
	Version PDF

Vitiligo vs. hypopigmented mycosis fungoides (histopathological and immunohistochemical study, univariate analysis)

European Journal of Dermatology. Volume 16, Number 1, 17-22, January-February 2006, Investigative report

Summary

Author(s) : Mohammad A El-Darouti, Salonaz A Marzouk, Omar Azzam, Marwa Mohsen Fawzi, Mona RE Abdel-Halim, Amira A Zayed, Tahra M Leheta , Dermatology department, Faculty of Medicine, Cairo University Egypt.

Summary : Vitiligo is a common skin disease characterized by the presence of well circumscribed, depigmented milky white macules devoid of identifiable melanocytes. On the otherhand, hypopigmented mycosis fungoides (MF) is a rare variant of MF which presents clinically as persistent hypopigmented macules and patches. Both disorders show a predominance of CD8+ T cells in tissue samples and hence the differentiation between the two diseases on clinical, histopathological and even immunohistochemical grounds may offer great diffculty. The aim of this work is to identity certain histopathological clues which might help to differentiate between the two diseases. The study included 54 patients (26 vitiligo patients and 28 patients with Hypopigmented MF). Skin biopsies were taken and examined by hematoxylin and eosin and CD3, CD4 and CD8 markers were performed for ten vitiligo and nine MF patients. We have found that epidermotropism, hydropic degeneration of basal cells, partial loss of pigment, preservation of some melanocytes, presence of lymphocytes within the papillary dermis, increased density of the dermal infiltrate and wiry fibrosis of the papillary dermal collagen were detected with a significantly higher incidence in hypopigmented MF rather than vitiligo (P-values <\; 0.0001, <\; 0.00011, <\; 0.00011, \= 0.001, \= 0.008 and \= 0.001 respectively). On the other hand, focal thickening of the basement membrane, complete loss of pigmentation, total absence of melanocytes, as well as absence or sparsness of lymphocytes in the dermal papillae were seen much more frequently in vitiligo. Statistical analysis of these differences was significant with P-values <\; 0.00011, <\; 0.00011, <\; 0.00011, \= 0.008 respectively, regarding these pathological criteria. We conclude that differentiation of hypopigmented MF from vitiligo is possible by relying on the histopathological clues described in this study. This is particularly useful in areas of the world where cost benefit is crucial.

Keywords : epidermotropism, histopathology, hypopigmented MF, immunohistochemistry, melanocytes, vitiligo

Pictures

ARTICLE

Auteur(s) : Mohammad A El-Darouti, Salonaz A Marzouk, Omar Azzam, Marwa Mohsen Fawzi, Mona RE Abdel-Halim, Amira A Zayed, Tahra M Leheta

Dermatology department, Faculty of Medicine, Cairo University Egypt

accepté le 31 Août 2005

Although hypopigmented mycosis fungoides (MF) is one of the less common presentations of MF, awareness of this clinical entity is increasing. This variant is usually observed in dark-skinned individuals, especially children, and often shows a T suppressor CD8-positive phenotype [1]. Overall, the real frequency of hypopigmented MF is still unknown, however to our knowledge there are 113 patients reported in the literature up to August, 2003 [2]. In fact, the incidence of hypopigmented MF is probably underestimated since most of the patients described in literature had had their lesions for several years before the final diagnosis of MF was made [2, 3]. The diagnosis of hypopigmented MF may also be delayed due to the prolonged non aggressive course of the disease, although it is now confirmed that hypopigmented MF shows a course similar to that of “classic” MF [4].Vitiligo (especially vitiliginous lesions with an inflammatory border) may, in particular, be clinically confused with hypopigmented MF. Moreover, it may show striking histopathological similarities to MF [5]. Cribier and co-workers [5] reported four cases of vitiligo whose histopathologic studies showed features resembling MF, with remarkable exocytosis of lymphocytes. Furthermore, immunohistochemical and molecular investigations showed a predominance of CD8+ T lymphocytes infiltrating the epidermis. Vitiligo was diagnosed in these four cases primarily because of the absence of melanocytes beyond the inflammatory area [5], however, in their study, hypopigmented MF could not be ruled out completely. In another study, vitiligo-like inflammatory macules that simulated MF were reported [6]. The authors emphasized that erythematous, papular borders surrounding the hypopigmented macules, CD8+ lymphocytic infiltrate and absence of T-cell clonal rearrangement were helpful clues to rule out MF.In challenging cases, T cell receptor (TCR) gene rearrangement studies with polymerase chain reaction (PCR) can be a useful tool for detecting early cases of hypopigmented MF that may be confused with vitiligo [2, 7]. However, PCR is an expensive technique. Moreover, TCR gene rearrangement studies by either PCR or Southern blot methods, may show monoclonality in diseases that are generally regarded as benign or inflammatory in nature such as lymphomatoid papulosis, Mucha-Habermann disease, psoriasis [8] and also inflammatory stages of vitiligo [5]. In contrast, clinically obvious T cell tumors can sometimes lack demonstrable clonal TCR gene rearrangements by either Southern blot or PCR analysis [9].Therefore, this study was performed to look for possible histopathological clues that may differentiate between hypopigmented MF and vitiligo.

Patients and methods

Patients

The study included 54 patients (26 patients with active vitiligo and 28 patients with hypopigmented MF) who attended the Dermatology department’s outpatient clinic, Kasr al Aini Hospital, Cairo University, in the period between May 2000 to May 2004 (table 1( Table 1 )). All 54 patients were biopsied (after signing a written consent). A five mm skin biopsy was taken from every patient, fixed in formalin and stained with hematoxylin and eosin for histopathological examination. In 19 patients (10 having vitiligo and 9 having persumed MF), immunohistochemical staining with CD3 (Pan T), CD4 and CD8 was done.
Table 1 Clinical data of the patients

	Hypopigmented MF	Vitiligo
Number of cases	28	26
Sex	21 males, 7 females	13 males, 13 females
Age	23 ± 10 years	33 ± 6 years
Duration of lesions	24 ± 9 months	30 ± 7 months
Sites	Trunk and proximal extremities	Trunk and proximal extremities
Size of lesions	2-10 cm in diameter	2-10 cm in diameter
Scaling	Fine scaling in 18 cases	No scales
Color	Hypopigmented	Milky white to hypopigmented

Methods

Histopathological examination of hematoxylin and eosin sections was carried out by 3 independent investigators to assess separately the following points: epidermotropism (amount and pattern), hydropic degeneration at the dermoepidermal junction, thickening of basement membrane, melanocyte content and epidermal pigmentation. The dermis was assessed for the pattern of dermal infiltrate, density of the infiltrate and the presence or absence of wiry fibrosis in the dermal papillae (table 2( Table 2 )). Staining with PAS and S100 was performed at a later stage for better evaluation of both basement membrane thickening and melanocytes, respectively.

Immunohistochemical staining was carried out using monoclonal antibodies for CD3, CD4 and CD8 and results were assessed as either weakly positive (25-30%), positive (> 30%) or negative (< 25%).

The data were coded and entered on an IBM compatible personal computer using the statistical package SPSS version 12. The data were summarized using the mean and standard deviation as well as percentages for qualitative variables. Differences between groups were assessed using the Chi square test. Appropriate risk estimates were conducted to determine the association impact between probable risk factors and the occurrence of vitiligo or MF. The level of significance for all used tests was at P-value < 0.05 [9].
Table 2 Histopathologic criteria assessed in vitiligo and hypopigmented MF

Epidermis	Dermis
Epidermotropism (amount: absent, rare, few or many) (pattern: absent, focal or diffuse)	Pattern of infiltrate 1) Site: perivascular or band like 2) Lymphocytes in papillary dermis: absent, few or many
Hydropic degeneration of basal cells (absent, focal or diffuse)	Density of infiltrate (absent, sparse, moderate or dense)
Thickening of basement membrane (absent, focal or diffuse)	Wiry fibrosis (absent, focal or present)
Melanocyte content (absent, markedly reduced, reduced or preserved)
Pigmentation in the epidermis (absent, decreased or preserved)

Results

The study included 54 patients (26 patients with vitiligo and 28 patients with hypopigmented MF). The vitiligo group included 13 males (50%), 13 females (50%) and the hypopigmented MF group included 24 males (85.7%) and 4 females (14.3%).

The results of assessment of histopathological criteria of H&E sections of all our patients, as well as immunohistochemical markers, are outlined in table 3( Table 3 ).

Epidermotropic cells were seen more often in MF than in vitiligo and the difference was statistically significant (< 0.0001). The pattern of epidermotropism was focal in the form of singly arranged cells in vitiligo, while a diffuse pattern was seen in addition to the focal arrangement in MF biopsies and the difference was statistically significant (< 0.0001) (( figure 1 )).

Hydropic degeneration at the dermoepidermal junction was seen more in MF, while basement membrane thickening was more detected in vitiligo and the differences were statistically significant (< 0.0001 for both) (( figure 2 )).

Melanocytes and pigmentation were mostly absent to markedly reduced (> 75% lost) in vitiligo but only moderately reduced (< 50% lost) to present in MF (< 0.0001 for both) (( figure 3 )).

The dermal infiltrate was mostly perivascular in both vitiligo and MF, but a band-like pattern was seen mostly in MF (5 out of 6 cases with band like infiltrate were MF cases). However, the difference was not statistically significant (0.06). The dermal infiltrate was sparse to moderate in vitiligo and denser in MF, with a statistical difference between both groups (0.008). Papillary dermal wiry fibrous dysplasia was absent in vitiligo and present in MF, with a statistical difference between both groups (0.001).

Immunohistochemical markers showed that both vitiligo and MF were mostly CD8 positive. The difference between both groups as regards the markers, CD3, CD4, CD8 was not statistically significant (0.217, 0.539, 0.929 respectively) (( figure 4 )).

Due to the marked statistical difference between both groups as regards the histopathological findings, regrouping was done in order to increase the discriminating ability of these factors between both diseases in terms of odds ratio. Table 4( Table 4 ) outlines the data of vitiligo and MF patients after regrouping.

The presence of epidermotropism was found to favour the diagnosis of MF (risk estimate of 3, P-value between both groups < 0.0001). On the other hand, the presence of a diffuse pattern of epidermotropic cells was a weak discriminating factor and did not increase the chance of diagnosis of MF (risk estimate of 0.563, P-value 0.01 between both groups).

The absence of hydropic degeneration favored the diagnosis of vitiligo (risk estimate of 3.9 and P-value < 0.0001 between both groups), while basement membrane thickening favored the diagnosis of vitiligo (risk estimate of 3.231 and a P-value < 0.0001 between both groups).

The absence of melanocytes and pigmentation was more predictive for the diagnosis of vitiligo (risk estimate of 2.857 and a P-value < 0.0001 for the former; risk estimate of 3.333 and a P-value < 0.0001 for the latter).

The absence of dermal lymphocytes within the papillary dermis favored the diagnosis of vitiligo (risk estimate of 2.4, P-value 0.007), while the presence of wiry fibrosis favored the diagnosis of MF (risk estimate of 2.744 and a P-value < 0.0001).

Finally, the pattern of the dermal infiltrate, as well as the markers CD3, CD4, CD8 did not have any discriminating ability for the differentiation between both diseases (P-values 0.136, 0.330, 0.510, and 0.906 respectively).
Table 3 Results of assessment of the histopathologic criteria in both MF and vitiligo (H&E and immunohistochemistry)

		Vitiligo	MF	P-value
Epidermotropism (amount)	No	12		< 0.0001
Rare	6	6
Few	6	2
Many	2	20
Epidermotropism(pattern)	Absent	12		< 0.0001
Focal	14	18
Diffuse		10
Hydropic degeneration	Absent	21	7	< 0.0001
Focal	5	20
Diffuse		1
Basement membrane thickening	Absent	5	21	< 0.0001
Focal	18	7
Diffuse	3
Melanocytes	Absent	2		< 0.0001
Markedly reduced	11	1
Reduced	10	4
Present	3	23
Pigmentation	Absent	14		< 0.0001
Reduced	8	16
Preserved	4	12
Dermal infiltrate (pattern)	Absent	3		0.06 (ns)
Perivascular	22	23
Band-like	1	5
Dermal infiltrate (density)	Absent	1		0.008
Sparse	18	9
Moderate	6	9
Dense	1	10
Lymphocytes in papillary dermis	Absent	6		0.001
Few	16	11
Many	4	17
Wiry fibrosis	Absent	14	2	0.001
Focal	4	9
Diffuse	8	17
CD3	Negative	1		0.217 (ns)
Weak positive	5	2
Positive	4	7
CD4	Negative	8	6	0.539 (ns)
Weak positive	2	2
Positive		1
CD3	Negative	2	2	0.929 (ns)
Weak positive	3	2
Positive	5	5

Table 4 Histopathologic criteria after regrouping

		Vitiligo	MF	P-value	Risk estimate
Epidermotropism (amount)	Absent	12		< 0.0001	3
Present	14	28
Epidermotropism (pattern)	Focal	14	18	0.01	0.563
Diffuse		10
Hydropic Degeneration	Absent	21	7	< 0.0001	3.9
Present	5	21
BM Thickening	Absent	5	21	< 0.0001	3.231
Present	21	7
Melanocytes	Absent	13	1	< 0.0001	2.857
Present	13	27
Pigmentation	Absent	14		< 0.0001	3.333
Present	12	28
Dermal infiltrate (pattern)	Perivascular	22	23	0.136	2.933
Band-like	1	5
Lymphocytes in papillary dermis	Absent	6		0.007	2.4
Present	20	28
Dermal wiry Fibrosis	Absent	14	2	< 0.0001	2.7
Present	12	26
CD3	Negative	1		0.33	2
Positive	9	9
CD4	Negative	8	6	0.51	1.429
Positive	2	3
CD4	Negative	2	2	0.906	0.938
Positive	8	7

Discussion

Hypopigmented MF is most frequently seen in individuals with dark complexions particularly children and adolescents [1]. Of note, Egyptians have dark skin types, mainly skin types III and IV. Most patients of hypopigmented MF are misdiagnosed as having other disorders such as vitiligo, atopic dermatitis, chronic pityriasis lichenoides, post inflammatory hypopigmentation, tinea versicolor, leprosy and pityriasis alba [2, 3]. Moreover, not every case diagnosed as hypopigmented MF is actually mycosis fungoides [10]. Vitiligo is frequently confused with hypopigmented MF in patients with dark complexions. The confusion extends to involve the microscopical diagnosis as both disorders can show similar histopathological features [5].

The characteristic loss of pigmentation in vitiligo is due to absence of melanocytes and melanin pigment. Moreover, the normal appearing skin adjacent to vitiliginous areas is characterized by degenerative changes of the basal cells [12], mild vesiculation in the epidermis [13], dermal and epidermal infiltration of lymphocytes [14] and macrophages in the upper dermis [15]. These features, particularly the epidermal hypopigmentation and the presence of intraepidermal lymphocytes, are similar to those features seen in hypopigmented MF. Furthermore, immunohistochemical studies of perilesional skin in widespread vitiligo have detected CD4 and CD8 positive T cells in the dermal infiltrate (with an increased CD8/CD4 ratio), both types of T cells had expressed activation molecules as interleukin 2 receptor, HLA-DR and major histocompatibility complex class II (MHCII) [16]. The predominance of CD8 T cells indicates that hypopigmentation in vitiligo results from the cytotoxic effect of CD8 cells on melanocytes.

Similarly, the loss of pigmentation in hypopigmented MF, though still unclear [17], is probably due to the cytotoxic effect of CD8 cells on melanocytes (at least in cases showing a CD8 positive phenotype). However, it remains unclear whether non neoplastic CD8 positive lymphocytes play a role in the formation of hypopigmented lesions in T helper CD4 positive cases [2, 17-20].

Thus, the pathomechanism of hypopigmentation in hypopigmented MF may be similar to that of vitiligo [21-26] and it is noteworthy that Goudie in 1991 categorized vitiligo as a “benign cutaneous lymphoma” [27].

The histopathological features of vitiligo (especially when biopsies are taken from the periphery of the lesions) may show striking similarity to hypopigmented MF. Both disorders show a large number of lymphocytes within the epidermis [5]. Moreover both vitiligo and, at least some cases of hypopigmented MF, show a predominance of CD8 cells [11, 16]. CD8 positivity was detected in nine out of 15 hypopigmented MF patients in one study [11] and in 70% of hypopigmented MF patients in another study [28]. A third study reported the predominance of CD8 cells in three out of seven hypopigmented MF patients [2]. Similarly, in our study the predominance of CD8 cells was detected in seven out of nine patients with hypopigmented MF and in eight out of ten patients with vitiligo. Thus, it is clear that CD8 predominance can not differentiate between vitiligo and hypopigmented MF lesions.

It should be remembered that although monoclinality has been detected in the majority of hypopigmented MF lesions, several studies have reported cases of hypopigmented MF with negative TCR gene rearrangement [3, 29]. Furthermore, monoclonality is not synonomous with malignancy. Many benign dermatoses including inflammatory stages of vitiligo can harbour monoclonal T cell populations [5].

Thus, monoclonality alone is not always enough for the differentiation of hypopigmented MF from vitiligo. Moreover, monoclonal studies are expensive techniques and are not readily available in the developing countries. Therefore our study was an attempt to search for easier microscopical clues that can be discerned upon examination of H&E sections.

As shown in table 5( Table 5 ), we found that epidermotropism, hydropic degeneration of basal cells, partial loss of pigment, preservation of at least some of the melanocytes, presence of lymphocytes within the papillary dermis, increased density of the dermal infiltrate and wiry fibrosis of the papillary dermal collagen were detected with a significantly higher incidence in hypopigmented MF (p-values < 0.0001, < 0.00011, < 0.00011, = 0.001, = 0.008 and = 0.001 respectively). On the other hand focal thickening of the basement membrane, complete loss of pigmentation, total absence of melanocytes, as well as, absence or sparsness of lymphocytes in the dermal papillae were seen much more frequently in vitiligo. Statistical analysis of these differences was significant with P-values < 0.00011, < 0.00011, < 0.00011, = 0.008 respectively regarding these pathological criteria. Furthermore, after regrouping for better assessment of the discriminating ability of each pathological criterion, by logistic regression loss of melanocytes was found to be a predictor for the diagnosis of vitiligo and exclusion of MF, while epidermotropism was a predictor for the diagnosis of MF.

Immunohistochemical analysis of both vitiligo and hypopigmented MF samples for CD3, CD4 and CD8 showed no significant difference between both groups (p-value 0.217, 0.539, and 0.929 respectively). Accordingly it seems that immunohistochemistry cannot be used as a method to differentiate between vitiligo and hypopigmented MF.

CD8 predominance in our study (eight of ten vitiligo patients, and seven of nine hypopigmented MF patients) was consistent with findings of other groups [15], but as mentioned before, CD8 predominance can be seen in both vitiligo and hypopigmented MF [2, 11, 16, 28].

In the clinical evaluation of hypopigmented MF, most authors agree that erythematous lesions may coexist with the predominant hypopigmented lesions at the time of presentation or may develop at a later stage [2]. On the contrary, we detected no erythematous lesions in any of our patients, however, this seems a helpful clinical clue that can differentiate hypopigmented MF from vitiligo.

In conclusion, we are aware of the importance of immunophenotyping and TCR gene rearrangement for the accurate diagnosis of hypopigmented MF, however, based on our findings, we feel that differentiation of hypopigmented MF from vitiligo is possible by relying on the histopathological clues described in this study. This is particularly useful in areas of the world where cost benefit is crucial.
Table 5 Suggested diagnostic histopathological criteria for differentiation between MF and vitiligo

	Vitiligo	Hypopigmented MF
Melanocytes	Total loss	Partial (focal) loss
Pigment loss	Total loss	Partial loss
Hydropic degeneration	Rare	More frequent
BM thickening	More frequent	Rare
Lymphocytes in papillary dermis	Less common	More frequent
Dermal infiltrate	Less common (low density)	More frequent (high density)
Dermal wiry fibrosis	Less common	More frequent

References

1 El Shabrawi-Caelen L, Cerroni L, Mederios LJ, McCalmont TH. Hypopigmented mycosis fungoides. Frequent expression of a CD8+ T-cell phenotype. Am J Surg Pathol 2002; 26: 450.

2 Ardig M, Borroni G, Muscardin L, Kerl H, Cerroni L. Hypopigmented MF in Caucasian patients: A clinicopathologic study of 7 cases. J Am Acad Dermatol 2003; 49(2): 264.

3 Neuhaus IM, Ramos-Caro FA, Hassanein AM. Hypopigmented mycosis fungoides in childhood and adolescence. Pediatr Dermatol 2000; 17: 403.

4 Robert C, Moulonguet I, Baudot N, Flageul B, Dubertret L. Hypopigmented mycosis fungoides in a light-skinned women. Br J Dermatol 1998; 139: 341.

5 Cribier B, Snatinelli F, Lipsker D, Gosshans E. Vitiligo with inflammatory raised border mimicking mycosis fungoides. A clinicopathological study of 4 cases. Am J Dermatopathol 2000; 4: 349.

6 Petit T, Cribier B, Bagot M, Wechsler J. Inflammatory vitiligo-like macules that stimulate hypopigmented mycosis fungoides. Eur J Dermatol 2002; 13(4): 410.

7 Volkenandt M, Soyer HP, Cerronl L, Koch OM, Atzpodin J, Kerl H. Molecular detection of clone-specific DNA in hypopigmented lesions of a patient with early evolving mycosis fungoides. Br J Dermatol 1993; 128: 423.

8 Weiss L, Wood GS, Trela M, Warnke RA, Sklar J. T cell populations in lymphomatoid papulosis. Evidence of a lymphoproliferative origin for a clinically benign disease. N Engl J Med 1986; 315(8): 475.

9 Vega F, Luthra R, Medeiros LJ, Dunmire V, Lee SJ, Duvic M, et al. Clonal heterogeneity in mycosis fungoides and its relationship to clinical course. Blood 2002; 100: 3369.

10 Werner B, Brown S, Ackerman AB. Hypopigmented mycosis fungoides is not always mycosis fungoides. Am J Dermatopathol 2005; 27: 56.

11 Dawson B, Trap RG. Basic and clinical biostatistics. 3rd edition Lange medical books. McGraw Hill, 2001.

12 Bhawan J, Bhutani LK. Keratinocyte damage in vitiligo. J Cut Pathol 1983; 10: 207.

13 Breathnach AS, Bor S, Wyllie LMA. Electron microscopy of peripheral nerve terminals and marginal melanocytes in vitiligo. J Invest Dermatol 1966; 47: 125.

14 Hann S, Park YK, Lee KG. Epidermal changes in active vitiligo. J Dermatol 1992; 19: 217.

15 Le Poole IC, van den Wijngaard RMJGJ, Westerhof W, Dutrieux R, Das PK. Presence or absence of melanocytes in vitiligo lesions: an immunohistochemical investigation. J Invest Dermatol 1993; 100: 816.

16 Horn TD. Analysis of the lymphocytic infiltrate in a case of vitiligo. Am J Dermatopathol 1997; 19: 400.

17 Akaraphanth R, Douglass MC, Lim HW. Hypopigmented mycosis fungoides treatment and a 6-year follow-up of 9 patients. J Am Acad Dermatol 2000; 42: 33.

18 Zackheim HS, Epstein Jr. EH, Grekin DA, McNutt NS. Mycosis fungoides presenting as areas of hypopigmentation. J Am Acad Dermatol 1982; 6: 340.

19 Breathnach SM, McKee PH, Smith NP. Hypopigmented mycosis fungoides: report of five cases with ultrastructural observations. Br J Dermatol 1982; 106: 643.

20 Goldberg DI, Schinella RS, Kechijian P. Hypopigmented mycosis fungoides. Am J Dermatopathol 1986: 326.

21 Fogelman JP, Bolognia J, Heald P, McNiff JM. Destruction of a congenital nevus by halo-like phenomenon secondary to cutaneous T-cell lymphoma. J Cutan Pathol 1997; 24: 97.

22 Mimouni D, David M, Feinmesser M, Coire CI, Hodak E. Vitiligo-like leukoderma during photochemotherapy for mycosis fungoides. Br J Dermatol 2001; 145: 1008.

23 Boulac A, Grange F, Delfau-Larue MH, Dieng MT, Tortel MC, Avril MF, et al. Leucoderma associated with flares of erythrodermic cutaneous T cell lymphomas four cases. Br J Dermatol 2000; 143: 832.

24 Castanet J, Ortonne J. Pathophysiology of vitiligo. Clin Dermatol 1997; 15: 845.

25 Palmero B, Campanelli R, Garblli S, Mantovani S, Lantelme E, Brazzelli V, et al. Specific cytotoxic T lymphocyte responses against Melan-AMART1, tyrosinase and gp100 in vitiligo by the use of major histocompatibility complex/peptide tetramersthe role of cellular immunity in the etiopathogenesis of vitiligo. J Invest Dermatol 2000; 117: 326.

26 Lang KS, Caroli CC, Muhm A, Wernet D, Moris A, Schittek B, et al. HLA-A2 restricted, melanocyte-specfic CD8+ T lymphocytes detected in vitiligo patients are related to disease activity and are predominantly directed against MelanA/MART1. J Invest Dermatol 2001; 116: 891.

27 Goudie RB. Two views of the origin of vitiligo. Lancet 1991; 338: 316.

28 Wain EM, Orchard GE, Whittaker SJ, Spittle MSCMF, Russell-Jones R. Outcome in 34 patients with juvenile-onset mycosis fungoides clinical, immunophenotypic and molecular study. Cancer 2003; 98(10): 2282.

29 Di L-A, Marchesi L, Naldi L, Motta M, Cainelli T. A case of hypopigmented mycosis fungoides in a young Caucasian boy. Pediatr Dermatol 1997; 14(6): 449.