Mapping the dermoscopic features of the normal face: Findings from 200 subjects with skin of colour

Introduction

Dermoscopy is a non-invasive diagnostic tool that allows the visualisation of epidermal and superficial dermal structures. It can be particularly useful for cosmetically sensitive sites such as the face, where more invasive diagnostic procedures, such as skin biopsy, are not preferred. Understanding normal dermoscopic variations is important for better interpretation of abnormal findings. However, there is surprisingly limited information on the dermoscopic features of normal facial skin, other than that a pseudonetwork is present instead of a true pigment network.¹ We conducted this study to describe the dermoscopic features of the normal face and assess the site-, gender-, and age-specific variations.

Methods

This was a prospectively conducted cross-sectional study undertaken from April 2023 to November 2024, after approval from the Institutional Ethics Committee (IECPG-76/07.03.2023, RT-08/23.03.2023).

We recruited a convenience sample of 200 participants, stratified by age and gender, with apparently normal facial skin or dermatology patients with skin lesions at extra-facial sites, after their informed consent. Non-probabilistic quota sampling method was used to recruit 100 male and 100 female subjects from the following age-groups: 0 to 14 (children), 15-34 (young adults), 35-64 (middle-aged), and >65 years of age (elderly). We aimed to include at least 20 subjects in each age group, with equal numbers of male and female subjects. Participants with active or incompletely resolved facial skin disease or those with a history of any topical medication use in the preceding 4 weeks were excluded.

Fitzpatrick skin phototype was determined based on the history of sun burn and tanning responses, as proposed by Eilers et al for skin of colour.² Dermoscopic examination of a priori selected seven facial sites (forehead, bilateral zygomatic areas, bilateral cheeks, nose, and chin) of normal-looking facial skin was independently done by two investigators (SA, VG) using a hand-held dermoscope (Heine delta 30) at 10x magnification [Figure 1]. Examination was done using light contact (without pressure), without interface fluid, and in both polarised and non-polarised modes. If a skin lesion was detected on careful ‘naked eye’ examination at a specific site, that site was excluded, and findings were recorded only from the remaining unaffected sites. The dermoscopic findings were recorded in a pre-designed proforma using the standard International Dermoscopic Society nomenclature.³ Dermoscopic images were taken by a smartphone (OnePlus 10 Pro) by apposing the rear camera to the eyepiece of the dermoscope in the polarised mode. Images were captured in standard settings of contrast and brightness, without a flash, and without zoom/magnification in the smartphone camera. All the proformas and dermoscopic images were reviewed at the end of study, and any discrepancies between the investigators were resolved by discussion and mutual agreement.

Close

Figure 1:

Schematic representation of a priori selected facial sites for dermoscopic examination. 1: Forehead; 2,3: Bilateral zygomatic area; 4,5: Bilateral cheeks; 6: Nose; and 7: Chin. Size of the examined area corresponds to the diameter of the Heine Delta 30 dermoscope lens (32 mm).

Export to PPT

The frequencies of various dermoscopic findings in different sub-groups (sites, gender, and age-groups) were compared using the Chi-square or Fisher Exact test, as applicable. A p-value of <0.05 was considered statistically significant.

Results

The age and gender distribution of study subjects are shown in Figure 2. The Fitzpatrick skin phototypes’ distribution was III (n=16, 8%), IV (n=28, 14%), V (n=102, 51%), and VI (n=54, 27%). A total of 1,320 sites in 200 subjects were examined, which included the forehead (n=199), right and left zygoma (n=395), right and left cheek (n=380), nose (n=181), and chin (n=165).

Close

Figure 2:

Age and gender distribution of study participants.

Export to PPT

Dermoscopic findings of different facial sites

The dermoscopic findings in the 200 subjects, along with site differences, are shown in Table 1.

Table 1: Site-wise variations of dermoscopic findings

Dermoscopic parameter	Facial site					p-value
	Forehead (n=199)	Zygoma (n=395)	Cheek (n=380)	Nose (n=181)	Chin (n=165)
Pseudonetwork and true network*
Pseudoreticular network, uniform	170 (85%)	250 (63%)	237 (62%)	74 (41%)	109 (66%)	<0.001
Pseudo reticular network, patchy	16 (9%)	95 (27%)	92 (27%)	23 (23%)	30 (22%)	<0.001
Pseudo reticular network, accentuated patchy	8 (4%)	62 (18%)	71 (21%)	11 (11%)	5 (4%)	<0.001
Absent pseudoreticular network	13 (7%)	44 (11%)	39 (10%)	82 (45%)	26 (16%)	<0.001
Pseudoreticular network, unfocused	35 (18%)	117 (30%)	140 (37%)	14 (8%)	45 (27%)	<0.001
True network present	92 (46%)	74 (19%)	35 (9%)	17 (9%)	35 (21%)	<0.001
Follicular structures and related findings
Follicular openings, variable size	59 (30%)	160 (40%)	167 (44%)	126 (70%)	83 (50%)	<0.001
Follicular openings, variable distance	39 (20%)	175 (44%)	166 (44%)	96 (53%)	61 (37%)	<0.001
Follicular plugs	5 (3%)	20 (5%)	26 (7%)	58 (32%)	9 (5%)	<0.001
Follicular brown pigmentation	17 (9%)	112 (28%)	139 (37%)	65 (36%)	44 (27%)	<0.001
Brown circles around follicular/eccrine openings	41 (21%)	92 (23%)	99 (26%)	75 (41%)	41 (25%)	0.012
Dot-in-circle	25 (13%)	119 (30%)	94 (25%)	24 (13%)	32 (19%)	<0.001
Perifollicular white colour	0	0	0	5 (3%)	4 (2%)	<0.001
Vessel patterns
Vessels present	17 (9%)	70 (18%)	184 (48%)	9 (5%)	44 (27%)	<0.001
Linear non-branching vessels	14 (7%)	67 (16%)	178 (47%)	8 (5%)	41 (25%)	<0.001
Linear branching vessels	5 (6%)	17 (4%)	80 (21%)	5 (3%)	19 (12%)	<0.001
Linear curved vessels	6 (3%)	29 (7%)	101 (27%)	6 (3%)	27 (16%)	<0.001
Polymorphous vessels	7 (4%)	34 (8%)	127 (33%)	8 (4%)	34 (21%)	<0.001
Other findings
Brown dots	0	3 (0.7%)	2 (0.5%)	1 (0.5%)	1 (0.6%)	0.831
Brown clods	0	7 (1.7%)	5 (1.3%)	4 (2.2%)	0	0.134
White clods	0	1 (0.2%)	5 (1.3%)	4 (2.2%)	4 (2.4%)	0.040
White scales	28 (14%)	80 (20%)	98 (26%)	37 (20%)	47 (28%)	0.004
White reticular lines	0	2 (0.5%)	9 (2.3%)	28 (15%)	2 (1.8%)	<0.001

* More than one pattern of pseudonetwork/true network can be present in a single facial site. Bold font signifies statistical signifance (ie, p<0.05)

Pseudonetwork and true network

The most common finding was a background uniform pseudonetwork at all sites [Figure 3a], but it showed a site-wise variation; being most common on the forehead (n=170, 85%) and least common on the nose (n=74, 41%, p<0.001). The pseudonetwork was present only patchily in 9-27% of all sites. Additionally, an accentuated (broader and darker, sharply defined) patchy pseudonetwork [Figure 3b] was seen, mostly on the cheeks (n=71, 21%) and zygoma (n=62, 18%, p<0.001). A small proportion of all facial sites lacked the pseudonetwork [Figure 3c], most commonly the nose (n=82, 45%, p<0.001).

Close

Figure 3a:

Background uniform pseudonetwork, in-focus, with uniform size and distance of follicle openings (forehead) (Heine delta 30, 10x, polarised).

Export to PPT

Close

Figure 3b:

Patchy accentuation of pseudonetwork (solar lentigo-like, red arrows) and background pseudonetwork (cheek) (Heine delta 30, 10x, polarised).

Export to PPT

Close

Figure 3c:

Absent pseudonetwork (cheek) (Heine delta 30, 10x, polarised).

Export to PPT

The pseudonetwork was in-focus [Figure 3a] in a majority (63-92%) of all sites, but the cheeks showed an unfocussed pseudonetwork (n=140, 37%, p<0.001) [Figure 3d] most frequently, followed by the zygoma (n=117, 30%) and chin (n=45, 27%).

Close

Figure 3d:

Unfocussed pseuodnetwork (cheek) (Heine delta 30, 10x, polarised).

Export to PPT

In addition, a true network [Figure 3e] was also seen patchily, most commonly on the forehead (n=92, 46%, p<0.001).

Close

Figure 3e:

True network (red *) and background pseudonetwork (green *) (forehead) (Heine delta 30, 10x, polarised).

Export to PPT

Follicular structures and related findings

A uniform size of follicular openings and uniform distances between them [Figure 3a] were seen commonly at all sites, except the nose, where the size and distance were more commonly variable (n=126, 70%, p<0.001; n=96, 53%, p<0.001, respectively) [Figure 4a]. Follicular plugs [Figure 4b], an infrequent finding otherwise (3-32% of all sites), were most common on the nose (n=58, 32%, p<0.001). Follicular brown pigmentation [Figure 4c] was most common on the cheeks and nose (n=139, 37% and n=65, 36%, respectively p<0.001). Brown circles around adnexal openings [Figure 4d], seen in 21-41% of all sites, were more common on the nose (n=75, 41%, p=0.012). ‘Dot-in-circle’ morphology [Figure 4e], present in 13-30% of all sites, was most commonly seen on the zygoma (n=119, 30%) and cheeks (n=94, 25%, p<0.001). Trichostasis spinulosa was incidentally detected on the nose of 4 (2%) participants [Figure 4f].

Close

Figure 4a:

Variable size and spacing of follicle openings (nose) (Heine delta 30, 10x, polarised).

Export to PPT

Close

Figure 4b:

Follicular plugs (nose) (Heine delta 30, 10x, polarised).

Export to PPT

Close

Figure 4c:

Follicular brown pigmentation (red arrows) (nose) (Heine delta 30, 10x, polarised).

Export to PPT

Close

Figure 4d:

Brown circles around adnexal openings (cheek).

Export to PPT

Close

Figure 4e:

‘Dot in circle’ sign (zygoma).

Export to PPT

Close

Figure 4f:

Trichostasis spinulosa (red arrows) seen as tufts of vellus hair from a follicular opening (nose) (Heine delta 30, 10x, polarised).

Export to PPT

Vessel patterns

Vessels [Figure 5] were seen variably in 5-48% of all sites, with cheeks (n=184, 48%) and chin (n=44, 27%) being the most common sites, and nose (n=9, 5%) and forehead (n=17, 9%) the least common.

Close

Figure 5:

Unfocussed linear branching vessels on cheek (Heine delta 30, 10x, polarised).

Export to PPT

A linear non-branching pattern of vessels was the most common (n=178, 47%, p<0.001), followed by linear curved (n=101, 27%, p<0.001) and linear branching (n=80, 21%, p<0.001) patterns. Dotted vessels were not seen at any site. Polymorphous vessels were most common on the cheek (n=127, 33%, p<0.001). All sites showed unfocused vessels only.

Other findings

Brown dots and clods were seen sporadically at all sites, and no site-specific variation was observed. Peri-follicular white colour, another uncommon finding, was seen only on the nose (n=5, 3%) and chin (n=4, 2%). White reticular lines [Figure 6] were seen almost exclusively on the nose (n=28, 15%).

Close

Figure 6:

White reticular lines on nose (Heine delta 30, 10x, polarised).

Export to PPT

Discussion

Knowledge of the dermoscopic variations in healthy skin based on site, age, gender, and ethnicity would help dermatologists in distinguishing the abnormal from normal findings, and increase their diagnostic accuracy. However, studies documenting normal dermoscopic features of healthy skin are limited in the literature; most studies have focused on dermoscopic findings in skin diseases. Pathological dermoscopic findings too may differ based on the location of the skin lesions.^4,5 Trichoscopic studies in healthy females and children have been conducted to establish normative data, allowing the formulation of diagnostic criteria for female pattern hair loss.^6–9 Recently, dermoscopic features of oral labial mucosa and nail fold capillaroscopy in healthy individuals have also been described.^10–12

To our knowledge, there are no previous studies exploring the variations in dermoscopic features of normal facial skin. The limited existing information on the dermoscopy of facial skin comes from the examination of facial pigmented skin lesions in fair-skinned patients.¹ Facial skin is typically described as having a pseudonetwork pattern instead of a true pigment network, because of more follicular openings and less well-developed rete ridges. Rarely, fine lines or a true network may also be seen on the face.¹

Our study documents the variations in dermoscopic features of the normal facial skin [Figure 7]. While dermoscopic findings showed a large overlap across facial sites (mainly a uniform background pseudonetwork), some notable site-specific differences were also observed. The cheeks and zygomas were more likely to have a patchy accentuated pseudonetwork and dot-in-circle morphology. Vessels were more common on the cheeks and chin as compared to other sites. The nose emerged as a special site with some unique dermoscopic findings, viz., follicular plugs, follicular brown pigmentation, brown circles around adnexal openings, and white reticular lines, while lacking the background pseudonetwork. There were a few gender and age-related variations as well. An accentuated patchy pseudonetwork, follicular plugs, and brown circles around adnexal openings were more common in males, while follicular brown pigmentation was more common in females. Accentuated patchy pseudonetwork, follicular plugs, follicular brown pigmentation, brown circles around adnexal openings, vessels, and white reticular lines became more common with advancing age.

Close

Figure 7:

Schematic representation of the site-specific dermoscopic variations of the normal face.

Export to PPT

Though a background pseudonetwork was typically present uniformly across all facial sites, it was sometimes patchy or not visible at some sites. Variable and distant spacing of follicular openings, as on the nose, disrupts their patterned arrangement, thus rendering the pseudonetwork indiscernible. Further, the pseudonetwork may be unfocused or even absent because of the non-visualisation of the adnexal openings or poor background contrast, as in participants with lighter skin complexion. This was seen mostly on the zygoma, cheeks, and chin; the reason for this is not clear, but it may be related to lower hair follicle density and thicker skin at these sites.^13,14 Additionally, the pseudonetwork may focally become darker, thicker, and sharper (accentuated), resembling a solar lentigo on the face.^15,16 This finding was seen mostly on the cheeks and zygoma of the middle-aged (34-44%) and elderly (11-44%) subjects, and may indeed represent a small incidental solar lentigo or its precursor not visible to the naked eye. Variations in pseudonetwork, which are often used as clues for facial pigmented skin lesions, should therefore be interpreted in the context of these observations. A uniform accentuation or a greyish pseudonetwork may carry more diagnostic importance than other variations.¹⁶

Apart from pseudonetwork variations, other normal dermoscopic findings on the face included follicular brown pigmentation, brown circles around adnexal openings, dot-in-circle structures, and vessels. Follicular brown pigmentation and brown circles around adnexal openings were seen in up to 37% and 41% of all facial sites respectively, with the cheek and nose being the more common sites. Follicular brown pigmentation was seen uncommonly in children, and brown circles became more common with age. Similar findings have been described in maturational hyperpigmentation¹⁷ and lichen planus pigmentosus¹⁸, and could represent an accentuated normal dermoscopic finding on the faces of individuals with darker skin types. Asymmetric peri-follicular pigmentation and circles, on the other hand, should be considered pathological and are a clue for melanoma.⁴ A dot-in-circle morphology was seen in 13-30% of all sites, with it being most common on the zygoma and cheek. It is seen as a central brown dot (or a small circle) surrounded by a hypopigmented halo. It has been previously reported in 13-16% cases of lichen planus pigmentosus as ‘targetoid’ or ‘owl’s eye’ structures.^18–20 Thought to represent follicular plugging, its exact histogenesis remains unclear. Such structures have also been described in other facial lesions, such as pigmented actinic keratosis and lentigo maligna.¹⁶ In the light of our findings, dot-in-circle structures may be better seen as a facial site-related finding than a diagnostic clue.

Vessels were seen variably at all sites, with cheeks (48%) being the most common sites across the gender and age groups. The lower frequency of vessels on the nose was unexpected but may be attributed to the smaller proportion of elderly participants in our study, as vessels were more common on the nose of older individuals than in other age groups (33% vs. 0–7%). The prevalence of vessels increased with age possibly because of photodamage. Various vessel patterns have been described in several facial dermatoses.^21–23 Notably, we did not see focused or dotted vessels in our study, and therefore they may be considered of clinical relevance when visualised on the face.

Some uncommon findings were site-specific, such as follicular plugs and white reticular lines (also known as chrysalis-like structures) seen on 32% and 15% of all nose sites, respectively. The white reticular lines are thought to represent fibrosis, and have been reported in several neoplastic and inflammatory skin lesions.^24,25 Their prevalence increased with age in our study, and they may represent age-related stiffening of dermal collagen.²⁶ Other uncommon findings included dots and clods (<1-2%), and peri-follicular white halos (2-3%). These findings may not be clinically relevant when seen sporadically, but may have diagnostic value when seen in multiple areas.

Limitations

Our study is limited by a small sample size but still comparable to previous studies describing findings on trichoscopy, mucoscopy, or nail fold capillaroscopy in healthy individuals. Only a small area from pre-selected facial sites was examined, as dermoscopic assessment of the entire facial skin was not practically feasible. The proportion of elderly subjects was lower than desired, as many had some clinically obvious facial lesions, such as sebaceous hyperplasia, seborrheic keratosis, maturational dyspigmentation, telangiectasias, and/or actinic keratoses, and were therefore excluded. Hence, the age-group comparisons in this study should be interpreted with caution. Additionally, most (78%) of our subjects had Fitzpatrick skin phototypes V-VI, and the validity of our observations needs to be tested in a fair-skinned population. Due to the under-representation of lighter-skinned subjects (skin phototypes III 8%, IV 14%), a post-hoc comparative analysis between different skin phototypes could not be carried out.

Conclusion

Our study describes the dermoscopic features of the normal face and documents the site, gender, and age-related differences therein. The pseudonetwork and its variations, unfocussed vessels, follicular brown pigmentation, brown circles around adnexal openings, and dot-in-circle structures were found to be common dermoscopic findings on the normal face. These results should be kept in mind by dermatologists while interpreting the dermoscopic findings in facial skin disease.