Dermoscopic Features of Erosive Pustular Dermatosis of the Scalp: A Comparative Multicentric Retrospective Study in Bald and Hairy Patients

Introduction

Erosive pustular dermatosis of the scalp (EPDS) is an uncommon inflammatory disorder affecting the scalp and typified by patches of hair loss along with erosions, thick yellow or yellow-brown crusts, superficial ulcers, and pustules. It progresses over months or years, eventually resulting in skin atrophy and scarring alopecia. Elderly individuals with a bald, sun-damaged scalp represent the most common involved population, yet instances affecting hairy subjects may also be seen.¹ EPDS can be difficult to diagnose, particularly in its early stages when the clinical presentation overlaps with other scalp disorders, such as scalp infections, bullous diseases, and cicatricial alopecias.¹ In this regard, dermoscopy has been proven to be a valuable diagnostic tool in assisting the recognition of various hair and scalp disorders by revealing specific clues, thus decreasing the number of cases needing biopsy.² However, data on dermoscopic features of EPDS is limited.^3,4

The aim of the present retrospective observational multicentric study was to compare the dermoscopic findings of EPDS in bald and hairy patients with those of the main differential diagnoses in order to improve the non-invasive recognition of such a condition.

Materials and Methods

This was a multicentric observational retrospective analysis involving three dermatological centers from Italy (Udine, Bologna and Trieste). The study population consisted of two cohorts, ie, consecutive bald (I) and hairy (II) patients diagnosed with EPDS according to histological assessment; controls were also considered by selecting subjects (bald and hairy) with hair disorders for which EPDS was included into the clinical differential diagnosis. We included only cases where high-quality dermoscopic images of the affected area were available, specifically the region from which the biopsy sample was taken or the most clinically representative lesion. Images were captured at 10× magnification under polarized light in a dry setting, with the possible additional use of a fluid interface based on the physician’s evaluation.

Histologically confirmed diagnosis of Erosive Pustular Dermatosis of the Scalp (EPDS) was needed for the enrollment. To ensure that treatment-related changes did not influence the findings, they must not have undergone systemic or topical treatment for scalp conditions in the three months preceding the dermoscopic evaluation. Additionally, complete clinical and demographic data, including age, sex, and medical history, had to be accessible. Only patients from one of the three participating dermatology centers—Udine, Bologna, or Trieste—who met all these criteria were included. Patients were excluded from the study if histopathological confirmation of EPDS was unavailable or if the dermoscopic images were of insufficient quality due to low resolution, poor focus, or non-standardized lighting conditions that could interfere with interpretation. Those who had used systemic or topical therapies, such as corticosteroids or immunosuppressants, within three months before the dermoscopic assessment were also excluded, as these treatments could alter the appearance of the skin and hair. Moreover, patients with coexisting scalp diseases, including autoimmune disorders like discoid lupus erythematosus or lichen planopilaris, as well as chronic infections affecting the scalp, were not considered eligible. Lastly, insufficient clinical data or incomplete patient records resulted in exclusion from the study.

Dermoscopic images were randomly evaluated by two independent experienced investigators (E.E. and N.P.), who assessed dermoscopic findings according to the International Dermoscopy Society criteria for non-neoplastic dermatoses⁵ integrated with parameters proposed by Rudnicka et al for follicular findings.⁶ Each feature was recorded and analyzed based on its prevalence in both EPDS and control groups. Comparative analyses were also performed separately for bald and hairy patient groups to assess any differences in dermoscopic features with controls. Interobserver agreement for EPDS cohorts was evaluated through Cohen’s kappa coefficient, whereas Fisher’s exact test with p-value set at 0.01 was used for comparative analyses. All statistical calculations were performed using Microsoft Excel, ensuring systematic data handling and accurate computation of values. Total duration of data collection was 6 months (January-June 2024).

Results

A total of 116 patients [53 bald patients (28 with EPDS and 25 controls) and 63 hairy patients (32 with EPDS and 31 controls)] were included in the study; mean age was 68 and 57 years for bald and hairy EPDS cases and 61 and 59 years for corresponding controls, respectively. In the control groups we included: 16 squamous cell carcinomas, six pemphigus vulgaris and three cicatricial pemphigoid (bald patients); 14 folliculitis decalvans, 11 discoid lupus erythematosus, five lichen planopilaris and two pityriasis amiantacea (hairy patients). For all instances, hand-held dermatoscope Dermlite DL5 (San Juan Capistrano, CA, United States) coupled with a high resolution camera or smartphone was employed.

Considering bald patients (Figure 1 and Table 1), the main dermoscopic features (prevalence greater than one-third) of EPDS turned out to be linear-curved vessels (unspecific distribution) (78.6%), orange structureless areas (diffuse) (46.4%), and yellow scales/crusts (focal) (39.3%); other relevant findings included white structureless areas (focal) (32.1%), non-follicular pustules (28.6%), and dotted, linear and linear with branches vessels (28.6–32.1%). Moving to hairy patients with EPDS (Figure 2 and Table 2), yellow scales/crusts (focal) and horizontally arranged hairs with the proximal part of the shaft being visible through a thinned epidermis at the edge of alopecic areas (“peripheral horizontal hair” sign) were the main dermoscopic features (both with a prevalence of 84.4%). Further common findings included linear vessels and linear vessels with branches (unspecific distribution) (56.3% and 40.6%, respectively), white structureless areas (focal) (53.1%), non-follicular pustules (40.6%), broken hairs (40.6%), and black dots (34.4%).

Table 1 Dermoscopic Findings of Included Instances of Erosive Pustular Dermatosis of the Scalp (EPDS) and Controls in Bald Patients, Along With Results of the Comparative Analysis

Table 2 Dermoscopic Findings of Included Instances of Erosive Pustular Dermatosis of the Scalp (EPDS) and Controls in Hairy Patients, Along With Results of the Comparative Analysis

Figure 1 Clinical and dermoscopic features of erosive pustular dermatosis of the scalp in three bald patients (a, c and e) compared to three controls (g, i and k). (a) Extensive erosions, erythema and yellow-white scales/crusts are visible on clinical assessment; (b) Dermoscopy displays diffuse focused linear-curved (black arrow) and branched vessels. (c) Clinical examination shows erosions with erythema and prominent yellow crusts; (d) In-focus linear and linear-curved vessels (black arrow) surrounding an Orange structureless areas are visible on dermoscopic examination. (e) Clinically, one major erosion with pustules and sero-hematic crusts are evident on the vertex region of the scalp; (f) Dermoscopy shows linear-curved vessels (black arrow) and non-follicular pustules (white arrow). (g) Cicatricial pemphigoid: clinical image; (h) Dermoscopic evaluation shows white scales, sero-crusts and linear/dotted vessels. (i) Squamous cell carcinoma of the vertex: clinical image; (j) Dermoscopy displays dotted and linear-curved vessels along with perivascular white halos, ulcerations and structureless white areas. (k) Pemphigus of the scalp: clinical image; (l) Dermoscopic features include linear erosions (black arrow), white-yellow scales and sparse dotted vessels.

Figure 2 Clinical and dermoscopic features of erosive pustular dermatosis of the scalp in three hairy patients (a, c and e) compared to three controls (g, i and k). (a) Erythema with extensive crusting and erosion with alopecia in affected areas on clinical examination; (b) Dermoscopy displays “peripheral horizontal hair” sign (visible hair shafts through the thinned epidermis) (black arrow), yellow scales/crusts, non-follicular pustules (white arrow) and linear-curved/branched vessels. (c) Clinically, crusting and erosion are present with initial sclero-atrophy; (d) Dermoscopy shows white/yellow scales, dilated vessel and “peripheral horizontal hair” sign (black arrow). (e) Clinical image with alopecia patches; (f) “Peripheral horizontal hair” sign (black arrow) along with linear-curved/branched vessels on an erythematous background are evident on dermoscopic examination. (g) Pityriasis amiantacea: clinical image shows thick adherent asbestos-like scales encasing hair shafts with underlying scalp erythema; (h) Dermoscopic features include thick yellowish-white scales, enveloping hair shafts, and focal scalp erythema. (i) Discoid lupus erythematosus of the scalp: clinical picture; (j) trichoscopy reveals white scarring areas, lack of hair, and peripheral brownish pigmentation. (k) Folliculitis decalvans: yellow crusting and patches of alopecia of the vertex on clinical assessment; (l) Dermoscopy shows tufted hairs, perifollicular pustules, yellowish perifollicular scaling and a prominent erythematous background with polymorphic vessels.

Turning to the comparative analyses, we found the following findings to be statistically more common in EPDS compared to controls in bald patients (Table 1): linear-curved vessels (p<0.001), orange structureless areas (diffuse) (p<0.001), and non-follicular pustules (p=0.005). Notably, this last feature was more frequent in EPDS affecting hairy patients compared to controls (p<0.001), along with yellow scales/crusts (focal), and “peripheral horizontal hair” sign (p<0.001 for both of them). On the other hand, perifollicular white scales (p=0.003) and follicular plugs (p<0.001) were more common in the control group. For further analytical details see Table 2.

In terms of inter-observer concordance about dermoscopic findings in EPDS, we found significant agreement between evaluators, with Kappa values being 0.81 and 0.83 for bald and hairy cohorts, respectively.

Discussion

Our study underlines that dermoscopy may be of help in assisting the recognition of EPDS in both bald and hairy patients. In particular, the diagnostic clues in the former population included linear-curved vessels, orange structureless areas (diffuse), and non-follicular pustules. The first two features were already reported as common findings in bald patients with EPDS histologically typified by lymphoplasmacellular infiltrate in a small case-series published by some of the authors of this study.⁴ Interestingly, orange structureless areas were thought to be related to either a compact cellular infiltrate in the dermis (“mass effect”) or dermal hemosiderin deposits.^5,7 Of note, similarly to Zoon’s balanitis, the presence of linear-curved vessels has been described as a possible dermoscopic clue of lymphoplasmacellular EPDS in bald population, in line with our results.⁴ This association is explained by the histological background as EPDS is characterized by atrophic epidermis, thus allowing the visualization of enlarged dermal vessels.^3,7 Importantly, in our study we did not perform a dermoscopic-pathological correlation analysis, thereby preventing us to establish a possible correlation between the aforementioned dermoscopic findings and the type of cellular infiltrate.

More recently, Gupta et al described a rare case of Erlotinib-induced EPDS, highlighting trichoscopic findings that further expand the spectrum of EPDS presentations.⁸ While our study focuses on spontaneous EPDS cases, the inclusion of drug-induced variants in future analyses may provide a broader understanding of its dermoscopic variations and pathophysiological mechanisms. These insights emphasize the need for further prospective studies, particularly integrating trichoscopy with histopathological correlation, to refine diagnostic accuracy and guide clinical management.

Another relevant feature emerging from our study was the association between non-follicular pustules resulting from the neutrophilic infiltration of the epidermis/upper dermis and EPDS in both bald and hairy patients. Besides such a dermoscopic clue, other typical findings of EPDS found in our hairy population included non-follicular focal yellow scales/crusts due to the drying of pustules as well as the presence of horizontally arranged hairs with proximal part of the shaft being visible through a thinned epidermis at the edge of alopecic areas (“peripheral horizontal hair” sign). According to our comparative analysis, the latter clue is indicative of EPDS in hairy patients as it was not seen in the main differential diagnoses. This is consistent with a case-series including 20 instances reported by some of the authors of the present analysis that found bulbs of the hair follicles visible through the thinned skin to be one of the most common dermoscopic findings of EPDS in hairy subjects.³ From a dermoscopic-pathological correlation point of view, it is possible that “peripheral horizontal hair” sign is due to the remarkable dermal fibrosis typical of EPDS that may give rise to displacement of the hair roots towards the atrophic skin surface at the periphery of the alopecic patches.^3,7 Notably, unlike our analysis, perifollicular yellow crusts turned out to be the most frequent scaling/crusting pattern in the previous study.³ Additionally, we also observed non-follicular pustules to be another distinctive feature of EPDS, while they were not reported in the prior analysis. It is likely that such differences might be due to the selection of lesions in different stages of development.

Limitations

The main limitations of the study include (I) the retrospective design, which is prone to recall and observation biases, that were addressed by involving evaluators who did not contribute to the sample collection; (II) the lack of dermoscopic-pathological correlation analysis; and (III) the potential variability introduced by having multiple investigators collecting the images at different centers, yet the use of the same dermoscopic device reduced the relevance of such a bias. Therefore, future studies addressing such points are needed to confirm our preliminary data.

Conclusions

This study highlights the diagnostic value of dermoscopy in EPDS, providing distinctive dermoscopic features that help differentiate EPDS from other scalp conditions. In bald patients, the most characteristic findings included linear-curved vessels, diffuse orange structureless areas, and non-follicular pustules. In hairy patients, EPDS was associated with focal yellow scales/crusts, non-follicular pustules, and the “peripheral horizontal hair” sign. The identification of these dermoscopic clues may assist dermatologists in reducing diagnostic uncertainty and potentially decreasing the need for invasive biopsies in suspected cases of EPDS.