Congenital naevocellular naevi (CNN) have classical histological criteria t
hat are thought to allow distinction from acquired naevi, These criteria ar
e mainly the horizontal distribution of melanocytes, forming 'Indian files'
, and a prominent adnexotropism. In order to check whether the previously d
escribed criteria were reliable, we analysed 1349 unselected consecutive ca
ses of naevocellular naevi excised in children under 16 years, during a 54-
month period. The slides were analysed in order to determine by histologica
l analysis only if they could be classified as CNN. These results were then
compared with the clinical files, in which only the most reliable data fro
m parental and/or medical observations were included. Of the 1349 naevi, 65
9 had the typical histological criteria of CNN, 32 of them being deep CNN,
characterized by massive involvement of the lower dermis and hypodermis. Th
e comparison with clinical data showed that 32 naevi with the histological
criteria of congenital naevi were actually acquired, and that 179 naevi pre
sent at birth did not fulfil these criteria. This study shows that the clas
sic histological criteria are not absolutely specific and are poorly sensit
ive as 36% of naevi present at birth lacked the classic criteria. The most
specific criteria of true CNN were the in involvement of eccrine glands and
presence of melanocytes in the septae. In the case of deep CNN which corre
sponded to large or very large naevi, the clinicopathological correlation w
as 100%. Deep CNN could easily be distinguished from superficial CNN and of
ten exhibited many histological changes consistent with a complex hamartoma
, such as presence of brown fat tissue, abnormal vessels and numerous large
terminal follicles, In conclusion, our study suggests that it is not possi
ble to predict, by histological analysis alone, that a naevus was present a
t birth, except in deep CNN which are likely to be a separate entity among
all congenital naevi. Studies dealing with congenital naevus-associated mel
anoma should take into account the lack of sensitivity of these criteria.