Primary cutaneous marginal zone lymphoma

INTRODUCTION — Primary cutaneous B cell lymphoma (PCBCL) refers to those cases of B cell lymphoma that present in the skin when there is no evidence of extracutaneous disease after the completion of an initial staging evaluation. There are three main subtypes of PCBCL [1]:

●Primary cutaneous follicle center lymphoma (PCFCL)

●Primary cutaneous large B cell lymphoma, leg type

●Primary cutaneous marginal zone lymphoma (PCMZL). In the revised 4^th edition of the World Health Organization classification, PCMZL is included in the broader category of extranodal marginal zone lymphoma [2] (see "Clinical manifestations, pathologic features, and diagnosis of extranodal marginal zone lymphoma of mucosa associated lymphoid tissue (MALT)")

These three subtypes appear to be unique entities with differing clinical presentation, pathologic features, and prognosis. They also require slightly different treatment approaches. This topic review will discuss PCMZL. Before 2008 some cases of PCMZL may have been classified as primary cutaneous immunocytoma or cutaneous plasmacytoma.

The other PCBCL subtypes and other forms of cutaneous lymphoma (eg, T cell lymphoma, mycosis fungoides) are presented separately. (See "Primary cutaneous follicle center lymphoma" and "Primary cutaneous large B cell lymphoma, leg type" and "Clinical manifestations, pathologic features, and diagnosis of mycosis fungoides".)

EPIDEMIOLOGY — Approximately 25 percent of patients with non-Hodgkin lymphoma will present at an extranodal site without systemic involvement. The skin is the second most common primary extranodal site, second in frequency only to the gastrointestinal tract. The overall incidence of primary cutaneous lymphoma in Western countries is estimated to be 0.5 to 1 case per 100,000 people annually, of which approximately 20 percent represent primary cutaneous B cell lymphoma (PCBCL) [1,3,4]. The incidence varies geographically, with lower rates in Japan and Korea [5,6]. The exact incidence of PCMZL is unknown.

PCMZL usually presents in the fifth or sixth decade of life, although patients as young as 15 years old have been reported [7-9]. Men are diagnosed approximately twice as often as women [9]. The vast majority of cases occur in non-Hispanic White individuals [4].

PATHOGENESIS — PCMZL has been associated with tattoo pigments, tick bites, and antigen injection, suggesting that they may develop from chronic antigenic stimulation by intradermally applied antigens [10]. However, the cause is unknown in most cases. There is no identifiable hereditary tendency.

An association between PCMZL and infectious agents such as Borrelia burgdorferi and autoimmune conditions has been postulated, as described below.

Infectious etiology

Borrelia burgdorferi — Pathogenetic links between primary cutaneous B cell lymphoma (PCBCL) and Borrelia burgdorferi infection have been established in Europe, but North American and Asian case series have thus far failed to demonstrate a similar relationship [11-14].

●Using enzyme-linked immunoabsorbent assays (ELISA) and Western Blot, positive Borrelia serologies were reported in 55 percent of patients with PCBCL, compared with 8 percent of patients with extranodal B cell non-Hodgkin lymphoma (NHL), none of 10 patients with cutaneous T cell NHL, and 3 percent of healthy blood donors [12]. The predominant strain identified was B. afzelii, a subtype that is a primary cause of acrodermatitis chronica atrophicans [15]. However, of the 12 patients with PCBCL and positive Borrelia serologies, only two had clinical manifestations of acrodermatitis chronica atrophicans, and only one had arthritis [12].

●In a similar series of 20 cases from Scotland, seven of the PCBCL lesions contained identifiable Borrelia DNA compared with only 1 of 40 control patients [16]. (See "Clinical manifestations of Lyme disease in adults", section on 'Acrodermatitis chronica atrophicans'.)

Borrelia infection has also been reported in sporadic cases of primary cutaneous follicle center lymphoma (PCFCL) and primary cutaneous large B cell lymphoma (PCLBCL), leg type. A possible link between Borrelia infection and PCBCL is further substantiated by reports of response of PCBCL to antibiotic therapy. However, this has been limited to case reports and does not appear to be as effective as Helicobacter pylori eradication in gastric MALT lymphoma. (See 'Treatment' below.)

Differences in B. burgdorferi strains between different endemic regions may provide some explanation; B. afzelii appears to be the predominant strain in Europe, while B. burgdorferi is more common in North America [17].

Viral etiologies — No definitive evidence exists for a link between PCBCL and viral infection (eg, Epstein-Barr virus, HIV, hepatitis C virus) despite several case reports that have suggested an association [18-23]. Treating hepatitis C virus infection does not appear to induce regression of PCBCL, as has been reported in splenic marginal zone lymphoma [24]. (See "Splenic marginal zone lymphoma".)

Autoimmune disorders — Although systemic marginal zone lymphoma has been reported in conjunction with various autoimmune disorders, no definitive association between autoimmune disease and PCMZL has been established [25-27]. (See "Clinical manifestations, pathologic features, and diagnosis of extranodal marginal zone lymphoma of mucosa associated lymphoid tissue (MALT)", section on 'Disease associations'.)

CLINICAL FEATURES — Patients with PCMZL present with red to violaceous papules, plaques, or nodules localized preferentially on the trunk or upper extremities (picture 1A-B). In contrast to primary cutaneous follicle center lymphoma (PCFCL), presentation with multifocal skin lesions is frequent. Ulceration is uncommon. PCMZL has a tendency to recur in the skin, but dissemination to extracutaneous sites is exceedingly rare [13,28]. (See "Clinical manifestations, pathologic features, and diagnosis of extranodal marginal zone lymphoma of mucosa associated lymphoid tissue (MALT)".)

The presence of "B" symptoms, abnormal blood counts, or an elevated lactate dehydrogenase (LDH) should raise suspicion of a systemic lymphoma.

The clinical features of PCMZL were best illustrated in an analysis of the Cutaneous Lymphoma Network database, which identified 137 consecutive patients with PCMZL and reported the following clinical features at the time of diagnosis [9]:

●Skin lesions consisted of tumors (84 percent), plaques (26 percent), and papules (19 percent)

●Lesions were characterized as solitary (51 percent), localized (26 percent), or multifocal (23 percent)

●The most common areas of involvement were the trunk (54 percent), upper extremities (36 percent), head/neck (23 percent), and lower extremities (19 percent)

●Extracutaneous involvement was uncommon (4 percent)

●Serum LDH and beta2-microglobulin were elevated in 7 and 4 percent, respectively

PATHOLOGIC FEATURES — The diagnosis of PCMZL is made based on a pathologic evaluation of a skin biopsy in a patient who has no evidence of systemic lymphoma on staging studies. (See 'Staging' below.)

Morphology — Skin biopsy specimens are characterized by nodular to diffuse infiltrates with sparing of the epidermis (picture 2). There is usually a prominent narrow layer (Grenz zone) beneath the epidermis that is not infiltrated or involved in the same way as the lower layers of the dermis.

The infiltrates are comprised of small lymphocytes, lymphoplasmacytoid cells, and plasma cells, admixed with small numbers of centroblast- or immunoblast-like cells and many reactive T cells (picture 3) [29]. Reactive germinal centers are frequently observed. Monotypic plasma cells are often located at the periphery of the infiltrates and in the superficial dermis beneath the epidermis. A subset of PCMZL is characterized by a predominance of monocytoid B cells instead of lymphoplasmacytoid cells [30,31].

PAS-positive, immunoglobulin containing, intranuclear (Dutcher bodies) or intracytoplasmic (Russell bodies) inclusions may be present in cases with a predominance of lymphoplasmacytoid cells. Unlike extranodal marginal zone lymphoma occurring at other sites, PCMZL rarely show colonization of follicular structures, lymphoepithelial lesions or transformation into a diffuse large B cell lymphoma, though a relative increase in large transformed cells can be seen in some cases [32]. (See "Clinical manifestations, pathologic features, and diagnosis of extranodal marginal zone lymphoma of mucosa associated lymphoid tissue (MALT)".)

Immunophenotype — Evaluation of the biopsy material with immunohistochemical stains is a key component of the diagnostic evaluation of cutaneous lymphoma. At this time, flow cytometry cannot be routinely recommended as a replacement for immunohistochemical studies. This is principally because of the difficulty associated with making a single-cell suspension with this tumor type.

The postulated normal counterpart to PCMZL is a post-germinal center B cell. The neoplastic cells express CD20, CD22, CD79a, and BCL2 and are typically negative for CD3, CD5, CD10, and BCL6 (picture 3). Reactive germinal centers typically express BCL6 and CD10, but not BCL2. There is cytoplasmic expression of immunoglobulin with light chain restriction on paraffin sections as well as monoclonal immunoglobulin gene rearrangements [10].

Two types of PCMZL are now recognized [30,31,33]. Unlike most other MALT lymphomas, the vast majority of PCMZL express class-switched immunoglobulins, including IgG, IgA, and IgE; do not express the chemokine receptor CXCR3; and show a predominance of T cells and only a small proportion of neoplastic B cells. A proportion of these PCMZL with plasmacytic differentiation (up to 39 percent) are IgG4-positive, while IgG4 is rarely expressed by noncutaneous MZL [34,35]. There was no evidence of systemic IgG4 disease in any of these patients pointing to a localized immunologic IgG4-driven process. These class-switched cases are considered by some authors to constitute a clonal chronic lymphoproliferative disorder rather than an overt lymphoma [30,31]. A small subset of (P)CMZL shows a diffuse proliferation or large nodules of neoplastic B cells, which express IgM and often CXCR3. These cases contain a much lower number of admixed T cells and more likely have extracutaneous disease [30,31]. It has been suggested that B. burgdorferi-associated PCMZL belong to this second group [33].

Genetics — Genotypic studies may be used to document clonality. Unlike other extranodal marginal zone lymphomas that demonstrate trisomy 3 or t(11;18), no reliable genetic lesion is demonstrable in PCMZL [29,36]. On cytogenetic and molecular testing, t(14;18), t(3;14), and t(11;18) have been reported in <25, 10, and 7 percent of cases, respectively [37-41]. MYD88 mutations have been found in 50 percent of non-class-switched cases of PCMZL, but not in class-switched cases [42]. Recurrent mutations in FAS were found in 24 of 38 (63 percent) patients with PCMZL [43].

DIAGNOSIS — The diagnosis of PCMZL requires a representative biopsy of involved skin and the exclusion of non-cutaneous disease. Excisional biopsies are preferred to punch biopsies, but if a punch biopsy is taken, the diameter should be at least 4 mm. (See 'Staging' below.)

The biopsy specimen is evaluated for morphology, growth pattern, and immunohistochemical studies. A typical biopsy of PCMZL demonstrates a B cell infiltrate comprised of small to medium sized lymphocytes, often with a plasmacytoid appearance in a reactive germinal-center-like structure. B cell lineage is confirmed by expression of CD20 or CD79a and the absence of CD3 expression.

DIFFERENTIAL DIAGNOSIS — Patients with PCMZL present with red to violaceous papules, plaques, or nodules that can mimic reactive and clonal skin disorders, all of which can be distinguished by pathologic review of a skin biopsy with morphologic evaluation and immunohistochemical stains (table 1).

This pathologic distinction is based on three main factors [44]:

●Identifying the infiltrate as clonal (not reactive), usually by demonstrating monotypic light chain restriction.

●Determining that the neoplastic cells are of B cell origin (not T cell or NK cell), usually by demonstrating the presence of B cell antigens (eg, CD20) and lack of T cell antigens (eg, CD3). Importantly, while the neoplastic cells are of B cell origin, there can be many admixed reactive T cells in the infiltrate.

●Confirming that the infiltrate has morphologic and immunophenotypic features of PCMZL, rather than another cutaneous B cell lymphoma (eg, primary cutaneous follicle center lymphoma; primary cutaneous diffuse large B cell lymphoma, leg type), as based on the morphology and immunophenotype.

In addition, staging studies are performed to distinguish PCMZL from systemic lymphoma. (See 'Staging' below.)

Reactive infiltrate — It is important to differentiate the neoplastic B cell infiltrate seen in primary cutaneous B cell lymphoma (PCBCL) from reactive B cell infiltrates. Entities that can mimic cutaneous B cell lymphoma include cutaneous lymphoid hyperplasia and cutaneous drug reactions (pseudolymphoma) [45]. In general, PCBCLs will demonstrate monotypic light chain restriction while reactive processes are polyclonal [46]. Both PCMZL and cutaneous B cell pseudolymphoma may develop from chronic antigenic stimulation by intradermally applied antigens (eg, tattoo pigments, tick bites, antigen injections). Both conditions may be manifestations of a continuous spectrum of cutaneous B cell proliferations [10]; as such, there has been debate over whether all cases currently classified as PCMZL should be considered overt malignancies [44].    

Systemic lymphoma — Patients with systemic B cell non-Hodgkin lymphoma (NHL) can develop cutaneous disease at some point in their illness. It is important to distinguish between patients with secondary cutaneous involvement by a systemic lymphoma and those with PCMZL since these entities have drastically different prognoses and management options.

Marginal zone lymphoma — Patients with cutaneous involvement of systemic marginal zone lymphoma (MZL) may present with skin lesions that clinically and histologically resemble those of PCMZL. While there are some clinical and pathologic findings that favor one diagnosis over the other, systemic evaluation for disease outside of the skin is necessary to stage PCMZL and to definitively rule out extracutaneous disease. (See "Primary cutaneous follicle center lymphoma", section on 'Staging'.)

When compared with patients with PCMZL, patients with secondary skin involvement of MZL tend to be older and have a distribution of lesions favoring the head and neck [8]. Skin biopsies of PCMZL lesions demonstrate similar histology to other types of extranodal MZL, but rarely show colonization of follicular structures or lymphoepithelial lesions and also differ in the expression of class-switching immunoglobulins and the composition of background inflammation [8,33]. While most noncutaneous MZLs express IgM, cases of PCMZL typically express IgG, IgA, or IgE. In addition, the inflammatory background in PCMZL most closely resembles that of a Th2 cytokine-influenced milieu whereas other extranodal MZL with secondary skin involvement have a Th1 cytokine inflammatory background. Colonization of follicular structures by neoplastic marginal zone cells, as often observed in extranodal MZL arising at other sites, is rarely seen in PCMZL. (See "The adaptive cellular immune response: T cells and cytokines", section on 'Th1' and "Clinical manifestations, pathologic features, and diagnosis of extranodal marginal zone lymphoma of mucosa associated lymphoid tissue (MALT)".)

Other systemic lymphomas — PCMZL may also be difficult to differentiate from mantle cell lymphoma or B cell lymphocytic lymphoma with skin involvement based on histology alone. Staining for CD5 and cyclin D1 may be useful to differentiate PCMZL (CD5-, cyclin D1-) from mantle cell lymphoma (CD5+, cyclin D1+) and skin localizations of small lymphocytic leukemia (CD5+, cyclin D1-). (See "Mantle cell lymphoma: Epidemiology, pathobiology, clinical manifestations, diagnosis, and prognosis", section on 'Diagnosis and classification' and "Clinical features and diagnosis of chronic lymphocytic leukemia/small lymphocytic lymphoma", section on 'Evaluation and diagnosis'.)

Other cutaneous lymphomas

Primary cutaneous diffuse large B cell lymphoma, leg type — PCMZL consists primarily of small cells with irregular nuclei. Lesions with a diffuse pattern or a monotonous proliferation of centroblasts and/or immunoblasts are classified as primary cutaneous diffuse large B cell lymphoma, leg type regardless of their anatomic location. (See "Primary cutaneous large B cell lymphoma, leg type".)

Primary cutaneous follicle center lymphoma — PCMZL may be difficult to distinguish from primary cutaneous follicle center lymphoma (PCFCL) since the reactive follicular structures of PCMZL may resemble the malignant follicles of PCFCL. The former disorder can often be differentiated from PCFCL based on differences in clinical presentation and immunophenotypic findings on the skin biopsy. A hallmark of PCMZL is monotypic immunoglobulin light chain expression by the lymphoplasmacytoid and plasma cells that are typically located at the periphery of the nodular infiltrates and subepidermally; this is rarely observed in PCFCL. Furthermore, PCFCL is often characterized by a diffuse infiltrate of large cleaved cells (large centrocytes) that express BCL6, but not BCL2, and have variable expression of CD10. In contrast, the neoplastic cells of PCMZL express BCL2, but not BCL6 or CD10.

Cutaneous T cell lymphoma — Primary cutaneous CD4+ small/medium T cell lymphoproliferative disorder is an uncommon lymphoma that can morphologically resemble PCMZL and often has many commingling B cells and in some cases may also show light chain-restricted plasma cells [44]. Unlike PCMZL, the medium-sized to large T cells in this entity express CD4, CXCL13, and PD-1. Cutaneous T cell lymphomas can usually be distinguished from cutaneous B cell lymphomas through their expression of T cell markers such as CD2, CD3, and CD5. In addition, cutaneous T cell lymphomas can be identified by T cell receptor (TCR) gene rearrangement studies. This is discussed in more detail separately. (See "Clinical manifestations, pathologic features, and diagnosis of mycosis fungoides", section on 'Diagnosis'.)

Intravascular lymphoma — Intravascular large B cell lymphoma, is a distinctly uncommon NHL variant that typically presents as multiple, erythematous tender nodules, tumors, or telangiectasias in older patients. Isolated involvement of the skin does occur, but the disease more frequently affects multiple organ systems, including the central nervous system, producing a wide variety of symptoms, ranging from generalized confusion or dementia to focal motor and sensory deficits. This is discussed in more detail separately. (See "Intravascular large B cell lymphoma".)

Extramedullary plasmacytoma — Extramedullary plasmacytoma, also known as extra-osseous plasmacytoma, can rarely present with isolated skin lesions [47,48]. Although previously considered separate entities, primary cutaneous plasmacytoma is now included in the category of PCMZL [49,50].

Skin involvement by extramedullary plasmacytoma typically demonstrates nodular or diffuse infiltrates of mature plasma cells that express CD38 and monotypic cytoplasmic immunoglobulin, but not CD20 or LCA; the plasma cells may be multinucleate [51]. Coexistent clonal lymphocytes are not found [29].

In the case of extreme plasma cell differentiation, staging studies to confirm the absence of disease outside of the skin should include a bone marrow biopsy and aspirate, skeletal survey, serum immunoglobulins, serum and urine protein electrophoresis, complete blood count, calcium, and serum chemistries, including renal function and alkaline phosphatase in addition to the staging studies described below. (See "Multiple myeloma: Clinical features, laboratory manifestations, and diagnosis", section on 'Evaluation'.)

STAGING — At the time of diagnosis, patients with suspected primary cutaneous lymphoma should undergo a complete history, physical examination, and staging evaluation (table 2). This evaluation both confirms the diagnosis of primary cutaneous lymphoma by excluding involvement of other sites and provides information to guide treatment.

We perform the following studies as part of this staging evaluation:

●Laboratory studies include a complete blood count, comprehensive serum chemistries, serum lactate dehydrogenase (LDH), and serum protein electrophoresis (SPEP). Flow cytometry of the peripheral blood mononuclear cells is indicated for patients with a lymphocytosis.

●Borrelia testing (serology and preferably Borrelia DNA testing on lesional skin by PCR) should be performed in endemic areas (parts of Europe, less relevant in United States and Asia)

●Radiographic imaging with a contrast-enhanced computed tomography (CT) scan of the chest, abdomen and pelvis; the neck should be included if there is palpable adenopathy. For patients with multifocal skin lesions, we perform positron emission tomography with computed tomography (PET/CT) to image the extremities and assess lymph nodes in order to exclude systemic disease. Magnetic resonance imaging (MRI) is often substituted for patients who are unable to safely undergo CT scan. Suspicious lesions should be biopsied.

●Bone marrow aspiration and biopsy is not performed routinely [52].

If any of the above studies demonstrates extracutaneous involvement in a newly diagnosed patient, strong consideration should be given to the diagnosis of systemic lymphoma. The presence of "B" symptoms, abnormal blood counts, elevated beta-2-microglobulin, or an elevated lactate dehydrogenase (LDH) should also raise suspicion of a systemic lymphoma. (See "Pretreatment evaluation and staging of non-Hodgkin lymphomas".)

Using information gathered from the above tests, patients are staged using the EORTC/International Society for Cutaneous Lymphomas (ISCL) TNM (tumor, node, metastases) staging system for primary cutaneous lymphomas other than mycosis fungoides and Sézary syndrome (table 2) [53].

TREATMENT — PCMZL is a rare disorder; accordingly, large trials are lacking. The published data regarding treatment consist entirely of retrospective reviews and anecdotal reports. The generally indolent nature of primary cutaneous B cell lymphoma (PCBCL) has shifted focus to localized, low morbidity modalities such as radiation therapy.

In general, our treatment approach depends on the number of lesions, their location, and the presence of symptoms attributable to the lesions (eg, pruritus):

●For patients with lesions that can be contained within one radiation field, we suggest initial treatment with local radiation therapy rather than observation, surgery, or more aggressive therapy (eg, combination chemotherapy, rituximab therapy). A radiation dose of 24 Gy is typically used and the treatment area should include a margin of healthy appearing tissue [54]. Solitary lesions in areas not amenable to radiation therapy may be excised surgically. This approach aims to maximize response in early stage disease where localized therapy is effective and has limited toxicities.

●For patients with asymptomatic multifocal disease, we suggest an initial observation period. For those with symptomatic multifocal disease, we suggest that treatment be directed at the symptomatic lesion(s) with intralesional triamcinolone, low-dose radiation therapy, or surgical excision rather than chemotherapy. This approach aims to limit exposure to potentially toxic systemic agents given the indolent nature of these tumors.

This approach is generally consistent with those proposed by the ISCL, International Lymphoma Radiation Oncology Group, and the National Comprehensive Cancer Network (NCCN) [54-57]. With this approach, extracutaneous spread is uncommon and estimated five-year survival rates are 98 to 100 percent [7,9,28,29].

Support for this approach comes from retrospective studies and case reports. The largest literature review included data on 288 patients with PCMZL and reported the following outcomes by treatment modality [55]:

●Radiation therapy – Compilation of data on 132 patients treated with 30 to 45 Gy of radiation revealed a 99 percent complete response (CR) rate [55]. Relapses were seen in 60 patients (46 percent), most of which were limited to the skin. Three patients had extracutaneous progression. High response rates and good in-field control have been reported in other studies [58]. One study demonstrated responses to lower doses (4 Gy in two fractions) of radiation [59].

●Excision – Of 75 patients treated with complete surgical excision, all but one obtained a CR [55]. During an undefined period of follow-up, skin relapse was seen in approximately 43 percent. In another study of 39 cases of PCMZL treated with excision, all but one achieved a CR, with relapses seen in approximately one-third.

●Intralesional triamcinolone – Case reports and small retrospective studies have reported that intralesional triamcinolone is a safe and effective treatment for low-grade primary cutaneous B cell lymphoma [60,61].

●Rituximab – A small number of patients treated with intralesional (nine patients) or systemic (three patients) rituximab have been reported with CR rates of 89 and 67 percent, respectively [55]. Relapses in the skin were seen in 50 to 60 percent of patients. No extracutaneous relapses were reported. Another trial of intralesional rituximab in 35 patients with PCMZL or PCFCL confirmed these early findings, reporting a CR in 71 percent with a median time to CR of eight weeks [62]. Median disease-free survival was approximately two years.

●Intralesional interferon alfa – Eight cases treated with intralesional interferon alfa (IFN) have been reported in the literature with all obtaining a CR after a median of eight weeks [55]. Two patients relapsed with both obtaining a second CR with repeat IFN treatment.

●Single agent or multiagent chemotherapy – Cumulative data on 14 patients treated with single agent chlorambucil showed responses in all patients, nine of which were CR (64 percent) [55]. Of the nine patients who obtained a CR, three relapsed. Multiagent chemotherapy demonstrated CRs in 28 of 33 patients (85 percent). Relapses were seen in a little over half.

●Antibiotics – Reports of antibiotic use in patients with coexisting PCMZL and Borrelia infection have had mixed results [55]. Six of 14 patients reported in the literature achieved a CR. Treatment appeared to be most successful with cephalosporin therapy. For those with PCMZL, antibiotics should be reserved for those with evidence of Borrelia infection on serologic testing or upon testing of lesional skin. Pathogenetic links between PCMZL and B. burgdorferi infection have been established in Europe, but North American and Asian case series have thus far failed to demonstrate a similar relationship. (See "Treatment of Lyme disease".)

ASSESSING DISEASE RESPONSE — We evaluate and restage patients 8 to 12 weeks after the completion of therapy. Restaging should consist of repeating laboratory and clinical examinations used to stage the disease at presentation. We typically do not perform a bone marrow biopsy unless there is suspicion of bone marrow involvement, such as an abnormal complete blood count. If the radiographic studies were normal at the time of diagnosis, we reserve repeat imaging for patients in whom there is a clinical concern for systemic spread based on symptoms, laboratory results, or physical examination.

Following the completion of therapy, restaging, and documentation of complete remission (CR), patients are seen at periodic intervals to monitor for treatment complications and assess for possible relapse. The frequency and extent of these visits depends upon the comfort of both the patient and physician.

For patients who are in CR, we re-evaluate with a physical examination every three months for the first two years following therapy. The frequency of clinic visits is then decreased thereafter. We do not use imaging studies or blood examinations in the routine follow-up of asymptomatic individuals. This is principally because recurrences are almost always detected by clinical examination. In a retrospective analysis that included 14 patients with PCMZL followed for a median of 44 months, there were 12 episodes of recurrence, all of which were identified by physical examination [63]. Neither blood examinations nor imaging procedures established recurrence or progression in any patient.

TREATMENT AT RELAPSE — Approximately half of patients initially treated with radiation therapy will relapse [9,55]. Most relapses are confined to the skin and occur outside of the radiation field [58]. If they appear outside of the prior radiation field, most will respond to low-dose radiation therapy (2 x 2 Gy) [59]. If they occur inside the radiation field, surgical excision or single agent rituximab may be used. While multiagent chemotherapy remains an option for difficult-to-treat disease, it is rarely necessary. (See 'Treatment' above.)

PROGNOSIS — Patients with PCMZL usually have an indolent clinical course. Cutaneous relapses are common, in particular in patients with multifocal skin lesions [9,28,64]. Extracutaneous dissemination is, however, uncommon. In two studies, only 6 of 71 (8 percent) and 9 of 144 (6 percent) patients with a histologically confirmed PCMZL developed extracutaneous disease [65,66]. The prognosis of patients with PCMZL is excellent, with five-year survival rates of 98 to 100 percent [7,28,29,64-66].

The largest series with long-term follow-up was an analysis of the Cutaneous Lymphoma Network database, which identified 137 consecutive patients with PCMZL followed for a median of 54 months (range 12 to 165 months) [9]:

●The majority was treated with surgical resection (44 percent), radiation (37 percent), or both (5 percent).

●A complete remission (CR) was achieved in 88 percent overall (93 percent for solitary or localized disease, 71 percent for multifocal disease).

●Of the 121 patients who achieved a CR, 53 patients (44 percent) had a cutaneous relapse (17 at initial site, 33 at distant site, and three at both). Eighteen patients had multiple relapses with 10 cases progressing to a higher stage (none with extracutaneous involvement).

●The median disease-free survival (DFS) was 47 months with estimated 5- and 10-year DFS of 46 and 17 percent, respectively. DFS was shorter in those with multifocal disease when compared with localized or solitary lesions (1.1 versus 3.8 and 6.6 years). Lesion type and location did not impact DFS.

●Six patients died during follow-up, with only one death secondary to lymphoma. Estimated overall survival rates at five and 10 years were 93 percent.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Primary cutaneous lymphoma".)

SUMMARY AND RECOMMENDATIONS

●Primary cutaneous B cell lymphoma (PCBCL) refers to those cases of B cell lymphoma that present in the skin when there is no evidence of extracutaneous disease at the time of diagnosis and after the completion of an initial staging evaluation. Primary cutaneous marginal zone lymphoma (PCMZL) is a PCBCL and is also included in the broader category of extranodal marginal zone B cell lymphoma. (See 'Introduction' above and 'Epidemiology' above.)

●There are no clearly defined risk factors for the development of PCMZL; however, an association between PCMZL and infectious agents, such as Borrelia burgdorferi, has been postulated. (See 'Pathogenesis' above.)

●Patients typically present with asymptomatic, localized, deep red to brown infiltrated plaques or grouped nodules surrounded by an area of diffuse erythema (picture 1A-B). Lesions are generally localized to one body region, often the extremities or trunk. Disseminated lesions are rare. (See 'Clinical features' above.)

●The diagnosis of PCMZL requires a representative biopsy of involved skin and the exclusion of non-cutaneous disease in the appropriate clinical setting (picture 2). A typical biopsy of PCMZL demonstrates a clonal B cell infiltrate comprised of small to medium sized lymphocytes, often with a plasmacytoid appearance, which express monotypic immunoglobulin light chains (picture 3). Reactive (nonclonal) germinal-center-like structures may be present. (See 'Pathologic features' above and 'Diagnosis' above and 'Differential diagnosis' above.)

●The staging evaluation both confirms the diagnosis by excluding involvement of other sites and provides information to guide treatment. (See "Primary cutaneous follicle center lymphoma", section on 'Staging'.)

●PCMZL is a rare disorder; accordingly, large trials are lacking. The published data regarding treatment consist entirely of retrospective reviews and anecdotal reports. In general, our treatment approach depends on the number of lesions, their location, and the presence of symptoms attributable to the lesions (eg, pruritus):

•For patients with focal disease (a solitary lesion or lesions that can be contained within one radiation field), we suggest treatment with radiation therapy rather than chemotherapy, surgical excision, or observation (Grade 2C). A radiation dose of 24 Gy is advised and the treatment area should include a margin of healthy appearing tissue. Surgical excision or observation are acceptable alternatives for lesions that are not amenable to radiation therapy (eg, in a location at which radiation toxicity is increased). (See 'Treatment' above.)

•For patients with asymptomatic multifocal disease, we suggest observation (Grade 2C). (See 'Treatment' above.)

•For patients with symptomatic multifocal disease, we suggest that treatment be directed at the symptomatic lesion with either intralesional triamcinolone, or low-dose radiation therapy (2 x 2 Gy), or excision rather than chemotherapy (Grade 2C). The choice between radiation therapy and excision may be made based on predicted toxicities for the affected site. (See 'Treatment' above.)

●Approximately half of patients initially treated with radiation therapy will relapse. Most relapses are confined to the skin and are treated with the same modalities as initial therapy. (See 'Treatment at relapse' above.)