Histopathology

Early lesions of mycosis fungoides reveal in the vast majority of cases a patchy lichenoid or band-like infiltrate in an expanded papillary dermis (Fig. 2.23). The epidermis may be hyperplastic, normal or atrophic (Figs 2.22a and 2.24) (the epidermis is atrophic as a rule in poikilodermatous lesions). Small lymphocytes predominate and atypical cells can be observed only in a minority of cases. Epidermotropism of solitary lymphocytes is usually found (Fig. 2.21), but Darier’s nests (Pautrier’s microabscesses) are rare (Fig. 2.25). Useful diagnostic clues are the presence of epidermotropic lymphocytes with nuclei slightly larger than those of lymphocytes within the upper dermis (Figs 2.26), the presence of lymphocytes aligned along the basal layer of the epidermis (“basilar epidermotropism”) (Fig. 2.27), and the presence of many intraepidermal lymphocytes in areas with scant spongiosis (Fig. 2.28) [45, 48–52]. This last feature is referred to as “disproportionate” epidermotropism, because a few intraepidermal lymphocytes may be observed in areas of spongiosis in many inflammatory dermatoses. In this context, it should be emphasized that in a few cases (about 5% of the total) epidermotropism may be minimal or missing altogether (Fig. 2.29), most likely related to previous treatment (particularly if the diagnosis is not yet known, biopsies are often taken during therapy because of poor response of the lesions, and are submitted as “therapy-resistant dermatitis”) (see also Teaching case 2.3). Small foci of epidermotropism may be found in deeper sections in these cases. The papillary dermis shows a moderate to marked fibrosis with coarse bundles of collagen and a band-like or patchy-lichenoid infiltrate of lymphocytes. Dermal edema is usually not found. Eosinophils may be present (Fig. 2.22b), but they are not a common finding in patches and plaques of mycosis fungoides [53]. Langerhans cells are usually increased in the epidermis and dermis in early lesions of mycosis fungoides (see the section on immunophenotype in this chapter).

Unusual histopathologic patterns of mycosis fungoides in early phases include the presence of a perivascular (as opposed to band-like) superficial infiltrate (Fig. 2.25), prominent spongiosis (Fig. 2.30), sometimes with Darier’s nests (Pautrier’s microabscesses) mimicking spongiotic vesicles (Fig. 2.31), an interface dermatitis, sometimes with several necrotic keratinocytes similar to the picture of erythema multiforme (Figs 2.32 and 2.33), marked pigment incontinence with melanophages in the papillary dermis (see the section on hyperpigmented mycosis fungoides in this chapter), prominent epidermal hyperplasia like in lichen simplex chronicus (Fig. 2.34), prominent sclerosis of the papillary dermis mimicking lichen sclerosus (Fig. 2.35), and prominent extravasation of erythrocytes simulating the picture of lichenoid purpura (see the section on pigmented purpura-like mycosis fungoides in this chapter). A pattern characterized by a markedly flattened epidermis, a lichenoid infiltrate in the dermis, and increased, dilated vessels in the papillary dermis is the histopathologic counterpart of poikilodermatous mycosis fungoides (see the specific section in this chapter). In rare cases the histopathological features may resemble those observed in pityriasis lichenoides and varioliformis acuta (PLEVA) (see the specific section in this chapter). This last pattern is found more frequently in children.

Plaques of mycosis fungoides are characterized by a dense, band-like infiltrate of lymphocytes within the upper dermis (Fig. 2.36). Intraepidermal lymphocytes arranged in Darier’s nests (Pautrier’s microabscesses) are a common finding at this stage (Fig. 2.37). Cytomorphologically, small/medium pleomorphic (cerebriform) cells predominate (Fig. 2.38).

In tumors of mycosis fungoides a dense, nodular, or diffuse infiltrate is found within the entire dermis, usually involving the subcutaneous fat (Fig. 2.39). Epidermotropism may be lost. Flat tumors are characterized histopathologically by dense infiltrates confined to the superficial and mid parts of the dermis (Fig. 2.40). In some cases flat tumors of mycosis fungoides may present with a predominantly interstitial infiltrate (see the section on interstitial mycosis fungoides in this chapter). Prominent involvement of the subcutaneous fat, simulating a subcutaneous panniculitis-like T-cell lymphoma, may be observed in some cases (Fig. 2.41), as well as very rarely angiocentricity/angiodestruction (Fig. 2.42). It should be underlined that tumors of mycosis fungoides are morphologically indistinguishable from those of other types of primary or secondary cutaneous NK/T-cell lymphoma.

Large Cell Transformation

In advanced stages, patients with mycosis fungoides may develop lesions with many large cells (immunoblasts, large pleomorphic cells, or large anaplastic cells) (Fig. 2.43) [54, 55]. Large cell transformation in mycosis fungoides is defined as the presence of large cells exceeding 25% of the infiltrate or of large cells forming microscopic nodules, and has been detected in more than 50% of patients with tumor-stage mycosis fungoides [54]. Clusters of large cells may sometimes be observed in plaques of mycosis fungoides (usually in patients having tumors at other sites of the body), and rarely even in thin patches of the disease (these patients, too, usually also have plaques and tumors of mycosis fungoides at other sites of the body) (see also Teaching case 2.2 in this chapter) (Fig. 2.44).

Tumors with a large cell morphology may or may not express CD30. Negativity for CD30 has been related to a poor prognosis in transformed mycosis fungoides [56]. Spontaneous regression of some lesions and/or CD30 expression by neoplastic cells may suggest the diagnosis of anaplastic large cell lymphoma or lymphomatoid papulosis in these patients, but such a diagnosis should be made only upon compelling evidence in patients with known mycosis fungoides. In this context, most cases of mycosis fungoides with CD30⁺ large cell transformation are negative for IRF4 translocations, whereas primary cutaneous large cell anaplastic lymphomas are constantly positive [57–59]. In addition, the percentage of CD30⁺ cells rarely reaches the 75% mark necessary for a diagnosis of anaplastic large cell lymphoma [56, 59a]. Although lymphomatoid papulosis may be associated with mycosis fungoides, the highest degree of caution should be exercised in making this diagnosis when mycosis fungoides precedes the putative lymphomatoid papulosis. In fact, one patient with mycosis fungoides whom we published several years ago as “PUVA-induced lymphomatoid papulosis in mycosis fungoides” [60] turned out to have a CD30⁺ large cell transformation with bad prognosis, and similar patients have been described as well [61, 62]. Large cell transformation of mycosis fungoides bears a poor prognosis and usually heralds the terminal stage of the disease.

Immunophenotype

Mycosis fungoides is characterized by an infiltrate of α/β T-helper memory lymphocytes (β F1⁺, TCR-γ⁻, CD3⁺, CD4⁺, CD5⁺, CD8⁻, CD45Ro⁺, TIA-1⁻) (Fig. 2.45). Only a minority of cases exhibit a T-cytotoxic (β F1⁺, TCR-γ⁻, CD3⁺, CD4⁻, CD5⁺, CD8⁺, TIA-1⁺) (Fig. 2.46) or γ/δ (β F1⁻, TCR-γ⁺, CD3⁺, CD4⁻, CD5⁺, CD8⁺, TIA-1⁺) T-cell phenotype (Fig. 2.47), which show no clinical and/or prognostic differences [63, 64]. In these cases, correlation with the clinical features is crucial, in order to rule out skin involvement by aggressive cytotoxic lymphomas such as cutaneous aggressive epidermotropic CD8⁺ cytotoxic T-cell lymphoma or cutaneous γ/δ T-cell lymphoma (see also Chapter 6).

In late stages there may be a (partial) loss of pan-T-cell antigens (CD2, CD3, CD5) and or CD4/CD8 expression (Fig. 2.48). Loss of pan-T-cell markers is a helpful sign for the diagnosis of mycosis fungoides (and of T-cell lymphomas in general), but in my experience it is very rare in early lesions of the disease. Rarely, early mycosis fungoides displays an aberrant CD4⁺/CD8⁺ or CD4⁻/CD8⁻ phenotype. Double negative cases may be positive for PD-1 [65]. Although this molecule is considered as a marker of follicular helper T-cells (T_FH), in my opinion and that of others [66] it is not specific for these cells. In general, PD-1 is negative in mycosis fungoides [67], but rare positive cases showing the complete phenotype of T_FH cells (PD-1⁺, Bcl-6⁺, CXCL13⁺, CD10⁺) have been described [68, 69].

Cytotoxic-associated markers such as T-cell intracellular antigen (TIA)-1, granzyme B, and perforin are negative in mycosis fungoides (with the obvious exception of cases with cytotoxic phenotype). In cases that displayed previously a conventional T-helper phenotype, sometimes biopsies of tumor lesions may reveal positivity for TIA-1 [70]. These cases should not be classified as cytotoxic lymphoma, but as tumor-stage mycosis fungoides with a cytotoxic phenotype. Cases with a cytotoxic phenotype may also express CD56 (Fig. 2.49), but CD56-positivity has been observed rarely also in specimens of otherwise conventional CD4⁺ mycosis fungoides [71].

As already mentioned, in plaque and tumor lesions (and rarely focally in patches, too) neoplastic T-cells may express the CD30 antigen (Fig. 2.50). It has been suggested that a higher expression of CD30 in nontransformed mycosis fungoides is associated with a worse prognosis [72]. On the other hand, CD30-negativity has been linked with a worse prognosis in transformed mycosis fungoides [56]. In my opinion, it is conceptually difficult to conceive of a marker that is associated with both good and bad prognosis depending only on the type of lesion where it is detected (transformed versus nontransformed, respectively), and I think that the prognostic value of CD30 expression (or lack of expression) should be critically considered.

Aberrant expression of the B-cell antigen CD20 by neoplastic T lymphocytes has been observed exceptionally [73] and seems to be more frequent in large cell transformation [73a]. In these cases, negativity for other B-cell markers such as CD79a or PAX-5 and positivity for T-cell markers helps in determining the exact lineage of the cells. Multiple myeloma-1/interferon regulatory factor-4 (MUM1/IRF4), a member of the interferon regulatory factor (IRF) family of transcription factors, is positive in cutaneous CD30⁺ lymphoproliferative disorders and in CD30⁺ large cell transformation of mycosis fungoides, but it is otherwise negative in mycosis fungoides.

T-regulatory (Treg) cells comprise a T-cell population that can influence other cell types with suppression of the immune response, and that shows mostly a CD4⁺/CD25⁺/FOX-P3⁺ phenotype. They are present in the reactive infiltrate of lesions of early mycosis fungoides but tend to disappear in the tumor stage, and in earlier reports were thought to be related to prognosis. However, studying sequential lesions of patients with mycosis fungoides we could not find a significant difference in the percentage of Treg cells between patch/plaque and tumor stage biopsies, nor any association with prognosis [74]. Additionally, neoplastic lymphocytes with the phenotype of Treg cells can be observed in mycosis fungoides (Fig. 2.51) [74, 75]. It has been suggested that the number of intraepidermal FOX-P3⁺ cells is higher in inflammatory conditions than in mycosis fungoides [76], but I would exert the greatest caution in using this criterion for the differential diagnosis of early mycosis fungoides.

An increased number of CD1a⁺ Langerhans cells as well as of other, mostly immature dendritic, cells can be observed in mycosis fungoides, particularly in early lesions, sometimes simulating the picture of a Langerhans cell histiocytosis (Fig. 2.52) [77, 78]. I have come across cases that, besides the typical intraepidermal Darier’s nests (Pautrier’s microabscesses), also revealed intraepidermal collections of Langerhans cells mimicking the histopathologic features of Langerhans cell histiocytosis (Fig. 2.53). These clusters represent most likely a reactive hyperplasia of Langerhans cells. In this context, it should be remembered that the occurrence of clonally related dendritic or histiocytic tumors has been observed rarely in mature B-cell neoplasms (so-called “transdifferentiation” – see Chapter 1), but has never been described in mycosis fungoides. Although “Langerhans cell microabscesses” are considered as a differential diagnostic tool for lymphomatoid contact dermatitis (see Chapter 26), they can be rarely observed in mycosis fungoides as well.

Lesions of mycosis fungoides, particularly in advanced stages, may show a prominent amount of reactive CD20⁺ B lymphocytes, even forming germinal centers (Fig. 2.54) [79–81]. The B lymphocytes may be prominent and mask the true T-cell nature of the neoplastic infiltrate. These cases should not be misinterpreted as B-cell lymphoma. On the other hand, it should be kept in mind that exceptional cases of composite lymphoma (mycosis fungoides and a primary cutaneous B-cell lymphoma within the same lesion) have been observed [43]. B-cell lymphomas composite with mycosis fungoides are mostly cases of B-cell chronic lymphocytic leukemia (B-CLL) colonizing the lesions of mycosis fungoides (see Teaching case 19.1 in Chapter 19) [82, 83]. As already mentioned, CD20 may also be expressed aberrantly by neoplastic T lymphocytes, thus underlining the need for complete phenotypic analyses of every case.

Molecular Genetics

There is a considerable body of evidence showing that epigenetic modifications such as aberrant gene methylation and histone deacetylation are involved in the pathogenesis of mycosis fungoides. Dysregulation of the pathway involving the nuclear factor κ-light-chain enhancer of activated B-cells (NF-κB) has been demonstrated in several studies [84, 85]. Downregulation of NFKBIZ has been suggested as the molecular basis for enhanced NF-κB activity [85].

Several studies have addressed early genetic events in mycosis fungoides [86, 87], but results have been sometimes contradictory [88, 89]. In this context, it should be underlined that molecular analysis of early lesions of mycosis fungoides is hindered by the intrinsic difficulties in the identification of neoplastic cells, which are a small minority and may look morphologically and phenotypically identical to reactive lymphocytes. As genetic techniques, especially on routinely fixed specimens, require a certain threshold of tumor DNA, it may be difficult to analyze early lesions of mycosis fungoides where only a few neoplastic cells can be observed. A study on early lesions showed that many genes belonging to pathways associated with inflammation, regulation of apoptosis, and immune activation are upregulated in both early mycosis fungoides and inflammatory disorders [17]. Only a few genes had a discriminatory diagnostic value, and among these TOX (a transcription factor) and PDCD1 (a pro-apoptosis regulator) were considered to be more helpful for identification of early lesions of the disease [17].

Using cDNA microarray analysis, a signature of 27 genes, including oncogenes and other genes involved in the control of apoptosis, has been identified in cases of early- and late-stage mycosis fungoides, and a six-gene prediction model capable of distinguishing mycosis fungoides from inflammatory diseases has been proposed, including FJX1, Hs.127160, STAT4, SYNE-1B, TRAF1, and BIRC3 [90]. In a similar study, consensus clustering revealed the presence of two clusters that tended to include mostly patients with mycosis fungoides in the early stages (Ia and Ib), and of a third cluster limited mostly to patients with more advanced disease [91]. However, the number of cases was relatively small and overlapping data were observed. Using reverse transcription-PCR, 593 genes overexpressed in mycosis fungoides were identified, allowing stratification of the disease in three different prognostic groups [92].

Oncogenes such as p16 and p53 do not show alterations in early lesions, but are often mutated in late (tumor) phases of the disease [93–95]. These results are in keeping with the demonstration that Twist, a transcription factor blocking p53 and inhibiting c-myc-induced apoptosis, is overexpressed only in biopsies from advanced stages [96]. Dysregulated expression of JUNB and JUND has been found in some cases [97] and FAS mutations in others [98]. Recurrent deletions of tumor suppressor genes BCL7A, SMAC/DIABLO, and RHOF have been detected by comparative genome hybridization in cases of early mycosis fungoides [99].

Chromosomal aberrations observed in large cell transformation of mycosis fungoides include alterations of chromosomes 1, 2, 7, 9, 10, 17, and 19, the most common imbalances being gain of chromosome regions 1p36, 1q25-31, 7p22-11.2, 7q21, 9q34, 17q12, 17q24-qter, 19, and loss of 2q36-qter, 9p21, 10p11.2, 10q26, and 17p [34,100]. Loss of heterozygosity but not microsatellite instability is associated with tumor progression [101–103].

Several studies focused on micro RNA (miRNA) expression profiling; miR-92a and miR-155 are upregulated and miR-203 and miR-205 are downregulated in mycosis fungoides as compared to inflammatory disorders [104–106]. Different miRNA expression profiles have also been found in mycosis fungoides when compared with cutaneous anaplastic large cell lymphoma [107], representing a potential tool in the differential diagnosis of CD30⁺ tumors arising in patients with mycosis fungoides.

A monoclonal rearrangement of the T-cell receptor (TCR) genes is commonly found in plaques and tumors of mycosis fungoides. Using a sensitive technique (Biomed-2) the majority (∼80%) of patch lesions also show monoclonality of T lymphocytes. Demonstration of T-cell clonality at the first diagnostic biopsy does not affect the prognosis of patients with early mycosis fungoides [63]. Development of “patient-specific” DNA probes can identify the neoplastic clone also in lesions that are not specific histopathologically [108], and may be a useful method (though not routinely available) for detection of residual disease.

The presence of a monoclonal population of T lymphocytes has been detected in the peripheral blood in patients with early-stage mycosis fungoides. In many of these patients the clone was different from that detected within the skin, but in some cases the same clone was present both in the peripheral blood and in the cutaneous lesions of mycosis fungoides, even after successful treatment and complete clinical remission [26]. The exact diagnostic and prognostic value of molecular genetic analysis of the TCR genes rearrangement within the peripheral blood in patients with early mycosis fungoides is still unclear. Detection of monoclonality within lymph nodes that are not affected clinically and/or histopathologically has been associated with a worse prognosis [109]. The analysis of the rearrangement of TCR genes in the peripheral blood and/or uninvolved lymph nodes is not part of routine investigations in early mycosis fungoides, but the determination of clonality is paramount in lymph nodes showing histopathologically the picture of dermopathic lymphadenopathy.

The “Parapsoriases” (Small- and Large-Plaque Parapsoriasis)

In 1902 Louis-Anne-Jean Brocq, a brilliant French dermatologist, introduced a concept and terminology that, in his eyes, were only provisional. He probably never dreamed that over 100 years after his publication experts of cutaneous lymphomas worldwide would still argue on what is parapsoriasis, and what its relationship is to mycosis fungoides. Besides semantic issues (the term parapsoriasis conveys a wrong sense of relationship to psoriasis – but mycosis fungoides, after all, is also a misnomer), the exact nosology of the parapsoriases (small- and large-plaque) is still a matter of discussion. In this context, the term “plaque” is also wrong, as we are referring to lesions that clinically are exclusively patches. Interestingly, already in 1938 Dr Harry Keil, an argute dermatologist practicing in New York City, wrote: “Parapsoriasis en plaques disséminées, parapsoriasis lichenoides of French authors, parakeratosis variegata, and the retiform type of parapsoriasis with their evolution into poikiloderma probably represent different phases of a single disease, which invariably progresses to mycosis fungoides. (…) The nomenclature of parapsoriasis, mycosis fungoides and poikiloderma vascolare atrophicans should be revised in the light of collected data.”

In my opinion, there is no doubt that large-plaque parapsoriasis belongs to the clinical spectrum of mycosis fungoides and cannot (and should not) be distinguished from other early manifestations of the disease (Fig. 2.55). Thus, in what follows I will only concentrate on small-plaque parapsoriasis, as this is a much more slippery subject. At the beginning, however, it should be underlined that distinction of “small-plaque” from “large-plaque” parapsoriasis is a matter of discussion, too, as (i) it only relies on the size of lesions (and size was never defined precisely) and (ii) often both small and large lesions are present in one and the same patient. Besides, it is well known that small patches of disease are common in patients with otherwise “conventional” mycosis fungoides. I refer to small-plaque parapsoriasis as a disease presenting exclusively with small patches (not larger than a few centimeters). Numbers defining exactly the size of the patches would only add a false sense of precision that in my opinion does not exist.

Patients presenting only with small, sometimes “digitated” patches, typically located on the trunk and upper extremities (Fig. 2.56), are variously diagnosed as “digitate” dermatosis, chronic superficial scaly dermatitis, small-plaque parapsoriasis, or mycosis fungoides. Molecular genetic techniques reveals that in some of these lesions a monoclonal population of T lymphocytes can be found. As already mentioned, the exact relationship between small-plaque parapsoriasis and mycosis fungoides is still a matter of debate: some think that they represent one and the same disease, whereas others maintain that they are completely unrelated [113, 114]. In Graz we have observed patients with small patches of “parapsoriasis en plaques” who on follow-up developed typical plaques and tumors of mycosis fungoides, leading us to conclude that small-plaque parapsoriasis represents an early manifestation of the disease. A similar observation has been reported by Väkevä and co-workers, who observed that 10% of their patients with small-plaque parapsoriasis developed mycosis fungoides during the course of the disease [115], and by Belousova and co-workers [116].

In the context of the discussion on small-plaque parapsoriasis, it should be noted that the boundary between “overt” cancer, “early” cancer, and “precursor” of cancer is getting more and more blurred by new molecular insights (see also Chapter 1). In this context, “small-plaque parapsoriasis” is as good (or as bad) a term as any other used in the field of “early” or “pre-malignant” lymphoproliferative disorders (e.g., monoclonal gammopathy of undetermined significance, monoclonal B-cell lymphocytosis, etc.), and can be used in order to avoid a diagnosis of “mycosis fungoides” for patients who will likely experience no serious problems from the disease. Regardless of the academic discussion, it is important to underline that patients with small-plaque parapsoriasis should not be treated aggressively, as progression of the disease is rare and, when it happens, takes place only after very long periods of time.

Once again, it must be clearly stated that “large-plaque parapsoriasis” represents one of the most typical early manifestations of “conventional” mycosis fungoides.

Pilotropic (Folliculotropic) Mycosis Fungoides (with or Without Follicular Mucinosis) and Relationship with So-Called “Benign” Follicular Mucinosis

Some patients with mycosis fungoides present with follicular papules and plaques characterized histopathologically by abundant deposits of mucin within hair follicles that are surrounded and infiltrated by T lymphocytes (Figs 2.57 and 2.58) [117]. Clinically, sometimes only small areas of patchy alopecia are visible (Fig. 2.59). At body sites with only vellus hairs, lesions may be very subtle, and it may be difficult to appreciate the extent of the disease (Fig. 2.60). Tiny keratotic spines may protrude from affected follicles and confer on the lesion some resemblance to lichen spinulosus (Fig. 2.61). In several cases the prominent follicular distribution of the lesions without involvement of the interfollicular epidermis produces a characteristic appearance of “agminated” lesions (Fig. 2.62). Rarely, pilotropic mycosis fungoides presents with solitary lesions [118].

Histopathologically, the hair follicles are infiltrated by neoplastic lymphocytes (“pilotropism”) (Fig. 2.63). The number of intraepithelial lymphocytes is variable. Large deposition of mucin may disrupt the follicle and make it more difficult to appreciate the lymphocytes. In later lesions, a prominent granulomatous reaction may rarely replace the entire follicle (Fig. 2.64). The onset of pustules, mimicking an inflammatory condition, is much more frequent in pilotropic mycosis fungoides than in other variants of the disease (Fig. 2.65). The epidermis is usually spared, but in rare cases a conventional picture of epidermotropic mycosis fungoides may be observed between affected follicles (Fig. 2.66). Eosinophils are present as a rule in cases characterized by deposition of mucin and may occasionally be a prominent feature (Fig. 2.67). Itching is severe and represents a major problem in this variant of mycosis fungoides, and may be nonresponsive to standard treatments. Prurigo-like lesions caused by scratching may complicate the clinical picture and lead to the erroneous diagnosis of atopic or other eczematous dermatitis (Fig. 2.68).

Deposition of mucin is found only in a proportion of cases of pilotropic mycosis fungoides (Fig. 2.69). In some patients follicular mucinosis may be observed in one biopsy but not in other specimens, confirming that pilotropic mycosis fungoides represents a spectrum characterized by prominent involvement of the hair follicles with or without mucin deposition [119, 120]. When present, deposition of mucin is usually restricted to the hair follicle, but I have observed cases with intraepidermal deposition as well (Fig. 2.70). Particularly in advanced stages, the typical histopathologic features of pilotropic mycosis fungoides tend to disappear and conventional tumors may be observed.

As already mentioned, alopecia due to destruction of the follicles is common (so-called alopecia mucinosa), either generalized or localized. The pattern of alopecia in mycosis fungoides is often distinctive and is characterized by irregular areas of hair loss, patchy erythema, and follicular papules and/or comedones (Fig. 2.71). In some patients alopecia areata-like lesions or even alopecia universalis may be observed [121]. Sometimes alopecia can be present in patients who do not have other lesions of follicular mycosis fungoides, but it is most prominent in the pilotropic variant of the disease [121].

In patients with marked involvement of the hair follicles, with or without deposition of mucin, a localized or rarely widespread eruption of small infundibular cysts and/or comedones infiltrated by neoplastic T lymphocytes can be observed (“mycosis fungoides with eruptive cysts and comedones”) (Figs 2.72, 2.73, and 2.74) [122]. The clinical picture is similar to that observed in so-called “milia en plaques”, a condition that besides mycosis fungoides can be associated to several inflammatory disorders. Prominent involvement of the hair follicles with follicular hyperplasia may lead to the formation of elevated lesions, clinically resembling tumors of the disease, even in the absence of true neoplastic tumors (“pseudotumoral mycosis fungoides”) [123].

There is no consensus on the exact nosology of so-called “idiopathic generalized follicular mucinosis” [117], and some authors do not believe that it represents a manifestation of mycosis fungoides. The term “pilotropic T-cell dyscrasia” has been introduced for this condition [124]. However, cases of “idiopathic generalized follicular mucinosis” with progression to advanced mycosis fungoides and death have been well documented (Fig. 2.75) [117]. I have observed patients with clear-cut pilotropic mycosis fungoides who entered clinical remission after conventional treatments, and who subsequently relapsed with skin lesions showing clinical and histopathologic features of “idiopathic” follicular mucinosis. Interestingly, generalized follicular mucinosis may rarely present with hypopigmented lesions, indistinguishable from hypopigmented mycosis fungoides (Fig. 2.76). In my opinion, “idiopathic generalized follicular mucinosis” represents a variant of mycosis fungoides and association with the disease has been observed even in children [125, 126]. The discussion is not only academic, however, as some authors reported a worse prognosis in patients with pilotropic mycosis fungoides as compared to those with “conventional” variants of the disease [127–130]. However, patients with idiopathic generalized follicular mucinosis were not included in these studies, thus introducing a selection bias. Other reports revealed a prognosis of pilotropic mycosis fungoides similar to that of the conventional variant of the disease [131].

The nosology of “benign” localized follicular mucinosis is unclear. Patients are usually children or young adults, presenting with a single infiltrated patch on the face and alopecia in the affected region (Fig. 2.77). Although “benign” localized follicular mucinosis may be compared conceptually to a “solitary” variant of mycosis fungoides with good prognosis, I have never observed progression to advanced stages of the disease. On the other hand, the clinicopathologic features are indistinguishable from those of pilotropic mycosis fungoides and monoclonality of T lymphocytes can be demonstrated in about half of the cases [117]. It has been suggested that current diagnostic criteria for early mycosis fungoides should not be applied to children with alopecia mucinosa [132], but I would agree only for those presenting with solitary lesions of the disease, as pediatric patients with generalized involvement of the skin in my opinion have mycosis fungoides. For practical purposes, I do not use the term “mycosis fungoides” for cases characterized by solitary lesions in children and adolescents. Management should be conservative and spontaneous resolution may occur.

The last variant of follicular mucinosis, the so-called acneiform type, is very rare. It is characterized by follicular, erythematous papules on the head and neck area and on the upper trunk (Fig. 2.78). In this type, too, the relationship to mycosis fungoides is unclear, but a follow-up study showed an invariably benign course [132a]. I have seen only a few patients, and none of them developed clear-cut mycosis fungoides. However, “acneiform” follicular lesions have been described in patients with mycosis fungoides [127], and a monoclonal T-cell population may be observed in the infiltrate [133].

Finally, it should be mentioned that deposition of mucin within the hair follicle is not synonymous with alopecia mucinosa/follicular mucinosis, and it may be an incidental finding in biopsy specimens of different conditions (Fig. 2.79) [133a].

Syringotropic Mycosis Fungoides

Some patients present with unusual lesions consisting of dense lymphoid infiltrates located mainly around hyperplastic eccrine glands and coils, often with syringometaplasia (Fig. 2.80) [134, 135]. The hair follicles are usually involved as well and are surrounded by dense lymphoid infiltrates with pilotropism. A concomitant band-like infiltrate in the papillary dermis is observed in the majority of the specimens. Particularly in cases without band-like infiltrate, however, the histopathologic diagnosis may be very difficult (Fig. 2.81). Clinically, in many cases distinction from pilotropic mycosis fungoides is not possible, and lesions may show a similar follicular plugging with agminated arrangement (Figs 2.82 and 2.83). In fact, syringotropic and pilotropic mycosis fungoides are closely related and should probably be regarded as a unique type of “adnexotropic” mycosis fungoides. Lesions may be generalized or solitary. Palms and soles are involved more commonly than in other variants of the disease (Fig. 2.84). In some cases, conventional patches of mycosis fungoides are present at other body sites.

The term “syringolymphoid hyperplasia with alopecia” has been used in the past to refer to these cases, but the relationship to mycosis fungoides has been well recognized [134].

Localized Pagetoid Reticulosis (Woringer–Kolopp Type)

Localized pagetoid reticulosis is a variant of mycosis fungoides presenting with psoriasiform, scaly erythematous patches or plaques, usually located on the extremities (Fig. 2.85). The lesions are either solitary or localized to one acral site. The clinical presentation can be deceptive, simulating that of benign conditions such as warts or eczematous dermatitis (Fig. 2.86). The histologic picture shows a markedly hyperplastic epidermis with striking epidermotropism of medium-sized pleomorphic lymphocytes (Fig. 2.87). Neoplastic lymphocytes usually have a T-cytotoxic phenotype (Fig. 2.88a), but a helper phenotype or double negativity for both CD4 and CD8 have been reported [136]. The lymphocytes may show a tropism for adnexal structures and grow along the eccrine coils (Fig. 2.88b), thus cautioning against superficial treatment modalities.

The term “pagetoid reticulosis” should be restricted to solitary or localized lesions only (Woringer–Kolopp type). Most patients presenting with the “generalized” form of pagetoid reticulosis (Ketron–Goodman type) have one of the cutaneous aggressive cytotoxic NK/T-cell lymphomas (cutaneous aggressive epidermotropic CD8⁺ cytotoxic T-cell lymphoma, cutaneous γ/δ T-cell lymphoma, extranodal NK/T-cell lymphoma, nasal type) (see Chapter 6).

It should be emphasized that a histopathologic picture characterized by markedly hyperplastic epidermis with striking epidermotropism of innumerable T lymphocytes is not restricted to pagetoid reticulosis or aggressive cytotoxic NK/T-cell lymphomas: other lymphomas may rarely present with similar features and a definitive diagnosis should be made only upon compelling clinicopathologic evidence and after complete phenotypic and genotypic studies [137]. Lymphomatoid papulosis, type D may be indistinguishable from pagetoid reticulosis histopathologically, but shows the typical waxing and waning course clinically, and usually does not arise on acral sites (see Chapter 4). Presence of intraepidermal CD30⁺ neoplastic cells in “pagetoid reticulosis” should raise suspicion of such a variant of lymphomatoid papulosis; a precise classification of these cases can be achieved only upon clinicopathologic correlation.

The prognosis of patients with solitary pagetoid reticulosis is excellent and involvement of internal organs has never been observed. Development of conventional mycosis fungoides has been documented in a few patients. The treatment of choice is surgical excision or local radiotherapy. Other skin-directed therapies may be used for more extensive lesions, including topical steroids, topical nitrogen mustard, PUVA, and narrow-band UV-B irradiation [138], but superficial treatment modalities may be ineffective in cases showing deeper extension along the adnexal structures (see Fig. 2.88).

Unilesional (“Solitary”) Mycosis Fungoides

Besides localized pagetoid reticulosis, a solitary variant of mycosis fungoides with clinicopathologic features similar to “common” mycosis fungoides has been described [139–141]. The literature about this subject is confusing, as in the past disparate cases were lumped together in this category, including pagetoid reticulosis and “mycosis fungoides en tumeur d’emblée” among others. In addition, many of the cases reported in the literature probably represent in truth examples of lichenoid keratosis or other pseudolymphomas (including some of the cases published by our group – see the section on solitary idiopathic B/T-cell pseudolymphoma in Chapter 26) [139], but I have rarely observed patients with solitary lesions of genuine mycosis fungoides (Fig. 2.89). Histopathological features are identical to those of conventional mycosis fungoides. Rarely, large cell transformation may be observed in solitary lesions, representing a sign of disease’s progression (see Teaching case 2.2 in this chapter).

It is yet unclear whether solitary mycosis fungoides has a better prognosis [142], but in some instances development of generalized lesions of the disease has been observed over time [143]. As in other variants of the disease, the presence of large cell transformation should be regarded as a bad prognostic sign.

Granulomatous Mycosis Fungoides

Granulomatous mycosis fungoides is an unusual histologic variant of the disease described first by Ackerman and Flaxman in 1970, who reported a patient with tumor-stage mycosis fungoides with histiocytic giant cells scattered within the dermal infiltrate [144].

Granulomatous lesions may either precede, be concomitant with, or follow “classic” mycosis fungoides; they can be observed in patches, plaques, and tumors of the disease, and in some cases even within affected lymph nodes [145]. Clinically there may be irregular areas of induration (Fig. 2.90), but in most cases the clinical aspect does not betray the presence of granulomas histopathologically. In my experience a granulomatous reaction is observed mostly in tumors of mycosis fungoides, and sometimes may be secondary to treatment. Granulomas can also be observed in the context of pilotropic mycosis fungoides, due to destruction of hair follicles (see the section on pilotropic mycosis fungoides in this chapter). Especially if the first manifestation of the disease shows prominent granulomatous features, the diagnosis of mycosis fungoides may be missed, and the histopathologic picture may be misinterpreted as that of a “granulomatous dermatitis.” Epidermotropism is a helpful clue (Fig. 2.91), but it is often minimal or missing in granulomatous tumors (Fig. 2.92).

Histopathologically, granulomatous mycosis fungoides may also be difficult or even impossible to distinguish from other cutaneous T-cell lymphomas with granulomatous features (e.g., CD4⁺ small/medium T-cell lymphoma) [146], and clinicopathologic correlation is crucial for the diagnosis.

Granulomatous mycosis fungoides seems to have a worse prognosis than that of conventional variants of the disease, and patients have a higher risk of developing a second lymphoma [147]. On the other hand, the amount of the granulomatous reaction is often variable in different biopsies from a single patient, suggesting that the term “granulomatous mycosis fungoides” should be used as a description of histopathologic specimens, and not for classifying the disease in a given patient. It is also important to realize that differentiation of “conventional” granulomatous mycosis fungoides from granulomatous slack skin cannot be achieved on histopathologic grounds only and that correlation with the clinical picture is crucial for a precise classification [148].

Granulomatous Slack Skin

A rare variant of granulomatous mycosis fungoides is represented by granulomatous slack skin, which is characterized clinically by the occurrence of bulky, pendulous skinfolds, usually located in flexural areas (Fig. 2.93) [149–151]. Granulomatous slack skin is considered as a specific manifestation of mycosis fungoides in most cases and is included as a variant of mycosis fungoides in the WHO classification [3]. Association with other lymphomas, including Hodgkin lymphoma, has been reported. In most of these cases, on the other hand, granulomatous slack skin probably represented a manifestation of mycosis fungoides associated with a second lymphoma, rather than a clinicopathologic variant of the other lymphoma.

The granulomatous infiltrate in granulomatous slack skin is usually diffuse and involves the subcutaneous tissues, in contrast to the patchy, dermal granulomas observed in “conventional” granulomatous mycosis fungoides (Fig. 2.94), but differentiation between the two variants may not be possible on histopathological grounds alone. Giant cells with many nuclei are a common finding (Fig. 2.95). Elastophagocytosis is also typically present. The same rearrangement of T-cells has been observed in skin and nodal lesions in one patient [152], and a t(3;9)(q12;p24) has been detected in another case [153].

As already mentioned, a diagnosis of granulomatous slack skin can only be made when the typical clinical presentation is observed, as histopathological differentiation from “conventional” granulomatous mycosis fungoides is not possible [148]. On the other hand, in some patients lesions located at typical sites of granulomatous slack skin (axillae, groins), but that do not show bulky, pendulous skin folds, may be considered as early manifestations of the disease if the characteristic histopathologic features described above are present (Fig. 2.96). Whether these patients, if untreated, progress to overt granulomatous slack skin is unclear.

Erythrodermic Mycosis Fungoides

Patients with mycosis fungoides may develop erythroderma during the course of their disease (Fig. 2.97). Rarely, swelling of the lymph nodes and presence of circulating neoplastic cells (“Sézary cells”) are observed as well, thus showing overlap of clinical features with Sézary syndrome. The histopathologic and phenotypic features are identical to those of conventional mycosis fungoides. After successful treatment patients may relapse with conventional patches, plaques or tumors of mycosis fungoides or with new flares of erythroderma. These cases should not be classified as Sézary syndrome but as erythrodermic mycosis fungoides, and the diagnosis of Sézary syndrome should be reserved to those patients presenting with the typical features of this disease from the outset [1, 3]. In fact, a strong support to the distinction of mycosis fungoides from Sézary syndrome has been provided by phenotypic and genetic analyses that showed major differences between the two diseases [34–36]. Precise history taking is crucial for proper diagnosis, and often only evidence of previous mycosis fungoides allows classification of these cases as erythrodermic mycosis fungoides.

Interstitial Mycosis Fungoides

Interstitial mycosis fungoides is a variant of the disease that does not have a characteristic clinical presentation, but shows histopathologically a pattern that may be misinterpreted as that of an inflammatory dermatosis (particularly interstitial granuloma annulare and inflammatory stage of localized scleroderma) [154]. It is observed usually in flat tumors of mycosis fungoides.

Epidermotropism and a band-like pattern may be missing, and histology shows dermal infiltrates of lymphocytes dissecting the collagen bundles (Fig. 2.98a and b). Immunohistology confirms that most interstitial cells are T lymphocytes, which is a helpful clue for the differential diagnosis with interstitial inflammatory dermatoses (Fig. 2.98c). “Free floating” collagen fibers surrounded by neoplastic lymphocytes have been observed in interstitial mycosis fungoides but not in granuloma annulare or morphea [155]. In this context, it should be remembered that granuloma annulare may rarely present with pseudolymphomatous infiltrates, thus complicating the differential diagnosis between the two diseases [156], and that a rare case of genuine granuloma annulare in a patient with mycosis fungoides has been reported [157].

Poikilodermatous Mycosis Fungoides (Poikiloderma Vasculare Atrophicans)

Poikilodermatous mycosis fungoides is one of the most common clinicopathologic variants of the disease. It is characterized clinically by atrophic, red-brown macules and patches with prominent teleangectasia (Figs 2.99 and 2.100). Sites of predilection are the breast and buttocks, but lesions may be generalized. Sometimes a “retiform” pattern can be observed (formerly referred to as “parakeratosis variegata”). Histology reveals an atrophic epidermis with loss of rete ridges, an interface dermatitis with a superficial, band-like infiltrate of lymphocytes with epidermotropism, and a thickened papillary dermis (Fig. 2.101). Necrotic keratinocytes and pigment incontinence may be a prominent finding. Dilated capillaries are present within the superficial dermis. This variant has also been referred to as the “lichenoid” type of mycosis fungoides, and in some cases distinction from pigmented atrophic lichen planus may be impossible without correlation with the clinical picture.

A cytotoxic phenotype seems to be common in poikilodermatous mycosis fungoides [158]. Recent data point to a better prognosis of poikilodermatous mycosis fungoides as compared to other variants of the disease [24].

Hypopigmented Mycosis Fungoides

Patients with mycosis fungoides may develop hypopigmented patches and plaques (Fig. 2.102). These lesions may be misinterpreted clinically as those of pityriasis versicolor, pityriasis alba, or vitiligo. Histology reveals typical features of mycosis fungoides. Hypopigmented mycosis fungoides is observed more frequently in dark-skinned individuals and is one of the most frequent variants seen in children [159, 160]. Repigmentation usually takes place after successful treatment of the lesions.

Although hypopigmented lesions may be the sole manifestation of mycosis fungoides, in many cases, especially in Caucasians, careful examination of the patients will allow the presence of erythematous lesions to be detected as well (Fig. 2.103). Interestingly, patients with hypopigmented mycosis fungoides tend to develop more hypopigmented lesions during the course of the disease, suggesting that this is indeed an unusual variant of mycosis fungoides.

A predominant CD8⁺ phenotype has been described in patients with hypopigmented mycosis fungoides, possibly underlying some pathogenetic similarities to vitiligo [159]. However, a CD4⁺ phenotype can be observed as well [160], and most cases of mycosis fungoides that are positive for CD8 are not hypopigmented clinically. It should be underlined that differential diagnosis of hypopigmented mycosis fungoides from its simulators, including particularly atopic dermatitis and the inflammatory stages of vitiligo, may be extremely difficult, and that at least some of the cases published in the literature may not represent genuine examples of hypopigmented mycosis fungoides. In my experience this problem is particularly evident in children, and sometimes only follow-up data provide a precise diagnosis (see also Chapter 26).

It is interesting to note that hypopigmented lesions have been described in so-called parapsoriasis en plaques [161] and that an identical clinical presentation may be observed also in “idiopathic generalized follicular mucinosis” (see Fig. 2.76), again suggesting that these are variants of mycosis fungoides.

Hyperpigmented Mycosis Fungoides

Rarely, the clinical picture of mycosis fungoides may be characterized by markedly hyperpigmented lesions, corresponding histopathologically to the presence of pigment incontinence and abundant melanophages in the papillary dermis (Figs 2.104 and 2.105). The phenotype of these lesions is more often cytotoxic than helper [162], possibly explaining the damage at the dermo-epidermal junction and subsequent pigment incontinence. On the other hand, it should be underlined that most cases of mycosis fungoides with cytotoxic phenotype do not present with hyperpigmented lesions.

Generalized hyperpigmented lesions covering the entire body have been described as “melanoerythroderma”, and are seen more frequently in patients with Sézary syndrome.

Purpuric Mycosis Fungoides

In some patients, lesions of mycosis fungoides are characterized clinically by a purpuric hue (Fig. 2.106) and histopathologically by the presence of many extravasated erythrocytes and siderophages (Fig. 2.107).

The differential diagnosis of purpuric mycosis fungoides from lichenoid purpura and lichen aureus can be difficult on histopathologic grounds, and correlation with the clinical features is crucial (see also Chapter 26). The distinction has been rendered even more difficult by the introduction of the term “atypical pigmented purpura” and by the suggestion of a possible relationship between some purpuric dermatoses and mycosis fungoides [163, 164]. Rare patients with “pigmented purpura” progressing to mycosis fungoides [165, 166] probably had purpuric mycosis fungoides from the outset. In this context, it should be emphasized that lichen aureus and lichenoid purpura are wholly benign disorders that should be clearly differentiated from mycosis fungoides.

Papular Mycosis Fungoides

In some patients with mycosis fungoides, small papules represent the predominant morphologic expression of the disease, and small and/or large patches are absent (Fig. 2.108) [167, 168]. The papules are not centered on the hair follicles, and these last are not involved histopathologically. Histopathologic examination reveals conventional features of mycosis fungoides, but the infiltrate is restricted to a part only of the biopsy specimen (Fig. 2.109). The absence of typical manifestations of the disease renders the clinical diagnosis very difficult. Moreover, differentiation from type B lymphomatoid papulosis can only be achieved by short-term follow-up of the patients, as the histopathologic features may be indistinguishable (CD30, however, is usually negative in papular mycosis fungoides). In fact, papular lesions of mycosis fungoides do not show the spontaneous regression typical of type B lymphomatoid papulosis.

Although we initially suggested that papular mycosis fungoides had the same prognosis as other variants of early mycosis fungoides [167], since the original publication I have observed patients who progressed to erythroderma or tumor stage within relatively short periods of time; thus great caution should be exerted in the management of these cases.

Bullous (Vesiculobullous) and Dyshidrotic Mycosis Fungoides

Some patients with mycosis fungoides present with vesicular or bullous lesions, these last usually associated with large, superficial erosions (Fig. 2.110). Histologic examination reveals bullous lesions characterized either by cleavage at the dermoepidermal junction or by confluence of intraepidermal vesicles (Figs 2.111 and 2.112). Sometimes the bullae are not visible on histopathologic specimens and only superficial erosions are present (see Teaching case 2.1 in this chapter). A “dyshidrotic” variant of the disease located on the palms and soles may also present with vesicular lesions (Figs 2.112 to 2.114). Typical features of mycosis fungoides are commonly present near to the vesicular and bullous ones, or at other sites of the body.

The reason(s) for the formation of vesicular and bullous lesions is unclear. Some cases may be due to pre-existent dyshidrotic or vesicular dermatitis and subsequent colonization by neoplastic cells. In other patients a direct cytotoxic effect may be responsible for the detachment of the epidermis from the dermis or pronounced spongiosis may lead to vesiculation, and eventually larger blisters may develop due to confluence of vesicles. Neoplastic cells may also induce directly and/or indirectly degeneration of collagen bundles and the blistering [169]. In some patients, onset of bullous lesions has been related to the use of interferon-α. Rarely, coexistence of mycosis fungoides and an autoimmune bullous disorder may be the reason for this peculiar clinical presentation.

Eczema herpeticatum developing on the background of bullous mycosis fungoides has been described [170].

Anetodermic Mycosis Fungoides

Two patients with secondary anetoderma developing on pre-existing lesions of mycosis fungoides have been described [171]. Some of the lesions depicted in this study resembled small tumors clinically, but represented probably just secondary anetodermic changes in flat lesions of mycosis fungoides, thus being a potential pitfall in clinical evaluation of the patients. Although loss of elastic fibers is observed also in granulomatous slack skin, anetoderma secondary to mycosis fungoides has completely different clinical and histopathologic features, and there is no relationship between these two variants of the disease.

I have observed anetoderma also in the context of generalized follicular mucinosis (Fig. 2.115), again underlining the fact that different variants of mycosis fungoides may show overlapping clinicopathologic features.

Pityriasis et Varioliformis Acuta (PLEVA)-Like Mycosis Fungoides

A variant of mycosis fungoides that simulates clinically and especially histopathologically PLEVA has been described in children (Fig. 2.116a) [172–174]. Most of these cases have a CD8⁺ cytotoxic phenotype (Fig. 2.116b), thus enhancing the similarities to PLEVA. I have observed biopsies from adult patients that histopathologically showed features identical to those described in children and similar cases have been described [175].

In the literature there are some reports of “atypical PLEVA,” sometimes with progression to mycosis fungoides [176]. It is unclear whether these patients had mycosis fungoides from the beginning or if an association between the two diseases exists (see also the section on mycosis fungoides in children and adolescents in this chapter and Chapters 25 and 26). In my opinion, great caution should be exerted in cases of “atypical” PLEVA and regular follow-up controls are mandatory (see Teaching case 25.1 in Chapter 25).

“Invisible” Mycosis Fungoides

Normal-looking skin in patients with mycosis fungoides may show histopathologic, electron microscopic, immunophenotypic, and/or molecular genetic evidence for the presence of neoplastic lymphocytes [177]. These cases have been termed “invisible” mycosis fungoides in the literature, because the lesions are clinically unapparent. The presence of such histologic features in normal-looking skin has been documented at the first diagnosis of mycosis fungoides, in patients with mycosis fungoides at sites distant from erythematous lesions, and in patients in complete clinical remission after treatment.

The finding of specific infiltrates of mycosis fungoides in skin that looks clinically normal poses two main questions: the first relates to the extent of the area that should be treated by skin-targeted therapies and the second concerns the proper monitoring of patients after treatment. At present it is difficult to answer either question, as the clinical implications (if any) of “invisible” persistent disease are not clear. In my experience, however, biopsies of normal-looking skin do not add information relevant for further management of patients with mycosis fungoides.

Other Unusual Clinical Presentations of Mycosis Fungoides

Other clinicopathological variants of mycosis fungoides have been described besides those discussed above in more detail (Table 2.3). Ichtyosiform mycosis fungoides is characterized by a cobblestone pattern of flat, geometric plaques that remind clinically of ichtyosis vulgaris (Fig. 2.117). The histopathologic features are identical to those of more conventional variants. Lesions restricted to palms and soles have been termed “mycosis fungoides palmaris et plantaris” and may resemble psoriasiform or dyshidrosiform dermatoses. Perioral and nasolabial papules may mimic perioral dermatitis (Fig. 2.118) [178]. In some patients, lesions may resemble clinically acanthosis nigricans. A condition reported in Japan as “papuloerythroderma Ofuji” [179] represents most likely a variant of mycosis fungoides characterized by a similar clinical presentation and protracted course.

Many of these variants have been observed in anecdotal cases only. On the other hand, they underline once again the protean clinicopathologic features of mycosis fungoides and the need for a high level of suspicion when diagnosing cutaneous eruptions that do not fit well into a precise category of inflammatory skin disease.