MATERIAL AND METHODS

Selection of cases

We screened the files from the division of dermatopathology at the New York Presbyterian Hospital, USA, in 1998, and the department of dermatology and allergology at the Technical University in Munich, Germany, between 1999 and 2001, for acquired melanocytic lesions measuring clinically ≤ 4mm in diameter. Initially excluded from the study were naevi in acral or genital locations. Also excluded after microscopic evaluation were blue naevi, Spitz and Reed naevi, naevi with congenital features, traumatised naevi, and naevi with incomplete biopsies precluding adequate evaluation of size. Two hundred and sixty one small diameter melanocytic lesions obtained from 206 patients remained for further analysis. Each specimen was routinely fixed in 3.7% buffered formalin, embedded in paraffin wax, sectioned at multiple levels, and stained with haematoxylin and eosin. The most “atypical” level was subjected to analysis. The diameter of ≤ 4 mm was reconfirmed microscopically in each case.

Definition of architectural and cytological features

The histopathological criteria of naevi with architectural disorder and the grading of cytological atypia were based on widely accepted, standard criteria and were adapted with minor changes.^8,14,18–22

Definition of a naevus with architectural disorder

According to Clemente et al, the diagnosis of an atypical naevus requires two of two major criteria and two of four minor criteria to be satisfied.¹⁸ Major criteria are: (1) basilar proliferation of atypical melanocytes, exhibiting either a lentiginous or a nested pattern, and (2) (when there is a dermal component) the basilar proliferation by at least three rete ridges beyond the dermal component, thus exhibiting a so called shoulder. Minor criteria are: (1) the presence of concentric eosinophilic fibrosis or lamellar fibroplasia, (2) neovascularisation, (3) bridging of adjacent rete ridges, and (4) a dermal cellular inflammatory response.

Additional architectural features

Apart from these, we looked for the presence or absence of symmetry, regression, thickening of the stratum corneum, circumscription (bilaterally not sharp or sharp on one side), and pagetoid spread of melanocytic cells in upper levels of the epidermis. The rete ridge pattern was evaluated with regard to an increase in number and elongation of rete ridges, distortion, and regularity versus effacement (flattening of at least one rete ridge, resulting in an increased length of a suprapapillary plate containing numerous single melanocytes). The orientation of most of the nests was assessed (round/oval, horizontal, vertical, horizontal/vertical).

Cytological features

The shape of the melanocytic cells was judged according to whether they were predominantly round cells, spindle cells, or equal proportions of both. In the same way, the proportion of solitary melanocytes to melanocytes in nests (defined as a group of three or more cells) was roughly estimated. The location of melanocytes at the rete ridges was also recorded: whether they were at the tips and lateral aspects of the rete ridges or also involved the suprapapillary plates.

Grading of cytological atypia

Although grading of nuclear atypia is difficult and may only be poorly reproducible, we tried to adapt for small naevi the grading scheme used by the groups of Rhodes and Weinstock.^21,22

Table 1 summarises the definitions of cellular morphology. The designation of slight, moderate, or severe atypia indicates the highest degree of atypia that applies to multiple melanocytes over at least two rete ridges at the epidermodermal junction, regardless of the manner in which they are grouped. No grading of dermal melanocytes was performed. The diagnosis of melanoma required a continuous proliferation of moderately to severely atypical melanocytes in a severely disordered pattern for a breadth of at least two suprapapillary plates connected by a flattened rete ridge.

Statistical analysis

All variables were analysed using χ² for trend, with significance set at p < 0.05.

RESULTS

A total of 261 acquired melanocytic lesions obtained from 206 patients (125 female and 81 male patients; mean age, 41.6 years; range, 13–79) met the inclusion criteria. Forty three per cent of all lesions were junctional and 57% compound. The cases were first classified into three groups: benign melanocytic naevi without architectural disorder, benign naevi with architectural disorder, and melanoma according to the criteria of Clemente et al.¹⁸ Almost all lesions exhibited a lentiginous proliferation of melanocytes as individual cells, or nested or mixed patterns. Because a high percentage of the ordinary benign melanocytic naevi showed some bridging of rete ridges, in addition to slight inflammation, meeting two of the minor criteria for atypia, we found it more important to use distortion of rete ridges (instead of bridging) as a differentiating criterion of architectural disorder. With this change, 73 lesions (28%) were benign and 188 (72%) showed architectural disorder, and we designated these atypical naevi. Only 34% of the benign naevi had a dermal component. In contrast, 66% of the atypical naevi were compound, all of them showing a shoulder phenomenon. According to their cytological changes, the atypical naevi were graded mild in 103 cases (55%), moderate in 73 (39%), and severe in 12 cases (6%) (fig 1). There was no recognisable variability in age and sex distribution within the different groups (data not shown). A slight female preponderance for severely atypical melanocytic naevi was not significant and could be explained by the low number of cases within this group. Table 2 shows the anatomical site distribution. Table 3 shows the frequencies of the most relevant criteria. Other criteria such as regression and hyperkeratosis observed in around 10% and 26% of all cases, respectively, did not show pronounced differences between the groups.

Some criteria were found to have a slightly increasing or decreasing frequency from benign to severe atypical naevi, but were not particularly useful for grading. For instance, most of the benign (90%), mild (77%), and moderate (62%) atypical lesions were symmetrical in contrast to the severe atypical lesions, which were asymmetrical in 54% of the cases. Ill defined borders were seen in 92% of benign, 89% of mild, 73% of moderate, and only 67% of severely atypical naevi. The orientation of nests was mostly round/oval, and the percentage of cases with horizontally oriented nests increased from 9% in benign, to 19% in mild, 38% in moderate, and 43% in severe atypical naevi. The cell shape of melanocytes was predominately round. A mixed pattern of round and spindle-type cells was seen in 18% of benign naevi, 27% of mildly atypical naevi, 48% of moderately atypical naevi, and 58% of severely atypical naevi. The naevus cells were preferentially located at the tips of the rete ridges in 88% of the benign and 82% of the mild cases. In addition, 48% of moderate and 75% of severe cases exhibited melanocytes in the suprapapillary plate regions. Some criteria, such as effacement of rete ridges or pagetoid spread, were mainly present in the centre of lesions in the severe group (50% and 67%, respectively) and almost non-existent in the others (benign, 0%; mild, 4% and 2%, respectively; moderate, 10% and 5%, respectively).

DISCUSSION

Our study focused on the histological appearance of small melanocytic lesions. Using widely accepted criteria for diagnosing and grading naevi with architectural disorder and cytological atypia (also known as “dysplastic naevi”), we found that small melanocytic naevi have the same complex spectrum as larger lesions, ranging from benign naevi to severely atypical naevi to malignant melanoma. An overall percentage of around 72% of the excised small naevi presenting with some atypical histological features is consistent with data established on naevi of any dimension.^15,17

For practical purposes, we found useful the following stepwise manner for classifying and grading. First, the distinction between ordinary benign and atypical naevi was made on architectural features. Most of the atypical naevi showed distortion of rete ridges and a cellular lymphoid immune response, less often fibrosis, and only rarely neovascularisation. Grading of atypia was mainly based on cytological criteria. A naevus cell nucleus more than twice the size of a keratinocyte nucleus was a reliable marker to distinguish between moderately and mildly atypical naevi. The visibility of nucleoli was also an easily recognisable feature seen in naevus nuclei with at least moderate atypia. Because the measurement of the chromatin pattern, nuclear shape, and contours was too difficult and time consuming, the overall visual assessment of nuclear variability turned out to be a fair method for differentiating between mild and moderate/severe atypia, although no significance difference could be reached between moderately and severely atypical naevi. Whereas in most mildly atypical naevi melanocytes were found in nests at the tips and lateral aspects of the rete ridges, moderately atypical naevi had increasingly more melanocytes as single units in between rete ridges. In severely atypical naevi and in melanoma in situ this increase of single melanocytes at the suprapapillary plate region merged into effacement of rete ridges and an irregularly spaced rete pattern. Pagetoid spread of single melanocytes into upper parts of the epidermis was seen focally in the centre of lesions in the severe group, thus allowing differentiation from moderate atypical naevi.

It was noteworthy that no single histological feature studied here was absolutely diagnostic by itself for a clear cut distinction between the different groups. It was always the combination of features, rather than a single parameter that allowed classification of small melanocytic lesions.

Grading of melanocytic lesions itself can be problematical. Several analyses of grading atypical melanocytic naevi larger than 5 mm in diameter have reported only a low interobserver concordance and poor reproducibility, even when simple scoring systems were used.^23–26 However, the clinical value of grading atypical naevi is not the subject of this study, grading only showed that melanocytic lesions do not automatically fit into the category of common benign naevi if they are small in size.

“In our opinion, if histological criteria have been formulated to define atypical naevi, then they should be used regardless of the size of the lesion”

The histopathological spectrum of small melanocytic lesions encompasses lesions that represent clearly benign naevi at one extreme and malignant melanoma at the other. In between is a large and heterogeneous group of small naevi exhibiting intermediate characteristics that fulfill the architectural and cytological abnormalities of atypical naevi. Several authors recommend that the diagnosis of an atypical naevus should be given only to lesions larger than 5 mm in size.^19,20 This restrictive approach in diagnosing atypical naevi is advised probably because of a better correlation between large atypical naevi and a personal or family history of melanoma, and because of the higher predictive value as a marker for the development of melanoma. Several studies in the past have indeed verified that clinical atypical naevi larger than 5 mm in diameter are markers of an increased risk of developing melanoma.^1,4,5 Numerous atypical naevi might be more predictive, but the number of small diameter naevi has also been found to be associated with an increased risk for melanoma. Recent studies indicated that the presence of more than 120 naevi less than 5 mm in diameter was associated with a relative risk for the development of melanoma of 19.6 compared with patients having fewer than 10 naevi.³ Tucker et al predicted a twofold higher risk for patients with more than 25 small naevi.²⁷ That study also found a significant 2.6-fold risk associated with naevi smaller than 5 mm that had the clinical appearance of atypical naevi, but did not meet the size criterion. All these studies exclusively judged the relative risk by counting the number of naevi on the whole body and subdividing them into common benign or atypical naevi according to their clinical appearance. A histological examination for defining atypical naevi has not been included and no study has directly determined the risk associated with histological atypia as a specific marker for melanoma development in comparison with clinical atypia. Annessi et al recently studied the correlation between clinical atypia and histological dysplasia in acquired melanocytic naevi and found an overall disagreement of 39.8%, in part reciprocally related to size.¹⁷ In particular, 69.6% of clinically non-atypical naevi were found to be atypical on histiological examination and 33.3% of naevi judged histologically to be not atypical were diagnosed clinically as atypical. Others have also demonstrated histological features of atypia in clinically common naevi.¹⁶ Moreover, Piepkorn et al estimated the prevalence of atypical architectural features in naevi to be as high as 53% in the general population, regardless the size of the lesions and without cytological atypia as a required criterion.²⁸ This poor correlation between clinical phenotype and histological appearance tends to question the concept of atypical naevus as a real entity and points towards the need for more exact definitions of both clinical and histological atypia.

The present study confirms a high proportion of histological atypical features in small naevi. Even though our study presents consecutive cases collected over several years, a percentage of 72% atypical naevi may reflect a certain bias because patients who have had small naevi removed may more often have a positive personal or family history of melanoma. If this is the case, small lesions may indeed reflect an increased risk to develop melanoma. Alternatively, it has been argued that atypical features seen in small naevi do not indicate persistent atypia, but may be related to an early evolving state of a lesion that is actively proliferating.²⁸ Because most of the small naevi in our study were excised because of changes over time or because they conformed to the “ugly duckling sign” (naevi that do not fit into the common profile of most naevi in a given patient),²⁹ we expected to see some signs of proliferation. The clinical relevance of these lesions as precursors or markers for the risk of malignant melanoma cannot be estimated without appropriate studies, including their histological appearance. In our opinion, if histological criteria have been formulated to define atypical naevi, then they should be used regardless of the size of the lesion. Therefore, we assign small naevi exhibiting atypical features on histological examination to the category of naevi with architectural disorder and mild, moderate, or severe cytological atypia of melanocytes.