Accuracy of dermoscopic criteria for the diagnosis of psoriasis [PDF]

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/221797319

Accuracy of dermoscopic criteria for the diagnosis of psoriasis, dermatitis, lichen planus and pityriasis rosea ARTICLE in BRITISH JOURNAL OF DERMATOLOGY · FEBRUARY 2012 Impact Factor: 4.28 · DOI: 10.1111/j.1365-2133.2012.10868.x · Source: PubMed

CITATIONS

READS

77

86

11 AUTHORS, INCLUDING: Athanassios Kyrgidis

Dimitrios Ioannides

Aristotle University of Thessaloniki

Aristotle University of Thessaloniki

111 PUBLICATIONS 1,068 CITATIONS

87 PUBLICATIONS 1,340 CITATIONS

SEE PROFILE

SEE PROFILE

Giuseppe Argenziano

Iris Zalaudek

Second University of Naples

Medical University of Graz

478 PUBLICATIONS 7,571 CITATIONS

373 PUBLICATIONS 4,966 CITATIONS

SEE PROFILE

SEE PROFILE

Available from: Athanasios Karatolias Retrieved on: 27 January 2016

BJD C L I N I C A L A N D L A B O R A T O R Y I N V E S TI G A T I O N S

British Journal of Dermatology

Accuracy of dermoscopic criteria for the diagnosis of psoriasis, dermatitis, lichen planus and pityriasis rosea A. Lallas, A. Kyrgidis,* T.G. Tzellos,– Z. Apalla, E. Karakyriou, A. Karatolias, I. Lefaki, E. Sotiriou, D. Ioannides, G. Argenziano and I. Zalaudek§ State Clinic of Dermatology, Hospital of Skin and Venereal Diseases, Delfon 124, 54643 Thessaloniki, Greece *Department of Otolaryngology–Head & Neck Surgery and First Dermatologic Department, Medical School, Aristotle University, Thessaloniki, Greece Dermatology Unit, Medical Department, Arcispedale Santa Maria Nuova, Reggio Emilia, Italy §Department of Dermatology, Medical University of Graz, Graz, Austria –Division of Evidence Based Dermatology, Departments of Dermatology, Venerology, Allergology and Immunology, Dessau Medical Center, Dessau, Germany

Summary Correspondence Aimilios Lallas. E-mail: [email protected]

Accepted for publication 24 January 2012

Funding sources None.

Conflicts of interest None declared. A.L. and A.K. contributed equally in the writing of this manuscript and share first authorship equally. DOI 10.1111/j.1365-2133.2012.10868.x

Background Dermoscopy is useful in evaluating skin tumours, but its applicability extends also to the field of inflammatory skin disorders. Plaque psoriasis (PP), dermatitis, lichen planus (LP) and pityriasis rosea (PR) are common inflammatory skin diseases, but little is currently known about their dermoscopic features. Objectives To determine and compare the dermoscopic patterns associated with PP, dermatitis, LP and PR and to assess the validity of certain dermoscopic criteria in the diagnosis of PP. Methods Patients with PP, dermatitis, LP and PR were prospectively enrolled. The single most recently developed lesion was examined dermoscopically and histopathologically. Variables included vascular morphology, vascular arrangement, background colour, scale colour, scale distribution and presence of Wickham striae. Univariate and adjusted odds ratios were calculated. Discriminant functions were used to plot receiver-operator characteristic curves. Results Eighty-three patients with PP and 86 patients with either dermatitis, LP or PR were included in the study. Dotted vessels in a regular arrangement over a light red background and white scales were highly predictive for the diagnosis of PP, whereas dermatitis more commonly showed yellow scales and dotted vessels in a patchy arrangement. PR was characterized by yellowish background, dotted vessels and peripheral scales; whitish lines (Wickham striae) were seen exclusively in LP. Conclusions PP, LP, PR and dermatitis show specific dermoscopic patterns that may aid their clinical diagnosis. Certain combinations of dermoscopic features can reliably predict the diagnosis of PP.

Plaque psoriasis (PP), dermatitis, pityriasis rosea (PR) and lichen planus (LP) are common inflammatory skin diseases. Their characteristic appearance allows a clinical diagnosis in a high proportion of patients.1–4 However, unusual presentations at times do exist and may cause difficulties in the differentiation among these entities.1–4 In those cases, histopathology contributes significantly to the accurate diagnosis.5 Dermoscopy as a noninvasive tool has become standard in the preoperative evaluation of skin tumours.6 Thereby dermoscopy offers the benefit to assess pigmented and vascular structures that are not visible clinically. Particularly the improved visualization of vessels, but also colour variegations that are difficult to recognize with the

naked eye, explain why dermoscopy is gaining increasing importance in the diagnosis of skin lesions in general dermatology: today dermoscopy is used in the realm of skin infections and infestations (entomodermoscopy), hair and scalp disorders (trichoscopy), nailfold capillary abnormalities (capillaroscopy) and inflammatory skin disorders.7 Dermoscopic patterns of red dots or globules arranged in a homogeneous, regular or ring-like fashion have been described as common findings in PP.8–10 However, there are limited data available concerning the dermoscopic pattern of dermatitis, PR and LP; moreover, sporadic cases of dermatitis reportedly show dermoscopic patterns similar to those in PP, namely predominantly red dots.11  2012 The Authors

1198

BJD  2012 British Association of Dermatologists 2012 166, pp1198–1205

Dermoscopy of inflammatory skin diseases, A. Lallas et al. 1199

Given that PP and other inflammatory skin diseases may sometimes be difficult to differentiate clinically, a more detailed determination of specific dermoscopic patterns of inflammatory skin diseases could be a valuable addition for the clinical assessment. The aim of this study was to determine the dermoscopic patterns associated with PP, dermatitis, LP and PR and to assess the validity of certain dermoscopic criteria in the diagnosis of PP.

Patients and methods This was a prospective study approved by the relevant authorities and the Human Ethics Review Committee of the School of Medicine of Aristotle University prior to patient enrolment. It was conducted in accordance with the Declaration of Helsinki and all patients provided written informed consent after the study was explained and their questions answered and before study procedures were initiated. The study was conducted at the Hospital of Skin and Venereal Diseases in Thessaloniki, Greece between January and May 2011. Consecutive patients presenting at the outpatient clinic were screened for eligibility to participate in the study. Inclusion criteria for study participation were a clinical diagnosis of PP, dermatitis (exogenous and endogenous types), LP and PR affecting the trunk and ⁄or upper or lower extremities. A large proportion of dermatitis cases belonged to the spectrum of nummular eczema, which is commonly included in the differential diagnosis of PP.1 Additionally, more generalized forms of dermatitis were included in the study. Exclusion criteria were topical or systemic treatment (ciclosporin, biologics, methotrexate, retinoids, corticosteroids) performed less than 1 and 6 months, respectively, before recruitment. Lesions located on the scalp, palms, soles, intertriginal and genital areas were excluded from the study. Patient demographics were recorded and the single most recently developed lesion was examined dermoscopically and histopathologically. Dermoscopy preceded histology, and no treatment was allowed in the interim. Patients received appropriate treatment after the histological report. Capture of dermoscopic images was performed using a digital dermoscopy system (Ecoscan II V0990Æ21, Canon 1000D, original magnification · 10; Canon, Tokyo, Japan). Minimal pressure was applied and ultrasound gel was used in order to preserve vessel morphology and ensure the best available projection. Dermoscopy image capturing was performed by a single practitioner to avoid diversification during the procedure. Dermoscopic evaluation was performed by two independent dermoscopists (A.L., E.K.), who were unaware of the histopathological diagnosis. Variables included in the dermoscopic evaluation were: (i) vascular morphology (dotted, linear, dotted + linear); (ii) vascular arrangement (regular, in clusters, patchy, peripheral, in rings); (iii) background colour (dull red, i.e. intense red colour, light red, i.e. fading red colour, yellowish); (iv) scale colour (white, yellow, white + yellow); (v) scale distribution (patchy, peripheral, diffuse, central); and (vi) presence of white crossing streaks (i.e. Wickham striae).  2012 The Authors BJD  2012 British Association of Dermatologists 2012 166, pp1198–1205

Selection of the dermoscopic variables included in the evaluation process was based on the available literature data, expertise and a pilot study (data not shown). Arrangement of the vessels was judged as regular when vessels were distributed uniformly throughout the lesion, clustered when they were aggregated in small groups and seen only in some areas of the lesion, patchy when they were arranged in an asymmetrical distribution which could not be classified as clustered, peripheral when they were observed predominantly at the periphery of the lesion, and in rings when they were arranged in irregular circles or rings (Fig. 1). Scale distribution was considered as an independent variable and was classified as diffuse when scales were arranged in a homogeneous pattern all over the lesion, patchy when scales covered a portion of the lesion asymmetrically, and central and peripheral when they were mainly arranged at the centre and periphery, respectively. Representative examples of criteria and colour variegations are shown in Figures 1 and 2, respectively. Similarly to the dermoscopic assessment, histopathological evaluation was also performed by two independent dermatopathologists (Z.A., I.L.), who were blinded to the dermoscopic diagnosis. Histological diagnosis of PP, dermatitis, PR and LP was based on the identification of the hallmarks of each disease under microscopic examination. Analytically, psoriasiform epidermal hyperplasia accompanied by tortuous, dilated capillaries in the superficial papillary dermis and a perivascular mononuclear cell infiltrate were the criteria for the histopathological diagnosis of PP.5 Criteria for dermatitis included presence of acanthosis, spongiosis and exocytosis of the epidermis, plus a mixed inflammatory cell infiltrate surrounding the superficial vascular plexus, composed of lymphocytes, histiocytes and eosinophils.5 Characteristic histopathological features of LP included wedge-shaped hypergranulosis, saw-toothed acanthosis, liquefaction degeneration of the basal layer and a lymphohistiocytic band-like infiltrate occupying the upper dermis.12 Histopathological criteria for diagnosis of PR included focal hyperkeratosis and angulated parakeratosis, hypogranulosis beneath the foci of parakeratosis, slight spongiosis and a lymphohistiocytic infiltrate surrounding the superficial vascular plexus.5 A negative periodic acid–Schiff stain to exclude superficial dermatophytosis was a prerequisite for the final histopathological diagnosis of the above-mentioned dermatoses. Discrepancies in the assessment of morphological criteria and diagnosis between the two dermoscopists or pathologists were resolved by consensus. All evaluators were unaware of patient clinical information. Statistical analysis The outcome dichotomous variable was set to definite histological diagnosis of psoriasis. All separate dermoscopic variables were included in the analysis.

1200 Dermoscopy of inflammatory skin diseases, A. Lallas et al.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

Fig 1. Representative dermoscopic examples of vessel arrangements: (a) regular (psoriasis), (b) in clusters (dermatitis), (c) patchy (dermatitis), (d) in rings (psoriasis) and (e) peripheral (lichen planus); and scale distribution: (f) central (psoriasis), (g) peripheral (pityriasis rosea), (h) diffuse (lichen planus) and (i) patchy (psoriasis).

(a)

(b)

(c)

Fig 2. Representative examples of colour variegations: (a) dull red (dermatitis), (b) light red (psoriasis) and (c) yellow (pityriasis rosea).

Colinearity was assessed via a correlation matrix, using Spearman’s rho correlation coefficient. Relative risks were calculated for all dichotomous variables; for categorical variables relative risks were approximated by odds ratios (ORs). Crude ORs, adjusted ORs and corresponding 95% confidence intervals (CIs) were calculated by univariate and conditional multivariate logistic regression, respectively. Forward inclusion with likelihood ratio criteria proved more parsimonious. Alpha level was set at 0Æ05 while an alpha level of 0Æ10 was used as cutoff for variable removal in the automated model selection for multivariate logistic regression. Logistic regression models were determined to account for outcome better than would be expected by chance

(P < 0Æ001) via Wald test and v2 likelihood ratio test. Goodness of fit was examined by adjusted R2. Tests for interactions were automatically commenced by the statistical package through the fitting procedure. We then used the logistic regression derived model to predict the event probability for a dichotomous outcome variable. Given the sets of independent variables produced by multivariate logistic regression, attempts to find linear combinations of those variables that best separate the groups of cases were made (discriminant analysis). Specificity and sensitivity were extracted from classification tables. Discriminant functions were saved and we then used receiveroperator characteristic (ROC) curves to choose between competing classification schemes.  2012 The Authors BJD  2012 British Association of Dermatologists 2012 166, pp1198–1205

Dermoscopy of inflammatory skin diseases, A. Lallas et al. 1201

Eighty-three patients with PP and 86 control subjects with dermatitis, LP or PR were recruited in this diagnostic accuracy study. Both false positives and false negatives would be classified as failures. Assuming 10% failures among controls and 28% failures among patients, we would be able to reject the null hypothesis that the failure rates for experimental and control subjects are equal with probability (power) 0Æ804. A corrected v2 statistic was used to evaluate this null hypothesis. The type I error probability associated with all tests in this study was set to 0Æ05. All statistical calculations were made with SPSS 17Æ0 (SPSS Inc., Chicago, IL, U.S.A.).

Results During the 4 months of study enrolment, 275 patients were screened for eligibility of study participation. Of those, 106 (38Æ5%) patients were excluded because of concurrent treatment (n = 72) or disease involvement exclusively on body sites excluded in the study design (n = 18), withdrew from study participation (n = 5) or lacked a definitive histopathological diagnosis (n = 11) resulting in a total of 169 (61Æ5%) patients included in the study. Of those, 83 (49Æ1%) patients had PP, 41 (24Æ3%) had dermatitis, 25 (14Æ8%) had LP and 20 (11Æ9%) had PR.

Table 1 Frequency of dermoscopic variables in psoriasis, dermatitis, lichen planus and pityriasis rosea Dermoscopic variables Background colour Light red Dull red Yellowish Type of vessels Dotted Linear Dotted + linear Pattern of vessels Regular In clusters Patchy Peripheral In rings Scale colour White Yellow White + yellow Scale distribution Patchy Peripheral Diffuse Central Wickham striae

Descriptive results of the dermoscopic analysis are quoted in Table 1. Dotted vessels were seen in all cases of PP and PR, in 39 ⁄41 (95%) cases of dermatitis and in 9 ⁄25 (36%) cases of LP. Even though dotted vessels were common among all entities, their arrangement and other associated criteria differed significantly between PP, dermatitis, LP and PR. Dotted vessels in PP were most commonly arranged in a regular distribution (73 ⁄83; 88%) and were associated with white scales (58 ⁄83; 70%) (Fig. 3a); in contrast, vessels in dermatitis appeared more commonly in a patchy distribution (24 ⁄41; 59%) and in association with yellow scales (25 ⁄41; 61%) (Fig. 3b). In PR, dotted vessels were mostly associated with a yellowish background colour (13 ⁄20; 65%) and a peripheral arrangement of scales (14 ⁄20; 70%) (Fig. 3c). Instead, white crossing lines (Wickham striae) were seen exclusively in LP (24 ⁄25; 96%) and occurred together with a dull red background colour and peripheral arrangement of vessels in 64% (16 ⁄25) and 60% (15 ⁄25), respectively (Fig. 3d). For the further statistical analysis to assess criteria predictive for PP, patients with dermatitis, LP and PR were matched as controls. Mean age of patients with PP was 47Æ8 years, and 44Æ3 years for the controls. Male to female ratio for PP and control group was 1Æ2 : 1 and 1 : 1, respectively. Table 2 summarizes the descriptive results of the univariate analysis.

Psoriasis (n = 83), n (%)

Dermatitis (n = 41), n (%)

Lichen planus (n = 25), n (%)

Pityriasis rosea (n = 20), n (%)

34 (41) 48 (58) 1 (1)

6 (15) 27 (66) 7 (17)

3 (12) 16 (64) 6 (24)

2 (10) 5 (25) 13 (65)

< 0Æ001

83 (100) 0 0

39 (95) 0 2 (5)

9 (36) 3 (12) 13 (52)

20 (100) 0 0

< 0Æ001 < 0Æ001

73 1 2 0 7

14 (34) 3 (7) 24 (59) 0 0

0 0 10 (40) 15 (60) 0

4 (20) 1 (5) 15 (75) 0 0

< 0Æ001 < 0Æ001 < 0Æ025

58 (70) 0 2 (2)

10 (24) 8 (20) 17 (41)

2 (8) 3 (12) 0

17 (85) 1 (5) 0

< 0Æ005 < 0Æ005 < 0Æ001

11 2 37 10 0

27 (66) 0 7 (17) 1 (2Æ4) 0

2 (8) 0 0 3 (12) 24 (96)

0 14 (70) 3 (15) 1 (5) 0

< 0Æ001 < 0Æ001 < 0Æ001

(88) (1) (2) (8)

(13) (2) (45) (12)

P-value < 0Æ001

< 0Æ001

< 0Æ001

Bold values denote positive significant correlation using v2 in comparison with the other diseases.

 2012 The Authors BJD  2012 British Association of Dermatologists 2012 166, pp1198–1205

1202 Dermoscopy of inflammatory skin diseases, A. Lallas et al.

(a)

(b)

(c)

(d)

Fig 3. Representative dermoscopic images of (a) psoriasis exhibiting regularly distributed dotted vessels over a light red background and white scales, (b) dermatitis revealing dotted vessels associated with yellow scales, (c) pityriasis rosea exhibiting yellowish background colour, peripheral arrangement of the scales and patchy distribution of loosely arranged dotted vessels, and (d) lichen planus showing Wickham striae.

Table 2 Univariate analysis: dermoscopy predictors of psoriasis from 169 patients

Predictor Background colour Light red Dull red Yellowish Type of vessels Dotted Linear Dotted + linear Pattern of vessels Regular In clusters Patchy Peripheral In rings Scale colour White Yellow White + yellow Scale distribution Patchy Peripheral Diffuse Central Wickham striae

Patients with psoriasis (n = 83), n (%)

Controls – nonpsoriasis (n = 86), n (%)

P-value

34 (41) 48 (58) 1 (1)

11 (13) 48 (56) 26 (30)

< 0Æ001 0Æ791 0Æ001

83 (100) 0 0

68 (79) 3 (3) 15 (17)

0Æ998 0Æ999 0Æ998

73 1 2 0 7

18 4 49 15 0

RR 4Æ731 1Æ086 0Æ028 0 0 0

95% CI 2Æ192–10Æ211 0Æ591–1Æ996 0Æ004–0Æ213 – – –

(21) (5) (57) (17)

< 0Æ001 0Æ439 < 0Æ001 0Æ998 0Æ999

27Æ578 0Æ506 0Æ019 0 0

11Æ898–63Æ921 0Æ090–2Æ841 0Æ004–0Æ081 – –

58 (70) 0 2 (2)

29 (34) 12 (14) 17 (20)

< 0Æ001 0Æ999 0Æ003

4Æ560 0 0Æ100

2Æ386–8Æ716 – 0Æ022–0Æ449

11 2 37 10 0

29 14 10 5 24

0Æ002 0Æ008 < 0Æ001 0Æ163 0Æ998

0Æ300 0Æ127 6Æ113 2Æ219 0

0Æ138–0Æ653 0Æ028–0Æ578 2Æ778–13Æ452 0Æ725–6Æ796 –

(88) (1) (2) (8)

(13) (2) (45) (12)

(34) (16) (12) (6) (28)

RR, relative risk; CI, confidence interval. Statistically significant values are shown in bold.

Univariate analysis yielded no statistical significance with regard to the types of vessels. However, PP was found to correlate significantly with a regular arrangement of dotted vessels (rho = 0Æ672, P < 0Æ001), a light red background colour (rho = 0Æ319, P < 0Æ001) and a diffuse distribution

(rho = 0Æ368, P < 0Æ001) of white scales (rho = 0Æ362, P < 0Æ001). Based on the univariate logistic regression analysis, regular vascular arrangement yielded a 27-fold higher probability of being PP (relative risk 27Æ58, 95% CI 11Æ90–63Æ92) compared  2012 The Authors BJD  2012 British Association of Dermatologists 2012 166, pp1198–1205

Dermoscopy of inflammatory skin diseases, A. Lallas et al. 1203 Table 3 Multivariate analysis with all dermoscopic variables entered: adjusted dermoscopy predictors for psoriasis from 169 patients Predictor Background colour Light red Dull red Pattern of vessels Regular In rings Scale colour White + yellow

P-value

RR

95% CI

< 0Æ001 0Æ008

182Æ535 22Æ107

11Æ785–2827Æ343 2Æ255–216Æ688

< 0Æ001 0Æ999

143Æ511 0

27Æ517–748Æ450 –

0Æ002

0Æ050

Table 4 Multivariate analysis with only variables with significant prognostic value in the univariate analysis included: adjusted dermoscopy predictors for psoriasis from 169 patients Predictor

0Æ008–0Æ322

Variables entered in model were background colour: light red, dull red, yellowish; type of vessels: dotted, linear, dotted + linear; pattern of vessels: regular, in clusters, patchy, in rings, peripheral; scale colour: white, yellow, white + yellow; scale distribution: patchy, peripheral, diffuse, central; Wickham striae. Relative risks (RRs) approximated by odds ratios, 95% confidence intervals (CIs) and P-values were calculated with conditional multivariate logistic regression backward elimination model according to likelihood ratio criteria. RRs mutually adjusted for variables in the model. Logit for outcome variable ‘psoriasis’ is estimated for psoriasis = no. Logit for all other independent dichotomous variables = no. Alpha level set to P < 0Æ05. Cut-off value set to 0Æ10. Statistically significant values are shown in bold.

Background colour Light red Yellowish Pattern of vessels Regular Patchy Scale colour White + yellow

P-value

RR

95% CI

0Æ001 0Æ015

11Æ340 0Æ060

2Æ603–49Æ403 0Æ006–0Æ578

< 0Æ001 0Æ017

17Æ188 0Æ099

4Æ551–64Æ916 0Æ015–0Æ665

0Æ002

0Æ051

0Æ008–0Æ338

Variables entered in model were background colour: light red, yellowish; type of vessels: dotted vessels, linear vessels, dotted + linear; pattern of vessels: regular, patchy; scale colour: white, white + yellow; scale distribution: patchy, peripheral, diffuse; Wickham striae. Relative risks (RRs) approximated by odds ratios, 95% confidence intervals (CIs) and P-values were calculated with conditional multivariate logistic regression backward elimination model according to likelihood ratio criteria. RRs mutually adjusted for variables in the model. Logit for outcome variable ‘psoriasis’ is estimated for psoriasis = no. Logit for all other independent dichotomous variables = no. Alpha level set to P < 0Æ05. Cut-off value set to 0Æ10. Statistically significant values are shown in bold.

with controls. Light red background colour, white or diffuse scales increased the likelihood for PP by 4-fold, 4-fold and 6fold, respectively (see Table 2). Multivariate analysis Two multivariate logistic regression models were performed. In the first, we entered as dichotomous independent variables all dermoscopic variables. Forward inclusion with likelihood criteria was utilized. This model yielded a specificity of 96Æ4% and sensitivity of 86Æ0% for the diagnosis of PP. Results are presented in Table 3. In the second model, only those variables with significant prognostic value in the univariate analysis were included (Table 4). Of note, both models yielded the same variables, a fact that signifies the independent prognostic value of the models. The specificity achieved was 88Æ0% while sensitivity was 84Æ9%. The ROC curve for this second model is presented in Figure 4.

Discussion Our study suggests significant differences in the dermoscopic patterns of PP, dermatitis, LP and PR, which may assist the clinical diagnosis in selected cases (Figs 5 and 6). Dotted vessels are a well-recognized criterion for the diagnosis of PP,8–11 and were effectively seen in all our cases of PP. However, our and previous studies showed that dotted vessels are not limited to PP but occur at variable frequency in  2012 The Authors BJD  2012 British Association of Dermatologists 2012 166, pp1198–1205

Fig 4. Receiver-operator characteristic (ROC) curve. Backward elimination according to likelihood ratio criteria. Area under curve (AUC) = 0Æ935, 95% confidence interval 0Æ896–0Æ974 (P < 0Æ001).

several other inflammatory and neoplastic lesions.7,11,13 Accordingly, dotted vessels as the only dermoscopic criterion are insufficient to distinguish between these different entities accurately. Besides the vascular morphology, the vascular arrangement and specific dermoscopic clues have been judged to be of equal importance in the differential diagnosis of nonpigmented skin lesions.9 This is further supported by our study, which revealed significant differences with respect to the distribution of vessels or additional criteria among PP, dermatitis, LP and PR.

1204 Dermoscopy of inflammatory skin diseases, A. Lallas et al.

(a)

(b)

(c)

(d)

In detail, the combination of regularly distributed dotted vessels over a light red background associated with diffuse white scales was highly predictive of PP and allowed a correct diagnosis with 88Æ0% specificity and 84Æ9% sensitivity; conversely, yellow scales, a patchy arrangement of vessels or yellow background colour decreased significantly the likelihood for PP by 20-fold, 10-fold and 17-fold, respectively. Yellow serocrusts have very recently been described as dermoscopic finding in two cases of nummular eczema.14 This, along with our findings, suggests that white vs. yellow scales along with regular vs. patchy distribution of dotted vessels may represent a valuable clue in the differential diagnosis of PP and nummular eczema, respectively (Fig. 5c, d). On the other hand, although red globular rings (i.e. red globules arranged in irregular circles or rings) as described previously by Va´zquez-Lo´pez et al.10 represented a highly specific feature for PP in our study, this pattern was seen in only a minority of cases (8%) in our series. Therefore, the

(a)

(b)

(c)

(d)

Fig 5. Side-by-side comparison of clinical images of plaque psoriasis (PP) (a) and nummular eczema (b). In both cases, a biopsy was required for the diagnosis. Squares in (a) and (b) indicate the corresponding dermoscopic image. Upon dermoscopic examination, PP shows regularly arranged dotted vessels and white scales (c); in contrast, nummular eczema reveals yellow diffuse scales and some dotted vessels (d).

value of this pattern in the diagnosis of PP remains to be elucidated further. Our study furthermore confirms preliminary observations on the dermoscopic patterns of PR and LP.10,11,15 As such, PR was typified by peripheral scaling (so-called collarette scales) around a diffuse and structureless yellowish centre; although dotted vessels were seen in all our cases of PR, they were generally much less evident and fewer in number compared with PP or dermatitis (Fig. 6c). By contrast, Wickham striae were seen exclusively in LP and observed in all but one case (Fig. 6d); thus our findings highlight the importance of Wickham striae in the diagnosis of LP. It still remains to be determined how much dermoscopy may aid the differential diagnosis of tumours from PP; this question seems particularly important when facing patients with diffuse PP, who received past psoralen plus ultraviolet A therapy and are at increased risk for nonmelanoma skin cancer.16,17 With regard to this, Pan et al.18 introduced a dermo-

Fig 6. Side-by-side comparison of clinical images of pityriasis rosea (a) and lichen planus (b). Squares in (a) and (b) indicate the corresponding dermoscopic image. (c) Dermoscopically, pityriasis rosea reveals peripheral scales over a yellowish background along with some loosely arranged dotted vessels. (d) Dermoscopy of lichen planus shows white crossing lines (Wickham striae) and dotted vessels.  2012 The Authors BJD  2012 British Association of Dermatologists 2012 166, pp1198–1205

Dermoscopy of inflammatory skin diseases, A. Lallas et al. 1205

scopic diagnostic model for differentiating solitary psoriatic plaques from intraepidermal carcinoma (IEC) and superficial basal cell carcinoma (sBCC). They concluded that red dots, homogeneous vascular pattern and light red background were significant dermoscopic features for psoriasis, yielding a diagnostic probability of 99% if all three features were present. In contrast, clustered vessels, glomerular vessels and hyperkeratosis yield a 98% probability for the diagnosis of IEC, whereas four of six described criteria (i.e. scattered vascular pattern, arborizing microvessels, telangiectatic or atypical vessels, milky-pink background, and brown dots ⁄globules) achieved a diagnostic probability of 99% for sBCC. As our dataset did not include tumours, the impact of dermoscopy in differentiating PP from skin tumours such as IEC and sBCC cannot be assessed by our study. Our study has some limitations: first, we included only white patients from a Mediterranean area. Given that distribution studies stress the significance of geographical and ethnic background in the clinical presentation of psoriasis,19 our results are limited to this population. Second, although recent studies suggest significant differences in the dermoscopic patterns between scalp psoriasis and seborrhoeic dermatitis,20 our study excluded lesions from the scalp, and acral, intertriginal and genital areas. Therefore, no conclusions about the impact of dermoscopy in the differential diagnosis of psoriasis at these sites from other erythematosquamous diseases can be drawn. Finally, our study design does not allow any conclusions about the impact of dermoscopy in the differential diagnosis of PP and nonmelanoma skin cancer; however, the data of Pan et al.18 point towards such a role. In conclusion, PP, dermatitis, LP and PR reveal specific dermoscopic patterns that may aid the clinical diagnosis. Besides its diagnostic purposes, dermoscopy might provide a useful tool for the evaluation of treatment outcome in patients with PP such as early detection of treatment response or unwarranted side-effects of long-term topical treatment (for example, the corticosteroid-induced atrophy in patients with psoriasis).21 The feasibility of our observations in clinical practice, as well as in studying the course of PP, warrants, however, further clinical studies.

What’s already known about this topic? • Plaque psoriasis, dermatitis, lichen planus and pityriasis rosea are common inflammatory skin diseases, but little is currently known about their dermoscopic features. • Applicability of dermoscopy is gradually expanding in evaluation of inflammatory skin disorders.

What does this study add? • Each of the above-mentioned diseases exhibits a characteristic dermoscopic pattern. • A certain combination of dermoscopic features is more predictive of the diagnosis of plaque psoriasis.

 2012 The Authors BJD  2012 British Association of Dermatologists 2012 166, pp1198–1205

References 1 James WD, Berger TG, Elston DM. Psoriasis. In: Andrews’ Diseases of the Skin: Clinical Dermatology, 10th edn. Philadelphia: Elsevier Saunders, 2006; 193–202. 2 James WD, Berger TG, Elston DM. Eczema. In: Andrews’ Diseases of the Skin: Clinical Dermatology, 10th edn. Philadelphia: Elsevier Saunders, 2006; 77–83. 3 James WD, Berger TG, Elston DM. Pityriasis rosea. In: Andrews’ Diseases of the Skin: Clinical Dermatology, 10th edn. Philadelphia: Elsevier Saunders, 2006; 208–9. 4 James WD, Berger TG, Elston DM. Lichen planus. In: Andrews’ Diseases of the Skin: Clinical Dermatology, 10th edn. Philadelphia: Elsevier Saunders, 2006; 217–23. 5 McKee PH, Calonje JE, Granter SR. Spongiotic, psoriasiform and pustular dermatoses. In: Pathology of the Skin with Clinical Correlations, 3rd edn. St Louis: Elsevier Mosby, 2005; 171–216. 6 Argenziano G, Soyer HP, Chimenti S et al. Dermoscopy of pigmented skin lesions: results of a consensus meeting via the internet. J Am Acad Dermatol 2003; 48:679–93. 7 Zalaudek I, Argenziano G, Di Stefani A et al. Dermoscopy in general dermatology. Dermatology 2006; 212:7–18. 8 Va´zquez-Lo´pez F, Manjon-Haces JA, Maldonado-Seral C et al. Dermoscopic features of plaque psoriasis and lichen planus: new observations. Dermatology 2003; 207:151–6. 9 Zalaudek I, Argenziano G. Dermoscopy subpatterns of inflammatory skin disorders. Arch Dermatol 2006; 142:808. 10 Va´zquez-Lo´pez F, Zaballos P, Fueyo-Casado A, Sa´nchez-Martı´n J. A dermoscopy subpattern of plaque-type psoriasis: red globular rings. Arch Dermatol 2007; 143:1612. 11 Va´zquez-Lo´pez F, Kreusch J, Marghoob AA. Dermoscopic semiology: further insights into vascular features by screening a large spectrum of nontumoral skin lesions. Br J Dermatol 2004; 150:226– 31. 12 McKee PH, Calonje JE, Granter SR. Lichenoid and interface dermatoses. In: Pathology of the Skin with Clinical Correlations, 3rd edn. St Louis: Elsevier Mosby, 2005; 217–60. 13 Argenziano G, Zalaudek I, Corona R et al. Vascular structures in skin tumors: a dermoscopy study. Arch Dermatol 2004; 140:1485–9. 14 Navarini AA, Feldmeyer L, To¨ndury B et al. The yellow clod sign. Arch Dermatol 2011; 147:1350. 15 Chuh AA. Collarette scaling in pityriasis rosea demonstrated by digital epiluminescence dermatoscopy. Australas J Dermatol 2001; 42:288–90. 16 Lim JL, Stern RS. High levels of ultraviolet B exposure increase the risk of non-melanoma skin cancer in psoralen and ultraviolet A-treated patients. J Invest Dermatol 2005; 124:505–13. 17 Stern RS, Lunder EJ. Risk of squamous cell carcinoma and methoxsalen (psoralen) and UV-A radiation (PUVA). A meta-analysis. Arch Dermatol 1998; 134:1582–5. 18 Pan Y, Chamberlain AJ, Bailey M et al. Dermatoscopy aids in the diagnosis of the solitary red scaly patch or plaque – features distinguishing superficial basal cell carcinoma, intraepidermal carcinoma and psoriasis. J Am Acad Dermatol 2008; 59:268–74. 19 Rigopoulos D, Gregoriou S, Katrinaki A et al. Characteristics of psoriasis in Greece: an epidemiological study of a population in a sunny Mediterranean climate. Eur J Dermatol 2010; 20:189–95. 20 Kim GW, Jung HJ, Ko HC et al. Dermoscopy can be useful in differentiating scalp psoriasis from seborrhoeic dermatitis. Br J Dermatol 2011; 164:652–6. 21 Va´zquez-Lo´pez F, Marghoob AA. Dermoscopic assessment of longterm topical therapies with potent steroids in chronic psoriasis. J Am Acad Dermatol 2004; 51:811–13.