ARTICLE
Epiluminescence microscopy, or dermoscopy, is a non invasive technique
used to observe in vivo the pattern of the intra- and extracellular
pigment in pigmented lesions. Under favorable conditions, it also allows
the identification of the intra- and extravasal blood pigment, the less
pigmented the lesion, the easier being such identification. These morphological
aspects are the expression of peculiar histopathological architectural
features. A correct interpretation of such features is therefore quite
a useful diagnostic tool for the clinician.
Dermoscopy is mainly used in the study of pigmentary lesions for the
early detection of melanoma. It is a fact that when diagnosed at an early
stage melanoma can be cured by surgery alone and that, on the other hand,
at a later stage both surgery and the adjuvant medical treatment are unable
to stop the fatal progression of this highly malignant tumor.
The role of dermoscopy in improving the clinical diagnosis of melanoma
is now recognized by all experts in the field. It is therefore important
to spread the use of this diagnostic tool in the fight against melanoma
and in this context some of us have felt the need to share their experience
and to offer a "key", making simpler and quicker the interpretation of
dermoscopy.
For this reason we have developed, using our experience and case records,
a method that combines good diagnostic sensitivity with simplicity of
use.
Materials and methods
This study concerns pigmented skin lesions observed by us between January
1992 and June 1997.
Each lesion was photographed clinically and dermoscopically. The dermoscopic
picture was obtained by means of a Dermaphot Heine at 10X, using a Kodak
Ektachrome 64 ASA film. The lesions were then excised and submitted to
histopathological examination.
Two kinds of analyses were carried out in sequence, namely a training
set to develop the diagnostic method and a test set to evaluate it.
Training set
In order to better evaluate the "diagnostic weight" of each dermoscopic
feature, as defined by the Consensus Conference held in Hamburg in 1989
[1] with regard to melanoma diagnosis, we checked their statistical significance,
sensitivity and specificity in our patient series (Table
I).
For that purpose we recorded all the dermoscopic features found in 218
pigmented lesions (training set) classified as follows:
45 melanomas, max diameter < 1 cm, 19 of which in situ and
26 < 0.75 mm thick (mean 0.36 mm, median 0.36 mm, min 0.10 mm, max
0.72 mm); 38 epithelioid and/or spindle cell nevi; 45 melanocytic nevi
undergoing regression; 45 mainly junctional melanocytic nevi; 45 mainly
dermal melanocytic nevi.
The different features of the pigment network defined by the Consensus
Conference were grouped as regular and irregular networks.
The regular network has thin lines, a close mesh net and is uniform
throughout the lesion. The irregular network is thick, has a wide mesh
net and shows varying features in a same lesions (Fig. 1).
To the classic dermoscopic features we have added two new ones, regression-erythema
and unhomogeneity. The term "regression-erythema" defines the disappearance
of dermoscopic features in a given area of the lesion, while diffuse erythema,
possibly with a few angiectases, is observed (Fig.
2).
"Unhomogeneity" is an asymmetrical or irregular distribution in the
lesion of at least two dermoscopic features not necessarily predictive
of malignancy (Fig. 3).
The sharp margin is regarded as such when it involves at least one fourth
of the lesion (Fig. 4).
Pseudopods are considered predictive of malignancy when they display
an irregular distribution (Fig.
5): in fact, epithelioid and/or spindle cell nevi usually present
pseudopods regularly distributed [2].
All the lesions were evaluated by three different observers according
to the above features. The observers were unaware of the clinical characteristics
of the lesions as well as of the patients' personal and case histories.
The dermoscopic features were evaluated as present or absent. In the
case of disagreement among the observers, the majority view prevailed.
The sensitivity and specificity of each feature for the diagnosis of
melanoma were evaluated.
The sensitivity of a dermoscopic feature is the proportion of all cases
of histologically proved melanomas that dermoscopically show the feature.
The specificity of a dermoscopic feature is the proportion of all cases
histologically proved not to be melanomas that dermoscopically do not
show the feature. Therefore a sensitivity of 8.8% for regular pigment
network means that 8.8% of melanomas exhibit the feature, while a specificity
of 61.8% for regular pigment network means that 61.8% of non melanomas
do not exhibit the feature.
The statistical significance of the malignancy predictive power of each
dermoscopic feature was measured by the chi square test with Yates' correction.
The Fischer's exact test was also applied for the dermoscopic features
showing an expected value < 6 in the 2 x 2 tables.
With regard to the diagnostic technique, the dermoscopic features were
chosen according to their reproducibility by the different observers and
by their relationships with histopathological criteria predictive of malignancy.
Seven dermoscopic features useful for the diagnosis of melanoma were
thus detected and used to develop a diagnostic method.
Test set
In the second part of the study, the features selected according to
the above mentioned criteria were used to develop a diagnostic test for
melanoma. With this method we evaluated 713 pigmented lesions consecutively
observed by us and regarded as being melanocytic in nature, following
the dermoscopic algorithm used to distinguish melanocytic from non-melanocytic
lesions.
The algorithm is as follows: the lesions showing network or globules
are regarded as melanocytic. The presence of horny pseudocysts and comedo-like
openings, without pigment network or globules, suggests seborrheic keratosis.
Maple leaf-like areas at the periphery suggest basal cell carcinoma, a
homogeneous blue coloring points to blue nevus, while red-blue areas are
typical of angioma and angiokeratoma.
When these features are absent, the lesion is regarded as melanocytic.
At the subsequent histopathological examination the lesions were classified
as follows:
Junctional melanocytic nevi: 92.
Mainly junctional compound melanocytic nevi: 37.
Compound melanocytic nevi: 224.
Congenital melanocytic nevi: 20.
Melanocytic nevi showing regression and inflammatory infiltrate: 102.
Combined melanocytic nevi: 8.
Epithelioid and/or spindle cell nevi: 53.
Lentigo simplex: 3.
Black reticulated solar lentigo: 1.
Seborrheic keratoses: 3.
Melanoacanthoma: 1.
Basal cell carcinoma: 1.
Melanomas: 168, of which: 29 in situ, 87, < 0.75 mm thick,
32, 0.76-1.5 mm thick, 17, 1.5-4 mm thick, 3, > 4 mm thick (mean 0.92
mm, median 0.52 mm, max 4.5 mm, min 0.10 mm).
The sensitivity, specificity, predictive values positive and negative,
and efficiency of the diagnostic test were calculated.
The sensitivity of the diagnostic method is the proportion of all cases
of histologically proved melanomas that were dermoscopically diagnosed
as melanomas. The specificity of the diagnostic method is the proportion
of all cases histologically proved not to be melanomas that were dermoscopically
diagnosed as not melanomas. The predictive value positive is the proportion
of all cases dermoscopically diagnosed as melanomas that were histologically
proved to be melanomas. The predictive value negative is the proportion
of all the cases dermoscopically diagnosed not to be melanomas that were
histologically proved not to be melanomas. The efficiency of the diagnostic
method is the proportion of all cases (melanomas and non-melanomas) dermoscopically
and histologically correctly diagnosed.
Results
Training set
In the first portion of the analysis the dermoscopic features, evaluated
in 218 lesions, showed the sensitivity, specificity and statistical significance
values reported in Table I.
The dermoscopic features that showed small expected frequencies (i.e.
< 6) in the 2 x 2 tables were also submitted to the Fisher's exact
test. The P value results do not differ significantly from those obtained
with the chi square test with Yates' correction. They are shown in Table
II.
Based on these results, we regarded as useful features for the diagnosis
of melanoma those with specificity > 80%, sensitivity > 5% and P
< 0.05.
The pigment elimination and globules features were not included since
they are poorly reproducible with two-dimension vision, the distinction
between pigment elimination and strongly pigmented thecae being sometimes
unclear.
According to such selection, the features regarded as essential for
the diagnosis of melanoma are those reported in Table
III and Fig. 1
to 7.
We further divided the malignancy predictive features into major (specificity
> 95%) and minor (specificity ¾ 95% and > 80%).
Following such selection we attributed a score 2 to the major features
and a score 1 to the minor features: the major features are: regression
erythema, radial streaming, gray-blue veil, irregularly distributed pseudopods;
the minor features are unhomogeneity, irregular pigment network, sharp
margin (see Table IV for
description and score). The lesions where the sum of the features gave
a score > or = 2 were diagnosed as being malignant, therefore to make
a diagnosis of melanoma, the presence of one major feature or the concurrent
presence of two minor features was regarded as sufficient.
We called this method the Seven Features for Melanoma (7FFM).
Test set
Using the diagnostic method developed with the training set, three different
observers, unaware of the clinical and personal characteristics of the
patients, recognized 159 out of 168 Melanomas and 466 out of 545 benign
lesions. The results of this study are reported in Table
V.
At histopathological examination, the 9 melanomas which proved to be
false negatives had a maximum thickness of 0.44 mm. The 79 false positives
were: 28 melanocytic nevi with regression and/or inflammatory infiltrate
features; 21 epithelioid and/or spindle cell nevi; 10 junctional melanocytic
nevi; 15 compound melanocytic nevi; 2 lentigoes; 1 melanoacanthoma; 1
combined nevus and 1 seborrheic keratosis.
Discussion
According to some statistics, the clinical diagnosis of melanoma has
a mean sensitivity of 67% (range 48% to 81% according both to the physician's
ability, dermatologists having better results than general practitioners,
and to the type of melanoma) [3]. Other authors before us have investigated
the extent to which dermoscopy is able to improve such percentages, based
on the degree of sensitivity and specificity of the different dermoscopic
features. They have, among other things, pointed out that none of the
features reaches a specificity of 100% and have made several suggestions
without actually developing a true diagnostic method [4-9].
Nachbar and Stolz in 1994 [10-12] developed a diagnostic method, the
ABCD rule of dermoscopy, that had a specificity of 92.8% and a sensitivity
of 91.2% and they were the first to demonstrate the usefulness of the
dermoscope, a tool suitable for use (unlike the big stereomicroscopes)
in out-patient facilities, in the screening of pigmented skin lesions.
Following such attempts and being both aware of the importance of this
diagnostic tool and convinced of its potential ability to improve the
clinical diagnosis of pigmented lesions, we took advantage of our experience,
dating back to the end of the Eighties, to develop a method based only
on the features with a malignancy predictive value.
In fact, in contrast to the method of Nachbar and Stoltz, all the dermoscopic
features selected in our method have a histopathological correlation with
malignancy.
This characteristic was shared also by the two new dermoscopic features
detected during our practical experience.
The regression-erythema feature corresponds to regression associated
with vasodilatation and neoangiogenesis, frequently found in melanoma
[13-15] while unhomogeneity corresponds to the histopathological architectural
disorder of the lesion.
Other authors in previous studies have proposed the concept of unhomogeneity
but in our opinion they did not define it in a way which is easy to understand
and reproduce.
Nilles et al. [5] considered an asymmetrical pigment distribution
(no relation with dermoscopic features) with four different grades of
severity. Kenet et al. [16] described a multicompetent pattern
which consists of three or more discrete regions with different ELM appearances,
including a darkly pigmented region with a broadened network. The multicompetent
pattern as described appears very different from unhomogeneity and a higher
magnification than that obtained with a dermatoscope is probably necessary
to detect it.
Regression-erythema, which in the first part of our study showed a specificity
of 100%, in the second half of the study was found in 6 out of 545 benign
lesions (specificity = 98.9%).
Therefore, as observed by other authors, our data confirm that none
of the dermoscopic features is 100% specific.
The fact that a dermoscopic feature has a histopathological correlation
with malignancy does not mean that the feature is present only in melanoma.
Dysplastic lesions such as atypical nevi, melanocytic nevi showing regression
and inflammatory infiltrate and epithelioid and/or spindle cell nevi can
present unhomogeneity and radial streaming, as shown by other authors
[17].
Our study confirms these data: most of the false positives obtained
with our method are melanocytic nevi showing regression and inflammatory
infiltrate and epithelioid and or spindle cell nevi.
A sharp margin can be observed in ink-spot lentigo and in benign palmar
or plantar lesions [17, 18].
Not even the features used to distinguish between melanocytic and non
melanocytic pigmented skin lesions are 100% specific: in fact in our study
3 seborrheic keratoses, 1 melanoacanthoma and 1 pigmented basal cell carcinoma
were regarded as being melanocytic. In a previous study we described a
melanoma showing horny pseudocysts and comedo-like openings [19].
In spite of these limitations, dermoscopy improves both the sensitivity
and specificity in the diagnosis of melanoma, as shown in the present
and in previous studies [4-12].
Our belief is that, in order to be used in clinical practice, a diagnostic
method must be easy to understand, to handle and to reproduce. This led
us to reduce to seven the features to be employed in dermoscopic diagnosis,
on the basis of the highest specificity (> 80%).
We have called this method the seven features for melanoma (7FFM).
The results of our efforts have been presented in an Atlas of Dermoscopy
first published in Italian in 1996 and in English in 1997 [20-21].
These data suggest that dermoscopy can be widely used in the out-patient
diagnosis of pigmented skin lesions.
REFERENCES
1. Bahmer FA, Fritsch P, Kreusch J, Pehamberger H, Rohrer C, Schindera
I. Terminology in surface microscopy. J Am Acad Dermatol 1990;
23: 1159-62.
2. Steiner A, Pahemberger H, Binder M, Wolff K. Pigmented spitz nevi:
improvement of the diagnostic accuracy by epiluminescence microscopy.
J Am Acad Dermatol 1992; 27: 697-701.
3. Piepkorn M, Odland PB. Quality of care in the diagnosis of melanoma
and related melanocytic lesions. Arch Dermatol 1997; 133:1393-6.
4. Steiner A, Binder M, Schemper M, Wolff K, Pahemberger H. Statistical
evaluation of epiluminescence microscopy criteria for melanocytic pigmented
skin lesions. J Am Acad Dermatol 1993; 29: 581-8.
5. Nilles M, Boedeker RH, Schill WB. Surface microscopy of naevi and
melanomas-clues to melanoma. Br J Dermatol 1994; 130: 349-55.
6. Soyer HP, Smolle J, Leitinger G, Rieger E, Kerl H. Diagnostic reliability
of dermoscopic criteria for detecting malignant melanoma. Dermatology
1995; 190: 25-30.
7. Menzies SW, Crotty KA, Ingvar C, et al. An atlas of surface
microscopy of pigmented skin lesions. Sidney, Australia: McGraw-Hill;
1996.
8. Menzies SW, Ingvar C, McCarty WH. A sensitivity and specificity analysis
of the surface microscopy features of invasive melanoma. Melanoma Res
1996, 6: 55-62.
9. Menzies SW, Ingvar C, Crotty KA, et al. Frequency and morphologic
characteristics of invasive melanomas lacking specific surface microscopic
features. Arch Derm 1996; 132: 1178-82.
10. Stolz W, Braun-Falco O, Bilek P, Landthaler M, Cognetta AB. Color
Atlas of Dermatoscopy. Blackwell Science Ltd, 1994.
11. Stolz W, Riemann A, Cognetta AB, Pillet l, Abmayr W, Holzel D,
Bilek P, Nachbar F, Landthaler M, Braun-Falco O. ABCD rule of dermatoscopy:
a new practical method for early recognition of malignant melanoma. Eur
J Dermatol 1994; 4: 521-7.
12. Nachbar F, Stolz W, Merkle T. The ABCD rule of dermatoscopy: high
prospective value in the diagnosis of doubtful melanocytic skin lesions.
J Am Acad Dermatol 1994; 30: 551-9.
13. Barnhill RL, Fandrey K, Levy MA, Mihm MC, Hyman B. Angiogenesis
and tumor progression of melanoma. Quantification of vascularity in melanocytic
nevi and cutaneous malignant melanoma. Lab Invest 1992; 67: 331-7.
14. Cockerell CJ, Sonnier G, Kelly L, Patel S. Comparative analysis
of neovascularization in primary cutaneous melanoma and spitz nevus. Am
J Dermatopathol 1994; 16: 9-13.
15. Barnhill RL, Levy MA. Regressing thin cutaneous malignant melanomas
(< or = 1.0 mm) are associated with angiogenesis. Am J Pathol
1993; 143: 99-104.
16. Kenet RO, Kang S, Kenet BJ, Fitzpatrick TB, Sober AJ, Barnhill RL.
Clinical diagnosis of pigmented lesions using digital epiluminescence
microscopy. Arch Derm 1993; 129: 157-74.
17. Yadav S, Vossaert KA, Kopf AW, Silverman M, Grin-Jorgensen C. Histopathologic
correlates of structures seen on dermoscopy (epiluminescence microscopy).
Am J Dermatopathol 1993; 15: 297-305.
18. Saida T, Oguchi S, Ishiara Y. In vivo observation of magnified
features of pigmented lesions on volar skin using video macroscope. Arch
Dermatol 1995; 131: 298-304.
19. Benelli C, Gianotti R, Dal Pozzo V, Roscetti E. Melanoma with dermatoscopic
features of seborrheic keratosis. Eur J Dermatol 1996; 6: 246-7.
20. Dal Pozzo V, Benelli C. Atlante di Dermoscopia. Edra Publishing
House, Milano, 1996.
21. Dal Pozzo V, Benelli C. Atlas of Dermoscopy. Edra Edizioni,
Milano, 1997.
|
Version imprimable
