285 (statistically significant).

Correlation coefficient R to age and gender = -0.563. In ANOVA, p-value for age group = 0.0003 (statistically significant); p-value for gender = 0.0975 (not statistically significant).

In two lower figures, data of female and male subjects were separately displayed. ... 28 Fig. 10. Intra-subject standard deviation of isolated dermal papilla height versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 8.315×10^-5 (statistically significant); P-value for gender = 0.4806 (not statistically significant). Correlation coefficient R to age = -0.540. In ANOVA, p-value for age group = 0.0051 (statistically significant); p-value for gender = 0.7703 (not statistically significant). ... 29 Fig. 11. Maximum of isolated dermal papilla height versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 5.077×10^-5 (statistically significant); P-value for gender = 0.3165 (not statistically significant). Correlation coefficient R to age = -0.551. In ANOVA, p-value for age group

= 0.0021 (statistically significant); p-value for gender = 0.5943 (not statistically significant). ... 29 Fig. 12. Average of isolated dermal papilla volume versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0037 (statistically significant); P-value for gender = 0.0930 (not statistically significant). Correlation coefficient R to age = -0.405. In ANOVA, p-value for age group

= 0.0128 (statistically significant); p-value for gender = 0.1604 (not statistically significant). ... 31 Fig. 13. Intra-subject SD of isolated dermal papilla volume versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0006 (statistically significant); P-value for gender = 0.5912 (not statistically significant). Correlation coefficient R to age = -0.482. In ANOVA, p-value

(11)

for age group = 0.0192 (statistically significant); p-value for gender = 0.7484 (not statistically significant). ... 31 Fig. 14. Maximum of isolated dermal papilla volume versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0005 (statistically significant); P-value for gender = 0.3244 (not statistically significant). Correlation coefficient R to age = -0.482. In ANOVA, p-value for age group = 0.0067 (statistically significant); p-value for gender = 0.4859 (not statistically significant). ... 32 Fig. 15. Depth of dermal papilla zone versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0009 (statistically significant); P-value for gender = 0.0989 (not statistically significant).

Correlation coefficient R to age = -0.457. In ANOVA, p-value for age group = 0.0095 (statistically significant); p-value for gender = 0.2017 (not statistically significant). .... 33 Fig. 16. Number density of dermal papillae versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.9503 (not statistically significant); P-value for gender = 0.4749 (not statistically significant). In ANOVA, p-value for age group = 0.8096 (not statistically significant);

p-value for gender = 0.3678 (not statistically significant). NS: not significant. ... 34 Fig. 17. Dermal papillae volume per unit area versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.1471 (not statistically significant); P-value for gender = 0.1287 (not statistically significant). In ANOVA, p-value for age group = 0.4008 (not statistically significant);

p-value for gender = 0.1561 (not statistically significant). NS: not significant. ... 35 Fig. 18. Average volume per dermal papilla versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.5962 (not statistically significant); P-value for gender = 0.1276 (not statistically significant). In ANOVA, p-value for age group = 0.3120 (not statistically significant);

p-value for gender = 0.0920 (not statistically significant). NS: not significant. ... 37 Fig. 19. Dermal papillae volume ratio within dermal papilla zone versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0143 (statistically significant); P-value for gender = 0.3796 (not statistically significant). Correlation coefficient R to age = 0.352. In ANOVA, p-value for age group = 0.0423 (statistically significant); p-value for gender = 0.2307 (not statistically significant). ... 38

(12)

subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 2.035×10^-5 (statistically significant); P-value for gender = 0.6453 (not statistically significant). Correlation coefficient R to age = -0.578. In ANOVA, p-value for age group

= 0.0013 (statistically significant); p-value for gender = 0.9041 (not statistically significant). ... 39 Fig. 21. Four sets of representative images from the ventral forearms of (a) 27-, (b) 24-, (c) 74-, and (d) 69-year-old subjects. The depths of images were labeled, and the depths of the first frames that showed the stratum corneum in respective stacks were set as 0 ȝm.

In (a) young and (c) elder skin, the cross-sections of isolated dermal papillae seemed to be larger and rounder; however, in other (b) young and (d) elder skin, the cross-sections of isolated dermal papillae appeared to be smaller and fragmentary. This difference between subjects did affect the number density of dermal papillae but was not related to age. ... 46 Fig. 22. Coefficient of variance of isolated dermal papilla height (left) versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0144 (statistically significant); P-value for gender = 0.3845 (not statistically significant). Correlation coefficient R to age = -0.352. Coefficient of variance of isolated dermal papilla volume (right) versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age

= 0.0376 (statistically significant); P-value for gender = 0.1055 (not statistically significant). Correlation coefficient R to age = -0295. ... 48 Fig. 23. Average section area of isolated dermal papilla versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.6223 (not statistically significant); P-value for gender = 0.1576 (not statistically significant). ... 49 Fig. 24. Diagrams showed how isolated dermal papillae changed with aging. The isolated dermal papillae were shorter and smaller in the elder skin (b), and the sizes became similar. However, the isolated dermal papillae did not extend their widths in the elder skin.

... 50 Fig. 25. Diagrams that showed how dermal papillae within dermal papilla zone changed with aging. In the elder skin, with the thinner dermal papilla zone, the dermal papillae within dermal papilla zone became shorter but extended in width and occupied more proportion of the volume of dermal papilla zone. ... 54

(13)

Fig. 26. 3D reconstructions of 9 XYZ image stacks from the ventral forearms of (a) 24-, (b) 27-, (c) 28-, (d) 47-, (e) 47-, (f) 57-, (g) 62-, (h) 69-, and (i) 74-year-old subjects. Only SHG signals within the dermal papilla zone were used in the 3D reconstructions, which revealed the structure of dermal papillae. In (d), (e), and (g), the step size between images was 2 ȝm, and the X-Y-Z dimensions of the 3D reconstructions were 240 ȝm × 240 ȝm × 84 ȝm. In the others, the step size was 5 ȝm, and the X-Y-Z dimensions were 240 ȝm × 240 ȝm × 85 ȝm. The coordinate axes were added beside the reconstructions, and z-axes pointed to the deeper dermis. The 3D reconstructions were built by ImageJ 3D Viewer. ... 55

(14)

Table Contents

Table 1. Summary of parameter definitions and notations in the in vivo analysis of dermal papillae within dermal papilla zone ... 18 Table 2. The summery of changes with increasing age of parameters in the analysis of dermal papillae within dermal papilla zone ... 50 Table 3. Depth of DPZ, DP volume per unit area, DP volume ratio within DPZ, and 3D interdigitation index for each image stack in Fig. 26 ... 55

(15)

ϭ

Chapter 1 Introduction

1.1 Skin Aging

Aging is a process of structural integrity loss and physiological changes caused by

both intrinsic and extrinsic factors. Intrinsic aging of skin is a natural process genetically

determined and inalterable. However, extrinsic aging of skin is affected by relatively

controllable factors, and the effects of sunlight exposure are estimated to account for up to

90% of the visible skin aging [1]. Morphologic changes related to intrinsic aging in older

skin are relatively subtle and consist of primary laxity, fine wrinkling, and a variety of

benign neoplasms [2]. The aging-related changes include fewer basal keratinocytes,

decreased vascularity, fewer sweat glands, and flattened dermal-epidermal junction [2].

These changes increase the susceptibility of skin and reduced the skin barrier function,

which increases incidence of inflammatory or infectious skin disorders in the elderly

people [2]. In fact, most people in America over 65 have at least one skin disorder, and

many have two or more [3]. As the proportion of people who are elderly is increasing, the

issue of skin aging is becoming more and more important.

(16)

Ϯ

1.2 Harmonic Generation Microscopy

Although skin biopsy is the most generally used technique to evaluate morphologic

changes in dermatological study, its invasive nature makes it not a suitable method to

investigate skin aging. The harmonic generation microscopy could be a more adaptive

technique for revealing the changes of intrinsic skin aging due to its capability of

non-invasive in vivo imaging with a high spatial resolution. The harmonic generations are

nonlinear optical processes, in which virtual-level transitions are involved. Through the

induction of electric polarization, the second-harmonic generation (SHG) process is the

generation of light with the frequency that is twice the frequency of the excitation

(fundamental) light, and the third-harmonic generation (THG) is the generation of light

with the frequency that is triple the frequency of the excitation light [4]. Due to only the

virtual-level transition involved, the SHG and THG processes leave no energy and cause

no photodamage [5, 6]. Moreover, the intensities of the SHG and THG generated were

proportional to square and cubic of the excitation light intensity respectively and both

signals are generated only in close proximity to the focal point. Therefore, the confined

excitation volume can provide high resolution of three-dimensional image [7, 8]. In

theory, SHG only occurs from optically non-centrosymmetric media, such as collagen

fibers and muscle fibers [9-11]. Therefore, SHG microscopy is ideal for investigate

(17)

ϯ

collagenous structures in the dermis. THG is a process that is dipole allowed and occurs

in all materials. Moreover, THG occurs at interfaces of any media and is free from the

constraint of a phase-matching condition and wavelength restriction [12]. Therefore,

THG can provide the information of bio-tissues and cellular morphology. In the previous

studies, THG was reported to arise from the cell membrane [13] and the cytoplasmic

organelles [8], and to be enhanced by melanin [14], oxy-hemoglobin [15], and elastin [16]

through resonance enhancement [17]. In our previous studies, a femtosecond Cr:forsterite

laser at 1230 nm was used as the excitation source of the harmonic generation

microscopy combining both SHG and THG, which was used for the non-invasive in vivo

optical virtual biopsy on human skin without damage. This harmonic generation biopsy

could clearly reveal in vivo skin structure from the outermost stratum corneum to the

upper dermis [16, 18]. In this study, the harmonic generation biopsy system was applied

to investigate structural changes of epidermis and dermis related to intrinsic aging.

(18)

ϰ

1.3 Comparison between SHGM and SHOCT

Second harmonic optical coherence tomography (SHOCT) is an extension of the

OCT that can provide additional information by collecting SHG signals for imaging

[19-22]. Compared with second harmonic generation microscopy (SHGM), SHOCT also

collects SHG signals from non-centrosymmetric media and possesses the advantages of

harmonic generation, such as no photodamage and confined excitation volume. For the

system setup, because SHOCT needs to split the source light first to generate a reference

second-harmonic light by a nonlinear crystal and then recombine it with the

second-harmonic signal generated from sample, SHOCT requires a more complicated

system setup. For the spatial resolutions, SHOCT could achieve axial resolution of

several micrometers determined by the coherence length and lateral resolution of few

micrometers determined by the spot size, such as 4.2 ȝm and 1.9 ȝm reported by Jiang et

al. [22]. However, using high NA objective, SHGM could achieve axial and lateral

resolution of ~1 ȝm and sub-micrometer [14, 18]. To improve the lateral resolution of

SHOCT, adopting a higher NA objective could narrow the spot size but results in a

smaller depth of focus [23]. To improve the axial resolution of SHOCT, the coherence

length should be reduced, and the coherence length lc was proportional to square of the

(19)

ϱ

center wavelength Ȝ0 and the inverse of bandwidth ǻȜ (_ୡ ൌ ^{ଶ ୪୬ ଶఒ}_గοఒ^బ^మ) [23]. In order to

prevent large scattering and absorption of the human skin, the center wavelength is

preferred in some ranges and could not be too small [23]. Therefore, wide bandwidth is

needed to acquire higher resolution. For example, to achieve axial resolution of ~1 ȝm,

bandwidth should be over 70 nm at center wavelength of 400 nm [24]. For the

penetration depth, OCT was generally reported to achieve penetration of few millimeters

[23]. For SHOCT, the penetration depth was reported to be 280-350 ȝm in hydrated

type I collagen [19] and ~700 ȝm in fish scales. For SHGM, the penetration depth was

reported to be ~1.5 mm through zebra fish embryo [25], ~700 ȝm inside the mouse eye

[26], and above 300 ȝm in the human skin [14, 18]. Therefore, SHOCT and SHGM seem

to provide similar depth of penetration in bio-tissues.

(20)

ϲ

Chapter 2 Materials and Methods

2.1 Study Population

48 Asian subjects with 7 females and 8 males aged 19-29 years, 13 females and 6

males aged 30-59 years, and 8 females and 6 males aged 60-79 years, of Fitzpatrick skin

phototype III or IV participated. One male subject was not included in the analysis of

viable epidermis thickness because his images of viable epidermis were not fully

obtained. Subjects with diabetes or skin diseases were excluded. This study was

conducted according to the Declaration of Helsinki Principles, and the protocol was

approved by the Institutional Review Board of National Taiwan University Hospital.

Informed consent was obtained from each subject prior to study entry.

2.2 Harmonic Generation Biopsy System

A 1230 nm femtosecond Cr:forsterite laser was used as the excitation to reduce the

attenuation from both scattering and absorption of the human skin and to increase the

imaging penetrability [27-29]. As described in the previous study [14], the harmonic

generation biopsy (HGB) system was adapted from a commercial confocal scanning

system (FV300, Olympus, Tokyo, Japan) combined with an inverted microscope (IX71,

(21)

ϳ

Olympus, Tokyo, Japan). A 60X water-immersion objective with NA 1.2 was used in this

study, and second- and third-harmonic generation signals were backward-collected by

two photomultiplier tubes (Fig. 1). Using this system, subjects only need to put their

forearms on the sample stage with ventral side downward, and the imaging depth in the

skin could be control by the z-motor with high resolution (~0.1 ȝm). A submicron lateral

resolution and ~1 ȝm axial resolution were achieved. The penetrability > 300ȝm could

be reached [14, 18].

Fig. 1. A schema of the 1230 nm-based harmonic generation biopsy system; PMT: photomultiplier tube;

DBS: dichroic beam splitter.

ϭϮϯϬŶŵġ

ƌ͗ĨŽƌƐƚĞƌŝƚĞ>ĂƐĞƌġ

ŽŶĨŽĐĂů^ĐĂŶŶŝŶŐ^ǇƐƚĞŵġ

&sϯϬϬġ

ġġ ġ

ġ ġ ġ

ġ WDdġ

^,'ġ ϲϭϱŶŵġ

WDdġ d,'ġ ϰϭϬŶŵġ

ġ ġ

ġ

&ŽƌĞĂƌŵġ

ŽŵƉƵƚĞƌġ

^ġ

KďũĞĐƚŝǀĞġ

ŽůŽƌ&ŝůƚĞƌġ

ĂŶĚͲWĂƐƐ&ŝůƚĞƌġ

ĞĂŵŽůůŝŵĂƚŽƌġ

(22)

ϴ

Fig. 2. A Part of a representative XYZ image stack combing SHG and THG signals from a 24-year-old female subject. The image stack recorded different skin layers including stratum corneum (SC), stratum granulosum (SG), stratum spinosum (SS), stratum basale (SB), papillary dermis (PD), and upper reticular dermis (RD). The depths of images were labeled, and the depth of the first frame of the stack that showed the stratum corneum was set as 0 ȝm. SHG and THG signals were represented by pseudo-colors green and purple. Image dimensions were 240 ȝm × 240 ȝm.

;ĂͿϭϬʅŵ ;ďͿϭϱʅŵ ;ĐͿϮϬʅŵ

;ĨͿϯϱʅŵ

;ĚͿϮϱʅŵ

;ĞͿϯϬʅŵ ;ŐͿϰϬʅŵ ;ŚͿϰϱʅŵ

;ŝͿϱϬʅŵ ;ũͿϱϱʅŵ ;ŬͿϲϬʅŵ

;ƉͿϴϱʅŵ

;ůͿϲϱʅŵ

;ƋͿϵϬʅŵ

;ŵͿϳϬʅŵ ;ŶͿϳϱʅŵ

;ƌͿϵϱʅŵ ;ƐͿϭϬϬʅŵ ;ƚͿϭϬϱʅŵ

;ŽͿϴϬʅŵ

^ ^' ^^

^^

^ ^ ^

W

Z

W

(23)

ϵ

2.3 Imaging Procedure

Observations were performed at the left ventral forearms about 15 cm proximal of

the subjects’ wrists. Using the XYZ mode of the in vivo HGB system, the objective

could move in the z direction with a specified step size, and image stacks of

equally-spaced parallel horizontal sections were acquired with step sizes 2 ȝm or 5 ȝm.

The XYZ image stacks of skin were obtained beginning from the stratum corneum,

through the stratum granulosum, stratum spinosum, stratum basale, papillary dermis, and

to the upper reticular dermis (Fig. 2). During the imaging process, the voltages of two

photomultiplier tubes were adjusted continuously to keep SHG and THG signals

visually bright enough but not saturated. Each image frame was with 512 × 512 pixels,

and the actual spatial dimensions were 240 ȝm × 240 ȝm or 200 ȝm × 200 ȝm. In this

study, the image stacks were excluded if the section images in them were not horizontal

to the skin, which obviously changed with depth non-uniformly in different corners.

Image stacks were also excluded if they had more than two successive blurred frames.

Total 200 XYZ image stacks including 14 stacks with step size 2 ȝm and 186 stacks

with step size 5 ȝm were analyzed. The maximum, minimum, and average numbers of

the image stacks obtained from one subject were 2, 10, and 4.167. The imaging trials of

13 subjects were carried out by me.

(24)

ϭϬ

For each subject, the total laser exposure time in the same area was limited to 30

minutes, and the average power measured right after objective was ~100 mW. Before

and after the HGB, the test sites of all volunteers were recorded by photographing. The

volunteers were informed to feel free to contact us if they felt any discomfort during or

after HGB. There were no cutaneous side effects observed such as erythema, pigment

alteration, ulceration or blister formation. The procedure was comfortable and caused no

itch or pain according to volunteers’ opinions.

2.4 Analysis Protocols

2.4.1 Height, Surface Area, and Volume Estimation

To understand the morphological features of viable epidermis and the dermal

papillae (DP), the analysis was performed on the image stacks of equally-spaced

parallel horizontal sections. From these image stacks, the height, the surface area, and

the volume were measured for analysis. Because the objective was water-immersed, of

which the refraction index was close to that of skin, the refraction of light was neglected.

Therefore, the distance between section images in the skin was approximated by the

step size of the objective movement. The height was calculated by counting the number

(25)

ϭϭ

of the consecutive frames and multiplying by the step size. For the surface area and the

volume, the section areas and their circumferences were measured in consecutive

frames, and the volume and the surface area were estimated by the rectangular

estimation, which approximated the real object by a pile of irregular cylinders (Fig. 3).

For the volume estimation, calculation was done by summing up the section areas

multiplied by the step size. With the same principle, the surface area was calculated by

adding all the products of the circumference and the step size, the area differences

between the pairs of section areas in two successive frames, and the top section area.

Fig. 3. A chart showed the rectangular estimation, which approximated an object by a pile of cylinders.

The volume and surface area of the object could be estimated by measuring the section areas, their circumferences, and step sizes between sections.

For the estimations of the volume and the surface area, which was more proper to

be called as the interface area, of dermal papillae, it should be noted that the tiny spaces

between collagen fibers and the capillaries inside the dermal papillae, which were

(26)

ϭϮ

observed as dark holes surrounded by the SHG signals, were all recognized as parts of

dermal papillae. In other words, if these tiny spaces and dark holes were encircled by

the SHG signals, they would be included in the measurements of the section areas and

their circumferences. Using the estimated height, the interface area, and the volume,

several parameters were defined: the thickness of viable epidermis, the height of

isolated dermal papilla, the volume of isolated dermal papilla, the depth of dermal

papilla zone, the number density of dermal papilla, the dermal papillae volume per unit

area, the average volume per dermal papilla, the dermal papillae volume ratio within

dermal papilla zone, and the 3D interdigitation index.

2.4.2 In Vivo Analysis of Viable Epidermis Thickness

Beneath stratum corneum lay the viable epidermis, which included stratum

granulosum, stratum spinosum, and stratum basale, where keratinocytes with nuclei

could be observed. Such as in Fig. 2(a), no cells with nuclei were observed in the stratum

corneum, and then cells with nuclei first appeared in the stratum granulosum in Fig. 2(b),

where was defined as the topmost layer of the viable epidermis. The thickness of viable

epidermis was measured from stratum granulosum to stratum basale. However, the

interface between epidermis and dermis was an undulated surface, and whether rete ridge,

(27)

ϭϯ

which were the protrusions of epidermis into dermis, should be included in the thickness

measurement was an issue. Therefore, measurements including and excluding the depth

of rete ridge were both performed. Therefore, such as in Fig. 2, the thickness excluding

rete ridge was the step size multiplied by the count of frames from (b) to (e), and the

thickness including rete ridge was the step size multiplied by the count of frames from (b)

to (q). Both thicknesses were measured for every image stack, and average and standard

deviation (SD) of a subject were calculated based on the respective thicknesses of all his

or her image stacks.

2.4.3 In Vivo Analysis of Isolated Dermal Papilla

2.4.3.1 Isolated Dermal Papilla

The dermal papillae were the individual protrusions of dermis to epidermis, and as

the observation got deeper into the skin, they joined together to form the papillary

dermis. In our XYZ image stacks, after the basal cells appeared in the upper layer, the

cross-sections of the upper-part dermal papillae were observed as independent round

areas (Fig. 4). As it got deeper, these independent round areas gradually became larger,

then connected with others, and finally joined together. These independent protrusions

of upper-part dermal papillae were called isolated dermal papillae and were focused in

(28)

ϭϰ

this part of analysis (Fig. 5(a)). In the XYZ image stacks, these independent round areas

were identified as the sections of the isolated dermal papillae before they connected

with others. If a cross-section of an isolated dermal papilla was cut by the border of

image, it would be excluded in the measurements of height and volume. These

independent cross-section areas of isolated dermal papillae in the frame where they first

appeared were recorded and assigned numbers respectively, and the cross-section areas

from the same isolated dermal papillae in the following frames, which were identified

by their locations, were assigned the same numbers and recorded. The measurement and

recording ended when the cross-sections connected to each other. The data of an

isolated dermal papilla with the same assigned number were used to calculate its height

and volume.

Fig. 4. Four consecutive frames of an image stack from the ventral forearm of a 24-year-old female subject. The cross-sections of isolated dermal papillae were observed as independent round areas (arrows), and connected with each other in the following frame (arrow heads). Image dimensions were 240 ȝm × 240 ȝm. The distances between images were 5 ȝm. SHG and THG signals were represented by pseudo-colors green and purple.

(29)

ϭϱ

Fig. 5. Diagrams of (a) isolated dermal papillae and (b) dermal papillae within dermal papilla zone (DPZ), which were denoted by the green areas. Dermal papilla zone was the region between two dotted lines in (b).

2.4.3.2 Height of Isolated Dermal Papilla

The respective heights of isolated dermal papillae observed in all image stacks from

a subject were measured. For each subject, at least 5 isolated dermal papillae were

included for height measurement, and the average, the intra-subject standard deviation,

and the maximum (top 20%) of isolated dermal papilla height were calculated based on

the results of all his or her isolated dermal papillae. It should be clarified that the

average, the intra-subject SD, and the maximum were not acquired for each stack and

then averaged for a subject, but directly calculated from the respective heights of all his

or her isolated dermal papillae.

ĞƌŵŝƐġ

ƉŝĚĞƌŵŝƐġ

;ĂͿ/ƐŽůĂƚĞĚĞƌŵĂůWĂƉŝůůĂĞġ ;ďͿĞƌŵĂůWĂƉŝůůĂĞǁŝƚŚŝŶWġ

ƉŝĚĞƌŵŝƐġ

ĞƌŵĂůWĂƉŝůůĂŽŶĞġ

(30)

ϭϲ

2.4.3.3 Volume of Isolated Dermal Papilla

The volume of each isolated dermal papilla in each image stack from a subject was

calculated by the rectangle estimation. For each subject, the average, the intra-subject

standard deviation, and the maximum of isolated dermal papilla volume were computed

as well based on all his or her isolated dermal papillae, and the number of isolated

dermal papillae should be no less than 5. It should be also clarified that that the average,

the intra-subject SD, and the maximum were directly calculated from the respective

volumes of all his or her isolated dermal papillae.

2.4.4 In Vivo Analysis of Dermal Papillae within Dermal Papilla Zone

2.4.4.1 Dermal Papilla Zone

In the human skin, epidermis and dermis interdigitated at the interface,

dermal-epidermal junction, and the dermal papilla zone was defined as the layer where

epidermis and dermis both existed. Therefore, in the XYZ image stacks, the dermal

papilla zone started when the dermis appeared in the shallower layer, and ended when

the epidermis disappeared (Fig. 5(b)). The features of dermal papillae within the dermal

papilla zone were focused in the following analysis. The total occupied section area and

its circumference of dermal papillae were recorded for each frame within dermal papilla

(31)

ϭϳ

zone in order to calculate following parameters. If part of the circumference overlapped

border lines of the frame, the length of that part was excluded in the measurement. As it

got deeper, dermal papillae would occupy more area in the frame. However, for most of

the XYZ stacks, the occupied section area of dermal papillae did not occupy whole

frame area in the deepest frame within the dermal papilla zone, and sweat glands

appeared occasionally, which were excluded in the measurement. Sweat glands were

observed as large dark holes surrounded by THG signals of cells in dermis. In addition,

the max occupied section area was generally found in the last frame within dermal

papilla zone, but slight decreases of the last frames in some stacks could be found due to

truly decreased occupied area, appearance of sweat glands, or errors of manual

selections. Therefore, the bottom area of the dermal papilla zone was defined as the max

occupied section area of the dermal papillae instead of the whole frame area.

(32)

ϭϴ

Fig. 6. Diagrams showed the required measurements for calculating the parameters defined in the analysis of dermal papillae within DPZ. The total number, the total interface area, the max occupied section area of dermal papillae within DPZ were specified in (a); the depth of DPZ and the total dermal papillae volume within DPZ were specified in (b).

Table 1. Summary of parameter definitions and notations in the in vivo analysis of dermal papillae within dermal papilla zone

Parameter Definition Notation Depth of DPZ

(TDPZ) _ୈ୔୞

Number Density of DP (D_N)

_{୲୭୲ୟ୪}

_୫ୟ୶ DP Volume per

Unit Area (D_V)

_{୲୭୲ୟ୪}

_୫ୟ୶ Average Volume

per DP (VDP)

_{୲୭୲ୟ୪}

_{୲୭୲ୟ୪} DP Volume

Ratio within DPZ (RV)

ൈ

_{୲୭୲ୟ୪}

_୫ୟ୶ൈ _ୈ୔୞

3D Interdigitation

Index (I_3D)

_{୲୭୲ୟ୪}

_୫ୟ୶

;ďͿĞƉƚŚŽĨWĂŶĚġ dŽƚĂůWsŽůƵŵĞǁŝƚŚŝŶWġ

;ĂͿdŽƚĂůEƵŵďĞƌ͕dŽƚĂů/ŶƚĞƌĨĂĐĞƌĞĂ͕

ĂŶĚDĂǆKĐĐƵƉŝĞĚ^ĞĐƚŝŽŶƌĞĂŽĨWġ

ĞƉƚŚŽĨW;dWͿġ dŽƚĂůW

sŽůƵŵĞ

;sƚŽƚĂůͿġ

DĂǆKĐĐƵƉŝĞĚ^ĞĐƚŝŽŶƌĞĂ;KŵĂǆͿġ dŽƚĂů/ŶƚĞƌĨĂĐĞƌĞĂŽĨW

;/ƚŽƚĂůͿġ

ϭġ

Ϭ͘ϱġ

Ϭ͘ϱġ dŽƚĂůEƵŵďĞƌ

ŽĨW;EƚŽƚĂůͿġ

(33)

ϭϵ

2.4.4.2 Depth of Dermal Papilla Zone

The depth of dermal papilla zone (TDPZ; Fig. 6(b)) was measured by counting the

frames from the first observation of dermal papillae to the last observation of basal cells

and multiplying by the step size for each image stack. For a subject, the respective depth

of dermal papilla zone for each of his or her stack was obtained. Based on these depths,

the depth of a subject was acquired from the average of stacks, and the intra-subject SD

was calculated. The change of the depth of dermal papilla zone could indicate whether

the dermal-epidermal junction flattened or sharpened.

2.4.4.3 Number Density of Dermal Papillae

To understand the distribution of dermal papillae, the number of dermal papillae

was acquired by simply counting the number of isolated dermal papillae. In each XYZ

image stack, the total number of dermal papillae (Fig. 6(a)) was counted, among which

an incomplete one was counted as 0.5, and the max occupied section area of dermal

papillae within dermal papilla zone was measured (Fig. 6(a)). The number density of

dermal papillae (DN) for a subject was calculated by the sum of the total number of

dermal papillae (Ntotal) obtained from all his or her stacks divided by the sum of the max

occupied section area (OAmax). For a subject, the respective number density of every his

(34)

ϮϬ

or her stack was also calculated, and the intra-subject standard deviation was obtained

between these number densities.

2.4.4.4 Dermal Papillae Volume per Unit Area

The dermal papillae volume per unit area (D_V) of a subject, which revealed how

much volume of dermal papillae spread on the skin, was calculated by the sum of the

total dermal papillae volume (V_total; Fig. 6(b)) from all his or her image stacks divided

by the sum of the max occupied section area of dermal papillae within dermal papilla

zone (OA_max; Fig. 6(a)). The volume per unit area of each stack was calculated, and the

intra-subject standard deviation for a subject was obtained between the respective

results of his or her stacks.

2.4.4.5 Average Volume per Dermal Papilla

In the analysis of isolated dermal papilla, only the upper-part dermal papillae that

did not connect to each other were analyzed. In this part of analysis, it was defined that

the bottom of each dermal papilla was as deep as the bottom of dermal papilla zone, so

the connected-part of dermal papillae within dermal papilla zone was included for

acquiring the volume of one dermal papilla. Average volume per dermal papilla (V_DP) of

a subject was estimated by the sum of the total dermal papillae volume within dermal

(35)

Ϯϭ

papilla zone (V_total; Fig. 6(b)) measured from all his or her image stacks divided by the

sum of the total number of dermal papillae (N_total; Fig. 6(a)). The average volume of

each stack was calculated, and the intra-subject standard deviation for a subject was

obtained between the respective average volumes of his or her stacks.

2.4.4.6 Dermal Papillae Volume Ratio within Dermal Papilla Zone

To indicate how much volume of dermal papilla zone was occupied by dermal

papillae, the dermal papillae volume ratio within dermal papilla zone (R_V) of a subject

was calculated from the sum of the total dermal papillae volume within dermal papilla

zone (V_total; Fig. 6(b)) in all his or her stacks divided by the sum of the volume of

dermal papilla zone, which was the max occupied section area of dermal papillae

(OA_max; Fig. 6(a)) multiplied by the depth of dermal papilla zone (T_DPZ; Fig. 6(b)). The

ratio of each stack was calculated, and the intra-subject standard deviation of a subject

was obtained between the respective ratios of his or her stacks.

2.4.4.7 3D Interdigitation Index

Timár et al. [30] introduced the idea of the interdigitation index to measure the

undulation of the dermal-epidermal junction. Vertical section of the skin was used, and

the interdigitation index was calculated by the length of the curved line along the

(36)

ϮϮ

dermal-epidermal junction divided by the length of the straight line between two end

points. By expanding their idea, 3D interdigitation index (I_3D) was defined to evaluate

the three-dimensional undulation, which was calculated by the sum of the total interface

area of dermal papillae (IA_total; Fig. 6(a)) obtained from all his or her stacks divided by

the sum of the max occupied section area of dermal papillae within the dermal papilla

zone (OA_max; Fig. 6(b)) for a subject. The index of each stack was calculated, and the

intra-subject standard deviation of a subject was obtained between the respective

indexes of his or her stacks.

2.5 Statistics Protocol

Statistical analysis was performed with IBM SPSS Statistics ver. 20. A set of

parameters and their intra-subject standard deviations was acquired for each subject by

the foregoing methods. To investigate age-related changes, the multiple linear

regression model with age and gender as variables was used for statistical analysis, and

P-values (denoted as P) were acquired for both variables. P-value < 0.05 was considered

to be statistically significant. If age and gender both showed statistical significances,

correlation coefficient (denoted as R) was acquired from the multiple linear regression

(37)

Ϯϯ

model. If only one variable showed statistical significance, correlation coefficient was

acquired from simple linear regression model with only one variable that showed

significance. If age and gender showed no statistical significances for a parameter, from

all the subjects, the average, the inter-subject standard deviation, and the coefficient of

variance (CV), which was the average divided by the standard deviation, were

calculated to observe the dispersion of the parameter. To explore the differences among

age groups, data were sort by age into groups of 19-29 years, 30-59 years, and 60-79

years. The analysis of variance (ANOVA) was performed with two factors, age group

and gender, and statistical significances of both factors (denoted as p) were acquired. If

ANOVA showed significance for either factor, post hoc analysis was performed by LSD

method to inspect difference of which two groups was significance. If gender showed

statistical significance in the multiple linear regression analysis or ANOVA, three age

groups were further divided by gender into six groups. The average and standard

deviation for each group were calculated and displayed as the average ± SD.

(38)

Ϯϰ

Chapter 3 Results

3.1 In Vivo Image Stacks of Human Skin

Using the harmonic generation biopsy system, a series of human skin images were

acquired in vivo starting from the stratum corneum to the upper reticular dermis. In the

image (a) to the image (e) of Fig. 2, all the layers of epidermis were clearly displayed by

the third harmonic generation signals. In the image (f), the second harmonic generation

signals showed the appearance of the papillary dermis distinctly. In the following

images, the epidermis seemed to diminish and the dermis increased their proportions of

image areas. Because the epidermis and dermis were showed by the THG and SHG

signals respectively, the interface of epidermis and dermis was clearly displayed.

Therefore, the structure of dermal papillae could be distinguished from the epidermis by

the SHG signals with fine resolution and surrounding THG signals.

3.2 In Vivo Analysis of Viable Epidermis Thickness

Although statistical significance of age group was found in ANOVA of the

thickness of viable epidermis excluding the depth of rete ridge (Fig. 7; p for age group =

(39)

Ϯϱ

0.0107 and p for gender = 0.8920), there was no significance of age nor gender found in

the regression analysis of the thickness of viable epidermis excluding the depth of rete

ridge (Fig. 7; P_excludeRR for age = 0.3614 and P_excludeRR for gender = 0.5475). The

thicknesses excluding rete ridge for age groups 19-29, 30-59, and 60-79 years were

22.76 ± 2.73 ȝm, 27.50 ± 4.19 ȝm, and 24.76 ± 5.14 ȝm. No trend of decrease or

increase with age could be observed in Fig. 7. Therefore, the thickness of viable

epidermis excluding the depth of rete ridge did not increase or decrease with age. The

average and inter-subject standard deviation of the thickness excluding rete ridge from

all the subjects was 25.23 ȝm and 4.49 ȝm, and the coefficient of variance was 0.1781.

Statistical significance was not found in ANOVA of viable epidermis thickness

including the depth of rete ridge (Fig. 8; p for age group = 0.1592 and p for gender =

0.4282). However, the result showed significance of age in the regression analysis of

viable epidermis thickness including the depth of rete ridge (Fig. 8; P_includeRR for age =

0.0487 and P_includeRR for gender = 0.3396). A small and negative correlation was found

between age and the thickness of viable epidermis including the depth of rete ridge (R =

-0.293). The thicknesses including rete ridge for age groups 19-29, 30-59, and 60-79

years were 91.91 ± 17.56 ȝm, 83.40 ± 15.97 ȝm, and 79.10 ± 17.71 ȝm. The thicknesses

including rete ridge seemed to decrease with age, but differences between subjects were

(40)

Ϯϲ

larger than the age-related decreases.

Both results together implied that not the thickness of viable epidermis but the depth

of rete ridge decreased with age. This implication would be confirmed later in the analysis

of the depth of dermal papilla zone, which was the same as the depth of rete ridge.

Fig. 7. Average thickness of viable epidermis (excluding rete ridge) versus age from the ventral forearms of 47 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.3614 (not statistically significant); P-value for gender = 0.5475 (not statistically significant). In ANOVA, p-value for age group = 0.0107 (statistically significant); p-value for gender = 0.8920 (not statistically significant).

Fig. 8. Average thickness of viable epidermis (including rete ridge) versus age from the ventral forearms of 47 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0487 (statistically significant); P-value for gender = 0.3396 (not statistically significant). Correlation coefficient R to age = -0.293. In ANOVA, p-value for age group = 0.1592 (not statistically significant); p-value for gender = 0.4282 (not statistically significant). NS: not significant.

(41)

Ϯϳ

3.3 In Vivo Analysis of Isolated Dermal Papilla

3.3.1 Height of Isolated Dermal Papilla

ANOVA showed statistical significances of age group for the three height

parameters, average, intra-subject standard deviation, and maximum of height,

respectively (Fig. 9-11; for age group, p_ave= 0.0003, p_SD = 0.0051, and p_max = 0.0021),

but no significances of gender were found (Fig. 9-11; for gender, p_ave = 0.0975, p_SD =

0.7703, and p_max = 0.5943). Regression analysis also showed significances between age

and the three height parameters, average, intra-subject standard deviation, and maximum

of height (Fig. 9-11; for age, P_ave = 0.0003, P_SD = 8.315ǘ10^-5, and P_max = 5.077ǘ10^-5),

respectively, but gender only showed significances in the regression analysis of the

average height (for gender, P_ave = 0.0285, P_SD = 0.4806, and P_max = 0.3165). The

intra-subject SD and the maximum of isolated dermal papilla height were negatively and

moderately correlated only to age (R_SD = -0.540; R_max = -0.551). The average of isolated

dermal papilla height was negatively and moderately correlated to both age and gender

(R_ave = -0.563). The overall average for male subjects of the average height (18.91 ȝm)

was higher than that for female subjects (15.54 ȝm). The average heights for age groups

19-29, 30-59, and 60-79 years were 21.90 ± 5.98 ȝm, 15.01 ± 4.85 ȝm, and 14.25 ± 4.11

(42)

Ϯϴ

ȝm. For respective age groups of females and males, the average heights for female age

groups 19-29, 30-59, and 60-79 years were 19.02 ± 4.77 ȝm, 14.97 ± 4.71 ȝm, and

13.40 ± 3.88 ȝm, and the average heights for male age groups 19-29, 30-59, and 60-79

years were 24.41 ± 6.06 ȝm, 15.10 ± 5.61 ȝm, and 15.38 ± 4.47 ȝm. The intra-subject

SDs of height for age groups 19-29, 30-59, and 60-79 years were 11.76 ± 4.37 ȝm, 8.10

± 5.14 ȝm, and 6.21 ± 2.69 ȝm. The maximum heights for subjects aged 19-29, 30-59,

and 60-79 years were 40.14 ± 11.30 ȝm, 28.69 ± 13.78 ȝm, and 23.98 ± 8.24 ȝm.

Fig. 9. Average of isolated dermal papilla height versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0003 (statistically significant);

P-value for gender = 0.0285 (statistically significant). Correlation coefficient R to age and gender = -0.563.

(43)

Ϯϵ

In ANOVA, p-value for age group = 0.0003 (statistically significant); p-value for gender = 0.0975 (not statistically significant). In two lower figures, data of female and male subjects were separately displayed.

Fig. 10. Intra-subject standard deviation of isolated dermal papilla height versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 8.315×10^-5 (statistically significant); P-value for gender = 0.4806 (not statistically significant). Correlation coefficient R to age = -0.540. In ANOVA, p-value for age group = 0.0051 (statistically significant);

p-value for gender = 0.7703 (not statistically significant).

Fig. 11. Maximum of isolated dermal papilla height versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 5.077×10^-5 (statistically significant); P-value for gender = 0.3165 (not statistically significant). Correlation coefficient R to age = -0.551. In ANOVA, p-value for age group = 0.0021 (statistically significant); p-value for gender = 0.5943 (not statistically significant).

(44)

ϯϬ

3.3.2 Volume of Isolated Dermal Papilla

ANOVA showed statistical significances of age group for the three volume

parameters, average, intra-subject standard deviation, and maximum of volume,

respectively (Fig. 12-14; for age group, pave = 0.0128, pSD = 0.0192, and pmax = 0.0067),

but no significances of gender were found (for gender, pave = 0.1604, pSD = 0.7484, and

pmax = 0.4859). There were statistical significances in the respective regression analyses

of age and the three volume parameters, average, intra-subject standard deviation, and

maximum of volume (Fig. 12-14; for age, Pave = 0.0037, PSD = 0.0006, and Pmax =

0.0005). The average, intra-subject SD, and maximum of isolated dermal papilla volume

were negatively related to age (Rave = -0.405; RSD = -0.482; Rmax = -0.482). In addition,

we found that none of them were correlated to gender with statistical significances (for

gender, Pave = 0.0930, PSD = 0.5912, and Pmax = 0.3244). The average volumes for age

groups 19-29, 30-59, and 60-79 years were 45667 ± 27010 ȝm³, 27147 ± 20758 ȝm³,

and 24291 ± 10794 ȝm³. The intra-subject SDs of volume for age groups 19-29, 30-59,

and 60-79 years were 49776 ± 28917 ȝm³, 33155 ± 30822 ȝm³, and 21219 ± 9538 ȝm³.

The maximum volumes for age groups 19-29, 30-59, and 60-79 years were 135624 ±

80970 ȝm³, 83118 ± 68373 ȝm³, and 58694 ± 23695 ȝm³.

(45)

ϯϭ

Fig. 12. Average of isolated dermal papilla volume versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0037 (statistically significant); P-value for gender = 0.0930 (not statistically significant). Correlation coefficient R to age = -0.405. In ANOVA, p-value for age group = 0.0128 (statistically significant); p-value for gender = 0.1604 (not statistically significant).

Fig. 13. Intra-subject SD of isolated dermal papilla volume versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0006 (statistically significant); P-value for gender = 0.5912 (not statistically significant). Correlation coefficient R to age = -0.482. In ANOVA, p-value for age group = 0.0192 (statistically significant); p-value for gender = 0.7484 (not statistically significant).

(46)

ϯϮ

Fig. 14. Maximum of isolated dermal papilla volume versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0005 (statistically significant); P-value for gender = 0.3244 (not statistically significant). Correlation coefficient R to age = -0.482. In ANOVA, p-value for age group = 0.0067 (statistically significant); p-value for gender = 0.4859 (not statistically significant).

3.4 In Vivo Analysis of Dermal Papillae within Dermal Papilla Zone

3.4.1 Depth of Dermal Papilla Zone

ANOVA of the depth of dermal papilla zone (T_DPZ) only showed the statistical

significance of age group (Fig. 15; p for age group = 0.0095 and p for gender = 0.2017).

The regression analysis was also statistically significant for age but not for gender (Fig.

15; P for age = 0.0009 and P for gender = 0.0989). The correlation was negative (R =

-0.457) for the depth of dermal papilla zone to age. The depths for age groups 19-29,

30-59, and 60-79 years were 74.22 ± 13.25 ȝm, 60.63 ± 16.32 ȝm, and 57.20 ± 14.29

(47)

ϯϯ

ȝm. The result of thinner dermal papilla zone in the aged skin indicated that the

dermal-epidermal junction flattened with aging.

Fig. 15. Depth of dermal papilla zone versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.0009 (statistically significant);

P-value for gender = 0.0989 (not statistically significant). Correlation coefficient R to age = -0.457. In ANOVA, p-value for age group = 0.0095 (statistically significant); p-value for gender = 0.2017 (not statistically significant).

3.4.2 Number Density of Dermal Papillae

No statistical significance of age group nor gender was found in ANOVA of the

number density of dermal papillae (DN; Fig. 16; p for age group = 0.8096 and p for

gender = 0.3678). Similarly, the regression analysis showed no statistical significance

for the number density of dermal papillae to age or gender (Fig. 16; P for age = 0.9503

and P for gender = 0.4749). Dermal papillae were not found to be more or fewer with

increasing age. The average and inter-subject standard deviation of number density from

all the subjects were 148.65 mm^-2 and 52.06 mm^-2, and the coefficient of variance was

(48)

ϯϰ

0.3502. The number densities for age groups 19-29, 30-59, and 60-79 years were 149.88

± 55.07 mm^-2, 149.54 ± 49.58 mm^-2, and 146.14 ± 55.85 mm^-2, which were quite similar

for different age groups. In Fig. 16, no tendency of age-related changes could be

observed, and data were scattered in all age groups.

Fig. 16. Number density of dermal papillae versus age from the ventral forearms of 48 Asian subjects (skin phototype III & IV). In linear regression analysis, P-value for age = 0.9503 (not statistically significant);

P-value for gender = 0.4749 (not statistically significant). In ANOVA, p-value for age group = 0.8096 (not statistically significant); p-value for gender = 0.3678 (not statistically significant). NS: not significant.

3.4.3 Dermal Papillae Volume per Unit Area

ANOVA of dermal papillae volume per unit area (DV) did not show statistical

significances of age group and gender (Fig. 17; p for age group = 0.4008 and p for

gender = 0.1561). There were no statistical significances in the regression analysis of

dermal papillae volume per unit area to age and gender (Fig. 17; P for age = 0.1471 and

P for gender = 0.1287). Therefore, the volume of dermal papillae within dermal papilla

zone per unit area did not increase or decrease with age. The volumes per unit area for