Dermoscopy is a non-invasive complementary examination that allows the examiner to observe structures in cutaneous lesions not visible to the naked eye. The recognition of these structures and their histopathological correspondence is of great diagnostic assistance in dermatology. Its main application is for early diagnosis of melanoma. Provided adequate training is available, the technique can offer increased diagnostic accuracy.1 In addition to aid in diagnosing tumors, dermoscopy has been increasingly used to identify inflammatory and infectious dermatoses, alopecia, and onicopathy.

Like other methods of morphological diagnosis, its feasibility, accuracy, and reproducibility are fundamental for it to be a useful resource. Therefore, the terminology to be used in the description of its diagnostic criteria is of particular relevance. The first International Consensus Conference on dermoscopy was held in 1989 in Hamburg.2 In 1993 the adopted dermoscopic terminology was first published in Brazilian Portuguese.3 The second International Consensus Conference was held online in 2003 in order to assess intra and interobserver concordance of several dermoscopic criteria and diagnostic algorithms exclusively for dermoscopy of pigmented skin lesions.4 In 2016 the third Consensus Report of the International Dermoscopy Society was published, whose objective was to establish a dictionary of standardized dermoscopic terms in English. In a survey, 1,093 professionals were asked whether they preferred metaphorical or descriptive terminology. Metaphorical language is the terminology traditionally applied that uses figurative language – such as leaf-like areas, streaks or veil – to describe structures. On the other hand, for descriptive terminology, the use of 60 terms to characterize the structures in their morphology was proposed based on five basic elements, referred to in English as lines, dots, clods, circles and pseudopods.5

The present study aimed to evaluate the reproducibility of the descriptive English terminology proposed at the third Consensus Report of the International Dermoscopy Society in Brazilian Portuguese.

Nine Brazilian dermatologists from different states (Rio de Janeiro, São Paulo, Paraná, Minas Gerais, and Rio Grande do Sul), with experience in dermoscopy independently evaluated the translation (suggested by the first author) into Brazilian Portuguese of the 60 descriptive and metaphorical terms proposed by the 3rd Consensus Conference of the International Dermoscopy Society (Charts 1 and 2). After that, a face-to-face meeting was held to discuss the choice of the suggested Portuguese terms and their clinical significance. In cases of three or more observers, Fleiss’ kappa concordance test (κ) was used to evaluate the interobserver agreement of the descriptive Portuguese terms. According to the values obtained, agreement was considered non-existent (> 0), weak (0 - 0.39), moderate (0.4 - 0.59), good (0.60 - 0.79) or optimal (> 0.8). The test has a null hypothesis where k = 0 when there is no agreement, and k > 0 when agreement is not random. The significance level for rejection of k = 0 was set at p < 0.05.

Interobserver agreement was considered weak (κ = 0.373, p < 0.05) regarding the acceptance of the descriptive terminology in Brazilian Portuguese. Additionally, a lack of concordance with terms related to clods criterion (κ = -0.02) and circles criterion (κ = -0.03) (Table 1) was observed.

In relation to the descriptive terminology used for the morphology and arrangement of vascular structures, the interobserver concordance was considered moderate (κ = 0.43, p < 0.05) (Table 2).

The descriptive terminology adopted by the English authors is based on five elements: lines, dots, clods, circles, and pseudopods. Paradoxically, two of these terms are metaphorical – clods and pseudopods – which makes the comparison between metaphorical and descriptive terminology incoherent. A metaphor, by definition, is a semantic resource for improving the expressiveness of what one wishes to express. Although most authors are quite familiar with metaphorical terms in dermoscopic terminology, one must be receptive to this new descriptive strategy should they become part of the professional environment. However, in our professional opinion, the five basic terms adopted in English are not adequate to be included in the Brazilian Portuguese terminology since their reproducibility was considered weak (κ = 0.373 p < 0.05).6

The typical appearance of a pigmented network is related both to its color and to the thickness of its lines. The distinction between broadened network and atypical network lies in the fact that atypical networks display an increased variability in color, with no thickness variation, whereas the broadened network, as the term suggests, displays a widening of the network lines. Both atypical and broadened networks are not unique to melanoma and can often be found in atypical melanocytic nevi. We observed a subtle variation of the terminology in Portuguese since three evaluators use the term rede pigmentar instead of rede pigmentada to refer to a pigmented network. Although inverse pigmented network is more frequently observed in melanoma and Spitz nevus, it can also be found in atypical and congenital nevi.

The fingerprint pattern is characteristic of solar lentigo. However, it can also be observed in seborrheic keratoses.

Maple leaflike structures, classically described in basal cell carcinoma, correspond to the intradermal masses of basaloid cells and do not display a linear aspect. They should not, therefore, be classified from the descriptive point of view as lines, but rather as masses or clods.

According to Japanese authors, the parallel ridge pattern is most frequently found in melanomas (99% in their series).7 According to the proposed algorithm, all plantar lesions exhibiting a parallel ridge pattern should be biopsied.8 Some exceptions are known such as paint pigmentation, Peutz-Jeughers syndrome, acral pigmentation due to the use of anticancer drugs, and pigmented plantar warts.8 Particularly in our milieu, racial melanosis is very frequent and may also exhibit a parallel ridge pattern.9,10

The literal translation of the English term “clod” – adopted by the consensus conference as being a descriptive term – would not make sense in Brazilian Portuguese. In our study, we suggested the terms massas, maciços, estruturas, aglomerados, and glóbulos (masses, lumps, structures, agglomerates, and globules, respectively). Therefore, the precise descriptive terms are not the best option to be included by in the Portuguese dermoscopic terminology. Thus, the metaphorical terms are more reproducible for this type of structure.

Although “gray dots” or “peppering” can be found in melanoma and benign lichenoid keratoses, they can also be observed in any condition with liquefaction of the basal layer and leakage in the dermis of melanin pigment.

The annular-granular pattern, despite being characteristic of lentiginous melanoma in situ, can also be identified in facial seborrheic and actinic keratoses.

Pseudopods found circumferentially in the classic “starburst” pattern of Reed nevus may be occasionally observed irregularly and asymmetrically in melanomas.

Centralized blotches (fried egg appearance) suggest dysplastic melanocytic nevi, and peripheral blotches suggest melanomas. In melanodermic patients, however, blotches can be observed in common melanocytic nevi at the center of the lesion.

Scarlike depigmentation may occasionally be observed in melanoma in areas where the controversial regression phenomenon is observed. However, this sign is not pathognomonic since it may also occur in areas of melanocytic or non-melanocytic traumatic lesions. Other names for this metaphorical term suggested by the authors were áreas de regressão, hipopigmentação tipo pseudocicatriz, and fibrose (regression zones, pseudo scar-like hypopigmentation, and fibrosis, respectively).

The structureless polychromatic zone, also referred to as a “rainbow pattern”, can be identified in several entities, but it is more significant in the diagnosis of Kaposi’s sarcoma.

The interobserver agreement with the descriptive terminology in Portuguese, when the vascular criteria were considered, was higher than when compared to the other criteria. In our study, it was classified as moderate (κ = 0.43; p < 0.05). “Looped” or “hairpin” vessels, in addition to being found in seborrheic keratoses and viral warts, may also be observed in keratoacanthomas and intradermal melanocytic nevi. When surrounded by a clear halo, they suggest the keratinocyte lineage of the tumor. Spiral or “corkscrew” helical vessels can also be found in cutaneous metastases of non-melanoma tumors.

Differently from the lack of consensus observed in the English language regarding the terms dermoscopy and dermatoscopy, in our environment there seems to be no disagreement as to the predilection of the well-established Portuguese term dermatoscopia, and not dermoscopia.5

The descriptive English terminology proposed by the 3rd Consensus Conference of the International Dermoscopy Society revealed weak reproducibility in Brazilian Portuguese and the descriptive dermatoscopic terminology related to morphology and arrangement of vascular structures showed moderate reproducibility. Despite small regional differences, metaphorical dermoscopic terminology seems to be the most useful and reproducible system to be adopted in Brazil.