Key points in dermoscopy for diagnosis of melanomas, including difficult to diagnose melanomas, on the trunk and extremities

Abstract

Early diagnosis and prompt surgical excision are the most important aims in the secondary prevention of cutaneous melanoma. Dermoscopy has increased the accuracy in the detection of melanoma because of dermoscopic-specific features that can be easily detected by trained dermoscopists. However, the classical melanoma-specific criteria such as multicomponent pattern, atypical pigmented network, irregular dots/globules, irregular streaks, multiple colors, blue-whitish veil or regression structures may not be present in all of these lesions. For some early melanomas change, as evidenced by sequential dermoscopic monitoring, may be the only feature suggesting malignancy. At present, even with dermoscopy, the diagnosis of these early melanomas remains to be a challenge for dermatologist. Patient education, digital dermoscopic follow up and consensus diagnosis have been proposed to overcome this problem.

Introduction

Several decades ago, the clinical diagnosis of melanoma was based on observable changes in a previously pigmented lesion such as bleeding, itching and ulceration, and the presence all of these at the time of diagnosis was associated with a worse prognosis. In the 1980s, when the “ABCD rule” was introduced, the early diagnosis of melanoma considerably improved.¹ This rule is based on the simple clinical and morphological features of melanoma: asymmetry, border irregularity, color variegation and a diameter of more than 5 mm. Later, a fifth criterion was added. It was called “E” for “evolution”, to describe the morphological changes of the lesion over time. Although this rule can be useful for large or thick melanomas, it has a low sensitivity for small lesions and very early melanoma.² Moreover, many unnecessary excisions of benign lesion may be carried out following the ABCD criterion.

The introduction of dermoscopy into the clinical practice of dermatology has helped to overcome those problems as it has disclosed a new and amazing morphological dimension of pigmented skin lesions. Dermoscopy is a non-invasive diagnostic technique for in vivo observation of the skin that uses optic magnification to permit the visualization of morphological structures that are not visible to the naked eye, thus opening a new dimension of the clinical morphological features of pigmented skin lesions.³ Its use is associated with both a significant increase of the sensitivity for melanoma diagnosis and a significant reduction of number of benign skin lesions unnecessarily excised.^4–6

Previous studies have demonstrated that dermoscopy improves accuracy in diagnosing pigmented skin lesions. A recent systematic review provided by Vestergaard et al.⁴ has reported that the relative diagnostic odds ratio for melanoma, for dermoscopy compared with naked eye examination, was 15.6 (P = 0.016). In this study, the average sensitivities for melanoma of the naked eye and dermoscopy examinations were 74% and 90%, respectively. These data are similar to other studies which estimated that dermoscopy allowed 10–27% higher sensibility than clinical diagnosis by the unaided eye.⁵ These results occurred without losing specificity, which means that better melanoma detection does not increase the number of unnecessary excision of benign tumors. On the contrary, the combination between the naked eye and dermoscopic examination determined a significant reduction in the percentage of patients referred for biopsy.^6,7

Why does dermoscopy allow better detection of melanoma? Argenziano et al.⁷ proposed three possible explanations: (i) with dermoscopy we can see early signs that become visible in melanoma much before clinical changes; (ii) clinicians are now more concerned with rigorous checking by dermoscopy of clinically banal-looking lesions, considering it as a first-level screening tool; and (iii) an improved attitude of clinicians to follow up their patients with digital dermoscopy.

Melanoma dermoscopic criteria

Although melanoma is a relatively uncommon tumor, moles, which are the main differential diagnosis, are extremely frequent. Distinguishing between benign melanocytic lesions and melanoma may be difficult. However, dermoscopy can help us in the management of these individual lesions.⁸ The first step when dealing with pigmented skin lesions should be differentiating between melanocytic and non-melanocytic lesions. There are some global features that are typical in melanocytic lesions such as pigmented network/pseudo-network, aggregated dots/globules, streaks, homogeneous blue pigmentation or parallel pattern. There are also some dermoscopic criterions for non-melanocytic skin lesions which can help us to rule out the melanoma diagnosis. Finally, the absence of any of these dermoscopic features leads us to think about a melanocytic lesion.

The most common global feature in melanocytic lesions is the reticular pattern, which is characterized by a pigment light-to-dark brown network with small symmetrical holes and thin network lines covering most parts of a given lesion.³ Histopathologically, the lines of the pigmented network more or less correspond to pigmented and elongated rete ridges and the small holes in the dermal papillae. A typical pigment network is a common finding in flat acquired melanocytic nevi, despite there being many variations on the theme of a typical pigmented network. Furthermore, the typical network is almost always present in lentigo simplex, solar lentigo and also in dermatofibroma. In the face, this network is called a “pseudo-network” due to the characteristic anatomy of the facial skin. In this location, the skin is devoid of rete ridges and there are many follicular infundibula located close to each other. With dermoscopy, it seems to be composed of round, equally-sized meshes corresponding to the pre-existing follicular ostia.

Although the typical network represents the hallmark of benign acquired melanocytic lesions, little changes can be appreciated in early melanomas. These changes lead to an atypical pigmented network characterized by a black, brown or gray network with irregular holes and thick lines irregularly distributed throughout the lesion and usually ending abruptly at the periphery. When the atypical pigmented network is present, the likelihood of melanoma is increased, with a calculated odds ratio of melanoma of 9.0 for an atypical pigmented network.⁸ However, this atypical network may be also present in benign melanocytic skin lesions, mainly Clark’s nevi (atypical melanocytic nevi). In fact, the differential diagnosis between these benign nevi with irregular pigmented network and early melanoma is difficult and still a challenge for dermatologists.

Another dermoscopic criterion of melanocytic skin lesions is the presence of aggregated dots/globules.³ Dots and globules are sharply circumscribed, usually round or oval, variously sized, and more or less clustered with various shades of brown and gray-black. They correlate with aggregations of pigmented melanocytes, melanophages or clumps of melanin within the cornified layer, the epidermis, the dermoepidermal junction or the papillary dermis. The color of the globules depends on the depth of these pigmented aggregates in the skin. In benign melanocytic lesions, dots and globules are regular in size and shape, and are evenly distributed. In melanoma, irregular dots and globules are usually located at the periphery, are of different size and shape and are asymmetrically distributed.

The architectural distribution of the streaks is more important than the morphology of a single streak in the management of pigmented skin lesions.³ Streaks are bulbous and often kinked or finger-like projections seen at the edge of a lesion. They are linear structures that may be observed throughout a lesion, but are more apparent in the periphery. They range in color from light brown to black, and can also be called radial streaming, radial streaks and pseudo-pods. They correspond to pigmented junctional nests of melanocytes that are disposed like tubules to be parallel to the skin surface. The presence of irregular streaks is strongly associated with melanoma, particularly when they are unevenly distributed in a pigmented skin lesion. However, regular and symmetrical streaks are typical of some benign lesions such as Spitz or Reed nevi.

The blue-whitish veil can be one of the key features for the diagnosis of melanoma. It is a confluent, irregular, structureless area of whitish-blue diffuse pigmentation associated with pigmented network, dots/globules and/or streaks. It is usually found in a clinically elevated part of the lesion and corresponds with nests of heavily pigmented melanocytes in the upper dermis which are under the acanthotic epidermis. Although it can be present in Spitz/Reed nevi, the presence of a blue-whitish veil is relatively frequent in melanoma. Furthermore, it is not present in Clark nevi.³

Regression structures are also a helpful clue for distinguishing melanoma from other benign lesions. These structures are often located in the flat part of the lesion and they appear as white scar-like areas, blue areas or a combination of both. The blue areas, also called “peppering”, are zones with small multiple blue-gray dots that correspond to a variable number of melanophages in the papillary dermis. Fibrosis and melanosis are usually found together during regression so the combinations of white and blue areas are commonly noted in the same lesion.

With dermoscopy, different vascular structures can be seen when evaluating pigmented skin lesions. The presence of irregular hairpin vessels, dotted vessels, linear irregular vessels or vessel within regression structures in a given lesion have been associated with melanoma. On the contrary, comma vessels are rare in melanoma and are mainly present in benign melanocytic lesions.⁸

For a better characterization of Clark’s nevi, a dermoscopic classification of these nevi has been previously reported.⁹ In this classification, three main dermoscopic patterns (reticular, globular and homogeneous) and two descriptors of pigmentation (hypo- or hyperpigmentation) have to be analyzed. There are also some dermoscopic criteria for the diagnosis of melanoma.^8,10,11 These criteria, also called “melanoma-specific criteria”, differentiate between both early melanoma/melanoma in situ and intermediate/thick melanoma (Figs 1,2). The dermoscopic appearance of these last is independent of the location of the lesion because the preformed anatomical structures are destroyed by thick melanomas. The melanoma-specific criteria are summarized in Tables 1 and 2.¹²

Figure 1.  Early invasive melanoma (Breslow index, 0.4 mm) exhibiting a dermoscopically atypical pigmented network in the right part of the lesion.

Figure 2.  Intermediate melanoma (Breslow index, 1.3 mm) showing irregular pigmentation, blue-whitish veil and irregular dots/globules.

Table 1.   Melanoma-specific criteria

Table 2.   Frequency of melanoma-specific criteria in thin and thick melanomas

Criteria	Thin melanoma: melanoma in situ and early invasive melanoma (Breslow index, <0.76 mm) (%)	Intermediate and thick melanoma (Breslow index, >0.75 mm) (%)
Atypical pigment network	>70	<30
Irregular dots/globules	50–70	50–70
Irregular streaks	50–70	50–70
Irregular pigmentation	50–70	50–70
Regression structures	30–50	<30
Blue-whitish veil	<30	>70
Irregular vascular pattern	<10	30–50

Difficult to diagnose melanomas

However, dermoscopy alone may not be sufficient to diagnose all early melanomas because some of them present an uncharacteristic dermoscopic appearance.^11,13–15 These “featureless melanoma” may lack those melanoma-specific criteria and may even appear as benign melanocytic lesions or atypical nevi. The traditional criteria correlating with those of melanoma such as multi-component pattern, atypical pigmented network, irregular dots/globules, irregular streaks, multiple colors, blue-whitish veil or regression structures may not be present. In addition, these melanomas may lack all of the features of the ABCD rule. Difficult to diagnose melanomas (DDM) have been reported at 5–10%. Nodular melanomas may just simulate benign tumors such as seborrheic keratosis, dermatofibroma or pyogenic granuloma. Some amelanotic melanoma and hypopigmented melanoma can be diagnosed only because of their irregular vascular structures. In many cases, the only clue to the diagnosis of some melanomas may be the history of change reported by the patients.

In a recent study, the characteristics of DDM were analyzed.¹³ In this study, the prevalence of DDM in women was significantly higher than male prevalence. The risk of DDM was increased with the reduction of the Breslow depth, with an odds ratio of 10.12 for melanoma of 1 mm or less and 30.82 for melanoma in situ. Most of the cases of DDM were early melanoma, and almost half of DDM were melanoma in situ. A previous study reported similar data of higher prevalence of DDM in early and thin melanomas.¹⁵ The absence or presence of regular pigmentation, the absence of a blue-whitish veil, the absence of regression structures and the presence of hypopigmentation were associated with increased risk of DDM. On the contrary, the melanoma-specific criteria were absent in these lesions.

The histopathological diagnosis of melanoma may also be complex and difficult, even more with DDM. In this sense, a diagnostic discrepancy among formally trained dermoscopists has been reported to be a predictor for diagnostic disagreement among histopathologists.¹⁶

Discussion

Therefore, how can we improve the diagnosis of early melanoma? One of the strategies may be patient education.¹¹ Patient concern about melanoma and risk of changing lesions will require earlier medical intervention. In fact, one of the main reasons for biopsy in DDM is a history of change, noted either by the patient or during digital dermoscopic follow up.

Short-term surveillance with dermoscopy and total body photographic images is another approach to detect minimal changes in the first stages of melanoma development.^11,13–15 Some melanomas can be diagnosed neither with the naked eye nor dermoscopically. These early and featureless melanomas can be recognized only by observing their dynamic evolution over time. The importance of this method for the early diagnosis of melanoma has been recently reported in a study in which 63 melanomas that were first indistinguishable from melanocytic nevi were only detected because of changes during digital dermoscopic follow up.¹⁵

For some, early melanoma change may be the only feature suggesting malignancy. Only when these lesions grow, some typical dermoscopic features associated with malignancy may become visible. Obviously benign lesions can also grow, but they often enlarge symmetrically. Melanoma, because of the loss of regulatory control, grows with an asymmetrical pattern.¹³ This irregular growing distinguishes melanoma from atypical nevi. The fact is that although many melanomas may be identified by using the described melanoma-specific criteria, subtle changes that lead to an equivocal dermoscopic appearance is the only reason for biopsy in other cases.^11,13–15 In these lesions, the architectural distribution of dermoscopic features is of the uppermost importance.

In a recent study, a diagnostic algorithm has been proposed to not miss DDM.¹⁴ In this algorithm, digital follow up is recommended when a lesion shows melanocytic pattern and the patient have multiple nevi. This management is not recommended in nodular lesions and also excision is recommended when a single atypical lesion is present as well as when any Spitzoid lesion is present in adults. Some conditions like recent changes in a melanocytic lesion, lesions with polymorphous vessels, gray to blue color in a lentigo/seborrheic keratosis, regression structures or peripheral blotches, and pyogenic granuloma/Spitz nevi-like lesions should increase our index of suspicion for melanoma. In all these cases, to make a dermoscopic diagnosis of melanoma may be less important than to decide to undergo a biopsy procedure.¹⁷

Finally, the diagnostic performance of dermoscopy improves when the diagnosis is made by a group of examiners in consensus.¹⁸ Nowadays, this is possible by electronic transmission of digital dermoscopic images. It has been reported that a consensus diagnosis increases the sensitivity of melanoma detection when it is made by a group of experts.⁸ Therefore, looking for a second opinion may be another helpful approach when dealing with these DDM.