- Email: [email protected]
Epiluminescence microscopy: Criteria of cutaneous melanoma progression
Epiluminescence microscopy: Criteria of cutaneous melanoma progression Giuseppe Argenziano, MD, a Gabriella Fabbrocini, MD, b Paolo Carli, MD, c Vincenzo De Giorgi, MD, c and Mario Delfino, MD a Naples and Florence, Italy
Background: Cutaneous melanoma develops through a series of evolutionary steps (intraepidermal, radial, and vertical growth phases) that are traceable in specific histologic features. Epiluminescence microscopy (ELM) is an in vivo technique that enables the visualization of morphologic structures in pigmented lesions correlated with specific histologic architectural characteristics. Many ELM criteria associated with cutaneous melanoma have been described, but their correlation with tumor progression has not yet been established. Objective: In this preliminary study our purpose was to explore the possibility of recognizing ELM criteria that allow the in vivo detection of the various phases of melanoma progression as well as tumor depth. Methods: Seventy-two cutaneous melanomas (41 "thin" melanomas [TnM], < 0.76 mm thickness, and 31 "thick" melanomas [TkM], > 0.75 mm thickness) were investigated with ELM for the presence of nine standard ELM criteria; their significance was determined by calculating the chi-square test of independence. Results: A significant association is found between the presence of pigment network and TnM and between the presence of gray-blue areas, vascular pattern, and TkM. Moreover, pigment network plus radial streaming is the most significant association of ELM criteria in TnM, whereas gray-blue areas plus vascular pattern is the greatest in TkM. Conclusion: This study shows a good correlation between certain ELM criteria and the histologic architecture of cutaneous melanoma for a preoperative evaluation of the tumor thickness. Further investigation is needed for verifying on a larger number of cases our pilot estimates of sensitivity and specificity of ELM criteria in thin and thick melanomas. (J Am Acad Dermatol 1997;37:68-74.)
Cutaneous melanoma develops through a series of evolutionary phases that are traceable in specific histologic features. 1 In the in situ melanoma phase, the atypical melanocytes are confined to the epidermis, whereas the radial growth phase includes superficially invasive tumor cells that extend into the papillary dermis. Vertical growth phase melanoma, moreover, is characterized by deeply invasive nests of
From the Institute of Dermatology a and Department of Molecular and Cellular Biology and Pathology,b Fedefico II University of Naples, and the Institute of Dermatology, University of Florence.c Accepted for publication March 12, 1997. Reprint requests: Giuseppe Argenziano, MD, Clinica Dermatologica, Universith "Fedefico II," Via S. Pansini 5, 80131 Napoli, Italia. Copyright © 1997 by the American Academy of Dermatology, Inc. 0190-9622/97/$5.00 + 0 16/1/81855
68
melanocytes that have become confluent to form sheets. 2 Epiluminescence microscopy (ELM) is an in vivo, noninvasive technique that enables the clinician to visualize morphologic structures in pigmented lesions that are correlated with specific histologic architectural features. 3,4 Numerous studies 5-10 have demonstrated that ELM improves the capacity to differentiate between benign and malignant melanocytic lesions, thereby promoting the early diagnosis of melanoma. Moreover, a previous study 11 referred to the possibility of preoperative tumor thickness measurement in malignant melanoma by high-frequency ultrasound and ELM. In this preliminary study we hypothesized that ELM criteria could be determined that allow the in vivo detection of the various phases of melanoma
Journal of the American Academy of Dermatology Volume 37, Number 1
Fig. 1. Thin melanoma (0.3 mm thick, Clark level II). Prevalence of pigment network, diffuse gray-black blotches, radial streaming, and structural asymmetry; black dots, brown globules, and regression pattern are also seen; absence of gray-blue areas and vascular pattern. (Original magnification xl6.)
Argenziano et al. 69
Fig. 2. Thick melanoma (1.4 mm thick, Clark level IV). Prevalence of gray-blue areas and structural asymmetry; dotted vascular pattern (arrow), black dots, radial streaming, regression pattern, and an almost completely unrecognizable pigment network are also seen. (Original magnification xl6.)
progression as well as the tumor depth. For this purpose, we evaluated the percentage expression o f certain E L M criteria in melanomas with a thickness o f less than or more than 0.75 mm. This permitted us to define the first appearance of the various E L M criteria in relation to the radial or vertical growth phases of the tumor. MATERIAL A N D M E T H O D S
A total of 72 patients whose ages ranged from 23 to 75 years (29 men and 43 women) affected by cutaneous melanoma were included in this study. Each lesion, after being covered by immersion oil and a glass slide (to render the stratum corueum translucent), was examined with a dermatoscope (Heine Delta 10, Optotechnik, Herrsching, Germany) and photographed with Dermaphot photo equipment (Heine Optotechnik) at a fixed magnification of ×10. In addition, 31 randomly selected cases were observed with a Wild M650 stereomicroscope (Heerbrugg AG) with 6 to 40 times magnification and equipped with a Nikon F3 camera; this way of proceeding was adopted to compare the two different viewing systems. ELM images, achieved with Dermaphot, were studied to evaluate the presence or absence of nine standard ELM criteria associated with melanoma and selected by us on the basis of their histologic significance. The following criteria were in accordance with those agreed on at the Consensus Meeting of Hamburg12: irregular, prominent, and broad pigment network; black dots; radial streaming; irregular brown globules; gray-blue areas; and white, scarlike areas. In addition, three more ELM criteria were selected: diffuse gray-black blotches, 4,9 linear or dotted vascular
Fig. 3. Thick melanoma (same lesion as Fig. 2 at higher magnification). Dense gray-blue areas and dotted vascular pattern (arrow). (Original magnification x25.)
pattern, 13,14 and structural asymmetry (described as asymmetric pigment distribution, 1° differential structures, 13 or multicomponent pattern15). The latter was defined as the presence of at least two ELM criteria irregularly or asymmetrically distributed throughout the lesion. 16 The evaluation of the presence or absence of the ELM criteria was carried out with the consensus of at least two of three different investigators. After ELM examination and photography the diagnosis of all the lesions was established histopathologically. The 72 melanomas included in the study (Table I) were subdivided, on the basis of their thickness, into the following two groups: 41 "thin" melanomas (TnM) with tumor thickness of less than 0.76 mm (Fig. 1), and 31 "thick" melanomas (TkM) with tumor thickness of greater than 0.75 mm. All TnMs were Clark level I (in situ) or II, except three cases of III
Journal of the American Academy of Dermatology
July 1997
70 Argenziano et al.
Table I. Tumor thickness, Clark level, and nevic remnants in observed cases of melanoma (N = 72) Clark level
I
Thickness
TnM<0.76mm TkM > 0.75 mm
II
HI
IV
Nevic remnants
No.
%
No.
%
No.
%
No.
%
No.
%
No.
41 31
57 43
13 0
18 0
25 1
35 1
3 19
4 27
0 11
0 15
8 4
%
11 6
TkM, Thick melanoma; TnM, thin melanoma.
Table lI. ELM criteria: Results of 72 melanoma cases ELM criterion
Pigment network Black dots Radial streaming Brown globules Diffuse gray-black blotches Gray-blue areas White scarlike areas Linear vascular pattern Dotted (or elongated loops) vascular pattern Structural asymmetry
Present in 41 TnM
Present in 31 TkM
No.
%
No.
%
39 25 34 32 25 10 23 6 1 31
95 61 83 78 61 24 56 15 2 76
21 20 22 22 20 29 12 6 14 27
68 64 71 71 64 94 39 19 45 87
TkM, Thick melanoma; TnM, thin melanoma.
level (mean tumor thickness: 0.42 mm); all TkMs were III or IV level, except one case of II level (mean tumor thickness: 1.5 ram). Only three TkMs were more than 3 mm thick; 17% of all melanomas arose in association with preexisting melanocytic nevi. A univariate statistical analysis was carried out for estimating the distribution frequency of the single clinical, histopathologic, and ELM variables (BMDP program, IBM, 1991 version). The analysis of cross tables was performed by the chi-square test of independence to evaluate the relation of ELM criteria between the two groups of melanoma. RESULTS
The results are shown in Table I1, where we report the frequency of occurrence of the standard ELM criteria (in Table II the vascular pattern is composed of two variables) in TnM and TkM. TriMs are characterized by a higher frequency of occurrence of pigment network (95% TnM vs 68% TkM), radial streaming (83% TnM vs 71% TkM) and white scarlike areas (56% TnM vs 39% TkM). In contrast, TkMs exhibit a greater frequency of occurrence of gray-blue areas (94% TkM vs 24% TnM), structural asymmetry (87% TkM vs
76% TnM) and vascular patterns (64% TkM vs 17% TnM). In particular, a linear vascular typology is noted in 19% of TkM versus 15% of TnM, whereas a dotted vascular pattern is observed in 45% of TkM and in just one case of TnM. Moreover, two melanomas with tumor thickness of more than 3 mm show a typical vascular pattern with an elongated loop conformation. Black dots, brown globules, and diffuse grayblack blotches do not show any difference between the two melanoma groups. The preliminary data of our study point out a significant statistical association (Table III) between the pigment network and TnM with 95 % sensitivity and 32% specificity. Pigment network plus radial streaming is the most frequent association between two ELM criteria in TnM; it is found in 80% of TnMs versus 52% of TkMs (Table IV). A significant statistical association is also seen (Table V) between the gray-blue areas and TkM, and between the vascular pattern and TkM. In particular the gray-blue areas show 94% sensitivity and 76% specificity, whereas the dotted (or elongated loop) vascular pattern exhibits 45% sensitivity but 98% specificity. Moreover,
Journal of the American Academy of Dermatology Volume 37, Number 1
Argenziano et al. 71
Table III. Sensitivity, specificity, and statistical significance of the pigment network in thin melanoma Sensitivity (%)
True positive/ total thin
Specificity (%)
True negative/ total thick
melanomas
95
p Value*
melanomas
39/41
32
10/31
0.002
*Chi-square significance.
Table IV. Most frequent associations between two ELM criteria in thin and thick melanomas ELM criteria
Thin melanoma (n = 41)
Pigment network plus radial streaming Gray-blue areas plus dotted (or elongated loops) vascular pattern
Thick melanoma (n = 31)
p Value*
No.
%
No.
%
33
80
16 52
0.009
1
2
13 42
< 0.0001
*Chi-square significance.
Table V. Sensitivity, specificity, and statistical significance of gray-blue areas and dotted vascular pattern in thick melanoma ELM criterion
Sensitivity (%)
True positive/ total thick
Specificity (%)
melanomas
True negative/ total thin
p Value*
melanomas
Gray-blue areas
94
29/31
76
31/41
< 0.0001
Dotted (or elongated loops) vascular pattern
45
14/31
98
40/41
< 0.0001
*Chi-square significance.
these two criteria are both present in 42% of TkMs versus only 2% of TnMs (Table IV). The epiluminescence stereomicroscope, used in addition in 31 randomly selected cases, in comparison with the dermatoscope, allowed us to better evaluate (because of a higher magnification and more intense lighting) the pigmentary and vascular structures in both heavily and slightly pigmented lesions. In fact, in heavily pigmented lesions the deep structures are less visible when the upper cutis is fully pigmented; on the other hand, in slightly pigmented lesions the cells and tissues are translucent because they are less colored by the melanin. Moreover, the stereomicroscope gave us a greater capacity to differentiate between the peppering and the gray-blue areas and between brown globules and black dots.
DISCUSSION Because evolutionary steps of melanoma are expressed by specific histologic patterns (Table VI), 2 we have selected their correlated ELM criteria 3,4,12-14 (Table VII). These criteria have been chosen to verify the hypothesis of the in vivo detection of the various phases of melanoma progression. On the basis of the ELM criteria incidence in TnM versus TkM, we have defined the first appearance of the different ELM features (Table VII). It is noteworthy that the ELM criteria mentioned are not exclusive of cutaneous melanomas but can also be seen with many types of pigmented lesions, such as dysplastic, congenital, lentiginous nevi and lentigines.
72 Argenziano et al.
Journal of the American Academy of Dermatology July 1997
T a b l e VI. Phases of histologic progression in cutaneous melanoma 2
Architectural features 1. Atypical melanocytic hyperplasia: Proliferation of atypical melanocytes disposed as solitary units along dermoepidermal junction 2. Pagetoid melanocytic infiltration: Progressive involvement of epidermis by atypical pagetoid melanocytes that are situated above dermoepidermal junction (sometimes including granular and cornified layers) both as solitary units and in nests 3. Progressive formation of irregular and confluent junctional nests of melanocytes 4. Involvement of dennis by invasive and irregular nests of melanocytes that have become confluent to form sheets descending in deeper cuffs. Mesenchymal features 1. Inflammatory infiltration of lymphocytes with a variable number of melanophages, which can occur in the in situ phase of melanoma 2. Regression (as a consequence of effects of lymphocytic infiltrates on melanocytes) with fibroplasia, melanophages, and telangiectases. Other diagnostic findings 1. Asymmetry 2. Poor circumscription 3. Failure of maturation of melanocytes with progressive descent into dermis
T a b l e VII. Malignant ELM criteria, their histologic correlates, and stage of appearance in relation to the radial or vertical growth phases of melanoma ELM criterion
Histologic correlates
Stage of appearance
1. Irregular, prominent, and broad pigment network
Atypical melanocytic hyperplasia with irregular, hyperpigmented and broadened rete ridges
TnM
2. Black dots
Pagetoid melanocytic infiltration with collections of melanocytes and/or melanin in the cornified layer
TnM
3. Radial streaming, pseudopods, irregular extensions 4. Irregular brown globules
Confluent radial junctional nests of melanocytes
TnM
Nests of melanocytes irregularly distributed throughout the dermoepidermal junction and/or in papillary dermis
TnM
5. Diffuse gray-black blotches
Hyperpigmentation throughout all levels of the epidermis and/or upper dermis (in melanocytes or melanophages) Pigmented melanophages or melanocytes of mid-reticular dermis location
TnM
6. Gray-blue areas 7. White scarlike areas with "peppering" and telangiectases 8. (a) Linear vascular pattern; (b) dotted (or elongated loops) vascular pattern 9. Structural asymmetry TkM, Thickmelanoma;TnM, thin melanoma.
Areas of melanoma regression with fibroplasia, melanophages, and telangiectases Neovascularization of (a) upper dermis location and (b) mid-reticular dermis location Architectural asymmetry
TkM, TnM TnM,
(a)TnM, (b) T~M TnM
Journal of the AmericanAcademyof Dermatology Volume37, Number 1
E L M criteria analysis 1. The pigment network, expression of pigmented and elongated fete ridges, in melanoma appears irregular, prominent, and broadened by the proliferation of atypical melanocytes disposed as solitary units along the dermoepidermal junction (atypical melanocytic hyperplasia) or arranged as irregular and confluent junctional nests. 2,3 Our data confirm that the pigment network appears in the early phases of the melanoma progression and is still observed in the majority of TkMs. Nevertheless in 5% of TnM cases and in 32% of TkM cases it is not visible, presumably because of a respective primary lack or loss of rete ridges. 11,14 We also observed that the pigment network irregularity could be less evident in the early phases of the tumor progression, probably because of the paucity and discontinuity of the cytoarchitectural atypias, 2 or when the melanoma arises in association with preexisting melanocytic nevus. 2. Black dots do not always assume the aspect of focal, round, black structures, but in our experience they could become confluent in greater and irregular accumulations or have a pulverulent appearance (numerous minute grains visible with greater magnification). Such structures, correlating with collections of "free" melanin or melanocytes in the cornifled layer,3,4 in melanomas could represent the atypical pagetoid melanocytic infiltration Of epidermis. Nevertheless, black dots were not always detected (present in 61% to 64% of our cases). This could be caused by the diverse capacity of the cells themselves to produce melanin. The first appearance of black dots occurs already during the in situ phase of the tumor. No differences were observed between the two groups of melanoma. 3. Radial streaming, pseudopods, irregular extensions, 3 and branched streaks 13 are morphologically dissimilar but all histologically correlate to confluent radial junctional nests of melanocytes. 12,17 Therefore we describe them as a single criterion. In this study we have chosen to use the term radial streaming because, in our opinion, this word better defines the histologic correlate of the structures. In contrast with other reports, 17 we observed radial streaming in the in situ melanomas (100% of cases). Moreover, they were more present in TnMs (83% TnM vs 71%
T~'Vl). 4. The brown globules are correlated in
Argenziano et al. 73
melanoma to nests of melanocytes irregularly distributed throughout the dermoepidermal junction or in the papillary dermis or both. 4 For this reason they appear in the radial growth phase of melanoma (78% of TnM). A similar number was found in TkM (71%). 5. Diffuse gray-black blotches are correlated to an intense melanic pigmentation throughout all levels of the epidermis (hyperpigmented keratinocytes or melanocytes) or upper cutis. 4 At this site the pigment could be related to melanophagic or melanocytic infiltration. Such a criterion is reported in 61% to 64% of the melanomas studied, without any difference between the two categories. 6. The gray-blue areas are observed when pigmented melanocytes or melanophages are located in the mid and reticular dermis. 4 Such a situation is found in 94% of TkMs, and we speculate that in these cases gray-blue areas are caused by the melanocytic infiltration of the mid-reticular dermis. When this criterion is found in TnM (24% of cases), we assume it is correlated to a deep melanophagic infiltration. In our experience the gray-blue areas observed in TkM show a greater density than that observed in TnM. 7. The white scarlike areas are histologically correlated to zones of melanoma regression. 4,12 The dermoepidermal pigmentation has disappeared, and in a fibroplastic dermis there is persistence of telangiectases and melanophages in the form of small pepperlike dots. 14,18 The regression pattern first appears in the radial growth phase of the tumor (56% of TnM cases) and is found in only 39% of TkM cases. 8. A vascular pattern is observed in the regressive areas but also in melanomas not heavily pigmented. 13 It can have various aspects; 15% to 19% of all melanomas observed had irregular, linear vessels (parallel to the skin surface) of upper dermis location, and 45% of TkM (and only 2% of TnM) had regular, dotted vessels (perpendicular to the skin surface) of mid-reticular dermis location. 1334 In some cases milky-red globules (reported by Stolz et al.13), as well as vascularized nests of amelanotic cells, were seen. In two of the three cases of melanoma with a thickness of more than 3 mm we noted the presence of a vascular pattern with elongated loops throughout the nodular component. 9. Structural asymmetry could be the ELM
74
Journal of the American Academy of Dermatology July 1997
Argenziano et al.
correlate o f the histologic architectural asymmetry of cutaneous melanoma. The first appearance o f the pattern is in the early phases of the tumor progression (76% of TnM), but it finds a greater incidence in T k M (87% o f cases); that fact is probably caused by coexistent radial and vertical growth patterns.
4.
5.
CONCLUSION The analysis of the results o f our pilot study shows a good correlation between the selected E L M criteria and the histologic architecture o f the cultaneous melanoma, for a preoperative evaluation of the tumor thickness. Pigment network and radial streaming indicate the radial growth phase of melanoma, whereas gray-blue areas and dotted vascular pattern (or the elongated loops) indicate the involvement of the mid and deep dermis and therefore the vertical growth phase o f the tumor. The lower frequency o f the regression pattern registered in T k M versus T n M (and in the opposite, the higher percentage of structural a s y m m e t r y in T k M ) c o u l d be explained by the emergence, in the latter, of more atypical cellular clones that have vertical growth and predominate the regression pattern. The gray-blue areas and the typical vascular pattern, individualized in our study as distinctive characteristics for the E L M diagnosis o f TkM, were both absent in only one case o f them (0.9 m m thick), whereas they were both present in only one case of TnM. Further investigation is needed to verify, on a larger number o f cases, our pilot estimates o f sensitivity and specificity of E L M criteria in thin and thick melanomas. REFERENCES
1. Barnhill RL, Fitzpatrick TB, Fandrey K, et al., editors. Color atlas and synopsis of pigmented lesions. New York: McGraw-Hill; 1995. p. 145-51. 2. Caputo R, Gelmetti C, Annessi G, editors. Pediatric dermatology and dermatopathology; vol. 3. Baltimore: Williams & Wilkins; 1995. p. 307-20. 3. Soyer HP, Smolle J, Hodl S, Pachernegg H, Kerl H.
6.
7.
8.
9.
10. 11.
12. 13. 14. 15.
16. 17. 18.
Surface microscopy: a new approach to the diagnosis of cutaneous pigmented tumors. Am J Dermatopathol 1989;11:1-10. Yadav S, Vossaert KA, Kopf AW, Silverman M, GrinJorgensen C. Histopathologic correlates of structures seen on dermoscopy (epiluminescence microscopy). Am J Dermatopathol 1993;15:297-305. Pehamberger H, Steiner A, Wolff K. In vivo epiluminescence microscopy of pigmented skin lesions. I. Pattern analysis of pigmented skin lesions. J Am Acad Dermatol 1987;17:571-83. Steiner A, Pehamberger H, Wolff K. In vivo epiluminescence microscopy of pigmented skin lesions. II. Diagnosis of small pigmented skin lesions and early detection of malignant melanoma. J Am Acad Dermatol 1987;17:584-91. Pehamberger H, Binder M, Steiner A, Wolff K. In vivo epiluminescencemicroscopy:improvement of early diagnosis of melanoma. J Invest Dermatol 1993;100(suppl): 356S-62S. Steiner A, Binder M, Schemper M, Wolff K, Pehamberger H. Statistical evaluation of epiluminescence microscopy criteria for melanocytic pigmented skin lesions. J Am Acad Dermatol 1993;29:581-8. Soyer HP, Smolle J, Leitinger G, Rieger E, Kerl H. Diagnostic reliability of dermoscopic criteria for detecting malignant melanoma. Dermatology 1995;190:2530. Nilles M, Boedeker RH, Schill WB. Surface microscopy of naevi and melanomas: clues to melanoma. Br J Dermatol 1994;130:349-55. Dummer W, Blaheta H-J, Bastian BC et al. Preoperative characterization of pigmented skin lesions by epiluminescence microscopy and high-frequency ultrasound. Arch Dermatol 1995;131:279-85. Bahmer FA, Fritsch P, Kreusch J, et al. Terminology in surface microscopy.J Am Acad Dermatol 1990;23:115962. Stolz W, Braun-Falco O, Bilek P, et al., editors. Color atlas of dermatoscopy. Oxford: Blackwell Science; 1994. Kreusch J, Rassner G, editors. Auflichtmikroskopie pigmentierter Hanttumoren: ein Bildatlas. Stuttgart: Thieme; 1991. Kenet RO, Kang S, Kenet BJ, et al. Clinical diagnosis of pigmented lesions using digital epiluminescence microscopy: grading protocol and atlas. Arch Dermatol 1993; 129:157-74. Dal Pozzo V, Benelli C, editors. Atlante di dermoscopia. Milano: Edra; 1996. Menzies SW, Crotty KA, McCarthy WH. The morphologic criteria of the pseudopod in surface microscopy. Arch Dermatol 1995;131:436-40. Menzies SW, Crotty KA, Ingvar C, et al., editors. An atlas of surface microscopy of pigmented skin lesions. Sydney, Australia: McGraw-Hill; 1996.
Background: Cutaneous melanoma develops through a series of evolutionary steps (intraepidermal, radial, and vertical growth phases) that are traceable in specific histologic features. Epiluminescence microscopy (ELM) is an in vivo technique that enables the visualization of morphologic structures in pigmented lesions correlated with specific histologic architectural characteristics. Many ELM criteria associated with cutaneous melanoma have been described, but their correlation with tumor progression has not yet been established. Objective: In this preliminary study our purpose was to explore the possibility of recognizing ELM criteria that allow the in vivo detection of the various phases of melanoma progression as well as tumor depth. Methods: Seventy-two cutaneous melanomas (41 "thin" melanomas [TnM], < 0.76 mm thickness, and 31 "thick" melanomas [TkM], > 0.75 mm thickness) were investigated with ELM for the presence of nine standard ELM criteria; their significance was determined by calculating the chi-square test of independence. Results: A significant association is found between the presence of pigment network and TnM and between the presence of gray-blue areas, vascular pattern, and TkM. Moreover, pigment network plus radial streaming is the most significant association of ELM criteria in TnM, whereas gray-blue areas plus vascular pattern is the greatest in TkM. Conclusion: This study shows a good correlation between certain ELM criteria and the histologic architecture of cutaneous melanoma for a preoperative evaluation of the tumor thickness. Further investigation is needed for verifying on a larger number of cases our pilot estimates of sensitivity and specificity of ELM criteria in thin and thick melanomas. (J Am Acad Dermatol 1997;37:68-74.)
Cutaneous melanoma develops through a series of evolutionary phases that are traceable in specific histologic features. 1 In the in situ melanoma phase, the atypical melanocytes are confined to the epidermis, whereas the radial growth phase includes superficially invasive tumor cells that extend into the papillary dermis. Vertical growth phase melanoma, moreover, is characterized by deeply invasive nests of
From the Institute of Dermatology a and Department of Molecular and Cellular Biology and Pathology,b Fedefico II University of Naples, and the Institute of Dermatology, University of Florence.c Accepted for publication March 12, 1997. Reprint requests: Giuseppe Argenziano, MD, Clinica Dermatologica, Universith "Fedefico II," Via S. Pansini 5, 80131 Napoli, Italia. Copyright © 1997 by the American Academy of Dermatology, Inc. 0190-9622/97/$5.00 + 0 16/1/81855
68
melanocytes that have become confluent to form sheets. 2 Epiluminescence microscopy (ELM) is an in vivo, noninvasive technique that enables the clinician to visualize morphologic structures in pigmented lesions that are correlated with specific histologic architectural features. 3,4 Numerous studies 5-10 have demonstrated that ELM improves the capacity to differentiate between benign and malignant melanocytic lesions, thereby promoting the early diagnosis of melanoma. Moreover, a previous study 11 referred to the possibility of preoperative tumor thickness measurement in malignant melanoma by high-frequency ultrasound and ELM. In this preliminary study we hypothesized that ELM criteria could be determined that allow the in vivo detection of the various phases of melanoma
Journal of the American Academy of Dermatology Volume 37, Number 1
Fig. 1. Thin melanoma (0.3 mm thick, Clark level II). Prevalence of pigment network, diffuse gray-black blotches, radial streaming, and structural asymmetry; black dots, brown globules, and regression pattern are also seen; absence of gray-blue areas and vascular pattern. (Original magnification xl6.)
Argenziano et al. 69
Fig. 2. Thick melanoma (1.4 mm thick, Clark level IV). Prevalence of gray-blue areas and structural asymmetry; dotted vascular pattern (arrow), black dots, radial streaming, regression pattern, and an almost completely unrecognizable pigment network are also seen. (Original magnification xl6.)
progression as well as the tumor depth. For this purpose, we evaluated the percentage expression o f certain E L M criteria in melanomas with a thickness o f less than or more than 0.75 mm. This permitted us to define the first appearance of the various E L M criteria in relation to the radial or vertical growth phases of the tumor. MATERIAL A N D M E T H O D S
A total of 72 patients whose ages ranged from 23 to 75 years (29 men and 43 women) affected by cutaneous melanoma were included in this study. Each lesion, after being covered by immersion oil and a glass slide (to render the stratum corueum translucent), was examined with a dermatoscope (Heine Delta 10, Optotechnik, Herrsching, Germany) and photographed with Dermaphot photo equipment (Heine Optotechnik) at a fixed magnification of ×10. In addition, 31 randomly selected cases were observed with a Wild M650 stereomicroscope (Heerbrugg AG) with 6 to 40 times magnification and equipped with a Nikon F3 camera; this way of proceeding was adopted to compare the two different viewing systems. ELM images, achieved with Dermaphot, were studied to evaluate the presence or absence of nine standard ELM criteria associated with melanoma and selected by us on the basis of their histologic significance. The following criteria were in accordance with those agreed on at the Consensus Meeting of Hamburg12: irregular, prominent, and broad pigment network; black dots; radial streaming; irregular brown globules; gray-blue areas; and white, scarlike areas. In addition, three more ELM criteria were selected: diffuse gray-black blotches, 4,9 linear or dotted vascular
Fig. 3. Thick melanoma (same lesion as Fig. 2 at higher magnification). Dense gray-blue areas and dotted vascular pattern (arrow). (Original magnification x25.)
pattern, 13,14 and structural asymmetry (described as asymmetric pigment distribution, 1° differential structures, 13 or multicomponent pattern15). The latter was defined as the presence of at least two ELM criteria irregularly or asymmetrically distributed throughout the lesion. 16 The evaluation of the presence or absence of the ELM criteria was carried out with the consensus of at least two of three different investigators. After ELM examination and photography the diagnosis of all the lesions was established histopathologically. The 72 melanomas included in the study (Table I) were subdivided, on the basis of their thickness, into the following two groups: 41 "thin" melanomas (TnM) with tumor thickness of less than 0.76 mm (Fig. 1), and 31 "thick" melanomas (TkM) with tumor thickness of greater than 0.75 mm. All TnMs were Clark level I (in situ) or II, except three cases of III
Journal of the American Academy of Dermatology
July 1997
70 Argenziano et al.
Table I. Tumor thickness, Clark level, and nevic remnants in observed cases of melanoma (N = 72) Clark level
I
Thickness
TnM<0.76mm TkM > 0.75 mm
II
HI
IV
Nevic remnants
No.
%
No.
%
No.
%
No.
%
No.
%
No.
41 31
57 43
13 0
18 0
25 1
35 1
3 19
4 27
0 11
0 15
8 4
%
11 6
TkM, Thick melanoma; TnM, thin melanoma.
Table lI. ELM criteria: Results of 72 melanoma cases ELM criterion
Pigment network Black dots Radial streaming Brown globules Diffuse gray-black blotches Gray-blue areas White scarlike areas Linear vascular pattern Dotted (or elongated loops) vascular pattern Structural asymmetry
Present in 41 TnM
Present in 31 TkM
No.
%
No.
%
39 25 34 32 25 10 23 6 1 31
95 61 83 78 61 24 56 15 2 76
21 20 22 22 20 29 12 6 14 27
68 64 71 71 64 94 39 19 45 87
TkM, Thick melanoma; TnM, thin melanoma.
level (mean tumor thickness: 0.42 mm); all TkMs were III or IV level, except one case of II level (mean tumor thickness: 1.5 ram). Only three TkMs were more than 3 mm thick; 17% of all melanomas arose in association with preexisting melanocytic nevi. A univariate statistical analysis was carried out for estimating the distribution frequency of the single clinical, histopathologic, and ELM variables (BMDP program, IBM, 1991 version). The analysis of cross tables was performed by the chi-square test of independence to evaluate the relation of ELM criteria between the two groups of melanoma. RESULTS
The results are shown in Table I1, where we report the frequency of occurrence of the standard ELM criteria (in Table II the vascular pattern is composed of two variables) in TnM and TkM. TriMs are characterized by a higher frequency of occurrence of pigment network (95% TnM vs 68% TkM), radial streaming (83% TnM vs 71% TkM) and white scarlike areas (56% TnM vs 39% TkM). In contrast, TkMs exhibit a greater frequency of occurrence of gray-blue areas (94% TkM vs 24% TnM), structural asymmetry (87% TkM vs
76% TnM) and vascular patterns (64% TkM vs 17% TnM). In particular, a linear vascular typology is noted in 19% of TkM versus 15% of TnM, whereas a dotted vascular pattern is observed in 45% of TkM and in just one case of TnM. Moreover, two melanomas with tumor thickness of more than 3 mm show a typical vascular pattern with an elongated loop conformation. Black dots, brown globules, and diffuse grayblack blotches do not show any difference between the two melanoma groups. The preliminary data of our study point out a significant statistical association (Table III) between the pigment network and TnM with 95 % sensitivity and 32% specificity. Pigment network plus radial streaming is the most frequent association between two ELM criteria in TnM; it is found in 80% of TnMs versus 52% of TkMs (Table IV). A significant statistical association is also seen (Table V) between the gray-blue areas and TkM, and between the vascular pattern and TkM. In particular the gray-blue areas show 94% sensitivity and 76% specificity, whereas the dotted (or elongated loop) vascular pattern exhibits 45% sensitivity but 98% specificity. Moreover,
Journal of the American Academy of Dermatology Volume 37, Number 1
Argenziano et al. 71
Table III. Sensitivity, specificity, and statistical significance of the pigment network in thin melanoma Sensitivity (%)
True positive/ total thin
Specificity (%)
True negative/ total thick
melanomas
95
p Value*
melanomas
39/41
32
10/31
0.002
*Chi-square significance.
Table IV. Most frequent associations between two ELM criteria in thin and thick melanomas ELM criteria
Thin melanoma (n = 41)
Pigment network plus radial streaming Gray-blue areas plus dotted (or elongated loops) vascular pattern
Thick melanoma (n = 31)
p Value*
No.
%
No.
%
33
80
16 52
0.009
1
2
13 42
< 0.0001
*Chi-square significance.
Table V. Sensitivity, specificity, and statistical significance of gray-blue areas and dotted vascular pattern in thick melanoma ELM criterion
Sensitivity (%)
True positive/ total thick
Specificity (%)
melanomas
True negative/ total thin
p Value*
melanomas
Gray-blue areas
94
29/31
76
31/41
< 0.0001
Dotted (or elongated loops) vascular pattern
45
14/31
98
40/41
< 0.0001
*Chi-square significance.
these two criteria are both present in 42% of TkMs versus only 2% of TnMs (Table IV). The epiluminescence stereomicroscope, used in addition in 31 randomly selected cases, in comparison with the dermatoscope, allowed us to better evaluate (because of a higher magnification and more intense lighting) the pigmentary and vascular structures in both heavily and slightly pigmented lesions. In fact, in heavily pigmented lesions the deep structures are less visible when the upper cutis is fully pigmented; on the other hand, in slightly pigmented lesions the cells and tissues are translucent because they are less colored by the melanin. Moreover, the stereomicroscope gave us a greater capacity to differentiate between the peppering and the gray-blue areas and between brown globules and black dots.
DISCUSSION Because evolutionary steps of melanoma are expressed by specific histologic patterns (Table VI), 2 we have selected their correlated ELM criteria 3,4,12-14 (Table VII). These criteria have been chosen to verify the hypothesis of the in vivo detection of the various phases of melanoma progression. On the basis of the ELM criteria incidence in TnM versus TkM, we have defined the first appearance of the different ELM features (Table VII). It is noteworthy that the ELM criteria mentioned are not exclusive of cutaneous melanomas but can also be seen with many types of pigmented lesions, such as dysplastic, congenital, lentiginous nevi and lentigines.
72 Argenziano et al.
Journal of the American Academy of Dermatology July 1997
T a b l e VI. Phases of histologic progression in cutaneous melanoma 2
Architectural features 1. Atypical melanocytic hyperplasia: Proliferation of atypical melanocytes disposed as solitary units along dermoepidermal junction 2. Pagetoid melanocytic infiltration: Progressive involvement of epidermis by atypical pagetoid melanocytes that are situated above dermoepidermal junction (sometimes including granular and cornified layers) both as solitary units and in nests 3. Progressive formation of irregular and confluent junctional nests of melanocytes 4. Involvement of dennis by invasive and irregular nests of melanocytes that have become confluent to form sheets descending in deeper cuffs. Mesenchymal features 1. Inflammatory infiltration of lymphocytes with a variable number of melanophages, which can occur in the in situ phase of melanoma 2. Regression (as a consequence of effects of lymphocytic infiltrates on melanocytes) with fibroplasia, melanophages, and telangiectases. Other diagnostic findings 1. Asymmetry 2. Poor circumscription 3. Failure of maturation of melanocytes with progressive descent into dermis
T a b l e VII. Malignant ELM criteria, their histologic correlates, and stage of appearance in relation to the radial or vertical growth phases of melanoma ELM criterion
Histologic correlates
Stage of appearance
1. Irregular, prominent, and broad pigment network
Atypical melanocytic hyperplasia with irregular, hyperpigmented and broadened rete ridges
TnM
2. Black dots
Pagetoid melanocytic infiltration with collections of melanocytes and/or melanin in the cornified layer
TnM
3. Radial streaming, pseudopods, irregular extensions 4. Irregular brown globules
Confluent radial junctional nests of melanocytes
TnM
Nests of melanocytes irregularly distributed throughout the dermoepidermal junction and/or in papillary dermis
TnM
5. Diffuse gray-black blotches
Hyperpigmentation throughout all levels of the epidermis and/or upper dermis (in melanocytes or melanophages) Pigmented melanophages or melanocytes of mid-reticular dermis location
TnM
6. Gray-blue areas 7. White scarlike areas with "peppering" and telangiectases 8. (a) Linear vascular pattern; (b) dotted (or elongated loops) vascular pattern 9. Structural asymmetry TkM, Thickmelanoma;TnM, thin melanoma.
Areas of melanoma regression with fibroplasia, melanophages, and telangiectases Neovascularization of (a) upper dermis location and (b) mid-reticular dermis location Architectural asymmetry
TkM, TnM TnM,
(a)TnM, (b) T~M TnM
Journal of the AmericanAcademyof Dermatology Volume37, Number 1
E L M criteria analysis 1. The pigment network, expression of pigmented and elongated fete ridges, in melanoma appears irregular, prominent, and broadened by the proliferation of atypical melanocytes disposed as solitary units along the dermoepidermal junction (atypical melanocytic hyperplasia) or arranged as irregular and confluent junctional nests. 2,3 Our data confirm that the pigment network appears in the early phases of the melanoma progression and is still observed in the majority of TkMs. Nevertheless in 5% of TnM cases and in 32% of TkM cases it is not visible, presumably because of a respective primary lack or loss of rete ridges. 11,14 We also observed that the pigment network irregularity could be less evident in the early phases of the tumor progression, probably because of the paucity and discontinuity of the cytoarchitectural atypias, 2 or when the melanoma arises in association with preexisting melanocytic nevus. 2. Black dots do not always assume the aspect of focal, round, black structures, but in our experience they could become confluent in greater and irregular accumulations or have a pulverulent appearance (numerous minute grains visible with greater magnification). Such structures, correlating with collections of "free" melanin or melanocytes in the cornifled layer,3,4 in melanomas could represent the atypical pagetoid melanocytic infiltration Of epidermis. Nevertheless, black dots were not always detected (present in 61% to 64% of our cases). This could be caused by the diverse capacity of the cells themselves to produce melanin. The first appearance of black dots occurs already during the in situ phase of the tumor. No differences were observed between the two groups of melanoma. 3. Radial streaming, pseudopods, irregular extensions, 3 and branched streaks 13 are morphologically dissimilar but all histologically correlate to confluent radial junctional nests of melanocytes. 12,17 Therefore we describe them as a single criterion. In this study we have chosen to use the term radial streaming because, in our opinion, this word better defines the histologic correlate of the structures. In contrast with other reports, 17 we observed radial streaming in the in situ melanomas (100% of cases). Moreover, they were more present in TnMs (83% TnM vs 71%
T~'Vl). 4. The brown globules are correlated in
Argenziano et al. 73
melanoma to nests of melanocytes irregularly distributed throughout the dermoepidermal junction or in the papillary dermis or both. 4 For this reason they appear in the radial growth phase of melanoma (78% of TnM). A similar number was found in TkM (71%). 5. Diffuse gray-black blotches are correlated to an intense melanic pigmentation throughout all levels of the epidermis (hyperpigmented keratinocytes or melanocytes) or upper cutis. 4 At this site the pigment could be related to melanophagic or melanocytic infiltration. Such a criterion is reported in 61% to 64% of the melanomas studied, without any difference between the two categories. 6. The gray-blue areas are observed when pigmented melanocytes or melanophages are located in the mid and reticular dermis. 4 Such a situation is found in 94% of TkMs, and we speculate that in these cases gray-blue areas are caused by the melanocytic infiltration of the mid-reticular dermis. When this criterion is found in TnM (24% of cases), we assume it is correlated to a deep melanophagic infiltration. In our experience the gray-blue areas observed in TkM show a greater density than that observed in TnM. 7. The white scarlike areas are histologically correlated to zones of melanoma regression. 4,12 The dermoepidermal pigmentation has disappeared, and in a fibroplastic dermis there is persistence of telangiectases and melanophages in the form of small pepperlike dots. 14,18 The regression pattern first appears in the radial growth phase of the tumor (56% of TnM cases) and is found in only 39% of TkM cases. 8. A vascular pattern is observed in the regressive areas but also in melanomas not heavily pigmented. 13 It can have various aspects; 15% to 19% of all melanomas observed had irregular, linear vessels (parallel to the skin surface) of upper dermis location, and 45% of TkM (and only 2% of TnM) had regular, dotted vessels (perpendicular to the skin surface) of mid-reticular dermis location. 1334 In some cases milky-red globules (reported by Stolz et al.13), as well as vascularized nests of amelanotic cells, were seen. In two of the three cases of melanoma with a thickness of more than 3 mm we noted the presence of a vascular pattern with elongated loops throughout the nodular component. 9. Structural asymmetry could be the ELM
74
Journal of the American Academy of Dermatology July 1997
Argenziano et al.
correlate o f the histologic architectural asymmetry of cutaneous melanoma. The first appearance o f the pattern is in the early phases of the tumor progression (76% of TnM), but it finds a greater incidence in T k M (87% o f cases); that fact is probably caused by coexistent radial and vertical growth patterns.
4.
5.
CONCLUSION The analysis of the results o f our pilot study shows a good correlation between the selected E L M criteria and the histologic architecture o f the cultaneous melanoma, for a preoperative evaluation of the tumor thickness. Pigment network and radial streaming indicate the radial growth phase of melanoma, whereas gray-blue areas and dotted vascular pattern (or the elongated loops) indicate the involvement of the mid and deep dermis and therefore the vertical growth phase o f the tumor. The lower frequency o f the regression pattern registered in T k M versus T n M (and in the opposite, the higher percentage of structural a s y m m e t r y in T k M ) c o u l d be explained by the emergence, in the latter, of more atypical cellular clones that have vertical growth and predominate the regression pattern. The gray-blue areas and the typical vascular pattern, individualized in our study as distinctive characteristics for the E L M diagnosis o f TkM, were both absent in only one case o f them (0.9 m m thick), whereas they were both present in only one case of TnM. Further investigation is needed to verify, on a larger number o f cases, our pilot estimates o f sensitivity and specificity of E L M criteria in thin and thick melanomas. REFERENCES
1. Barnhill RL, Fitzpatrick TB, Fandrey K, et al., editors. Color atlas and synopsis of pigmented lesions. New York: McGraw-Hill; 1995. p. 145-51. 2. Caputo R, Gelmetti C, Annessi G, editors. Pediatric dermatology and dermatopathology; vol. 3. Baltimore: Williams & Wilkins; 1995. p. 307-20. 3. Soyer HP, Smolle J, Hodl S, Pachernegg H, Kerl H.
6.
7.
8.
9.
10. 11.
12. 13. 14. 15.
16. 17. 18.
Surface microscopy: a new approach to the diagnosis of cutaneous pigmented tumors. Am J Dermatopathol 1989;11:1-10. Yadav S, Vossaert KA, Kopf AW, Silverman M, GrinJorgensen C. Histopathologic correlates of structures seen on dermoscopy (epiluminescence microscopy). Am J Dermatopathol 1993;15:297-305. Pehamberger H, Steiner A, Wolff K. In vivo epiluminescence microscopy of pigmented skin lesions. I. Pattern analysis of pigmented skin lesions. J Am Acad Dermatol 1987;17:571-83. Steiner A, Pehamberger H, Wolff K. In vivo epiluminescence microscopy of pigmented skin lesions. II. Diagnosis of small pigmented skin lesions and early detection of malignant melanoma. J Am Acad Dermatol 1987;17:584-91. Pehamberger H, Binder M, Steiner A, Wolff K. In vivo epiluminescencemicroscopy:improvement of early diagnosis of melanoma. J Invest Dermatol 1993;100(suppl): 356S-62S. Steiner A, Binder M, Schemper M, Wolff K, Pehamberger H. Statistical evaluation of epiluminescence microscopy criteria for melanocytic pigmented skin lesions. J Am Acad Dermatol 1993;29:581-8. Soyer HP, Smolle J, Leitinger G, Rieger E, Kerl H. Diagnostic reliability of dermoscopic criteria for detecting malignant melanoma. Dermatology 1995;190:2530. Nilles M, Boedeker RH, Schill WB. Surface microscopy of naevi and melanomas: clues to melanoma. Br J Dermatol 1994;130:349-55. Dummer W, Blaheta H-J, Bastian BC et al. Preoperative characterization of pigmented skin lesions by epiluminescence microscopy and high-frequency ultrasound. Arch Dermatol 1995;131:279-85. Bahmer FA, Fritsch P, Kreusch J, et al. Terminology in surface microscopy.J Am Acad Dermatol 1990;23:115962. Stolz W, Braun-Falco O, Bilek P, et al., editors. Color atlas of dermatoscopy. Oxford: Blackwell Science; 1994. Kreusch J, Rassner G, editors. Auflichtmikroskopie pigmentierter Hanttumoren: ein Bildatlas. Stuttgart: Thieme; 1991. Kenet RO, Kang S, Kenet BJ, et al. Clinical diagnosis of pigmented lesions using digital epiluminescence microscopy: grading protocol and atlas. Arch Dermatol 1993; 129:157-74. Dal Pozzo V, Benelli C, editors. Atlante di dermoscopia. Milano: Edra; 1996. Menzies SW, Crotty KA, McCarthy WH. The morphologic criteria of the pseudopod in surface microscopy. Arch Dermatol 1995;131:436-40. Menzies SW, Crotty KA, Ingvar C, et al., editors. An atlas of surface microscopy of pigmented skin lesions. Sydney, Australia: McGraw-Hill; 1996.