Marginal Zone Lymphomas: Management of Nodal, Splenic, and MALT NHL

Marginal zone lymphomas are indolent B-cell lymphomas that putatively originate from the marginal zone of B-cell follicles, which can be found in the spleen, lymph node, and mucosal lymphoid tissues. The follicle consists of a germinal center surrounded by a corona that is divided into the mantle zone and the marginal zone. The marginal zone forms the outermost portion of the corona and is most developed in lymphoid tissues subjected to constant antigen exposure, such as the spleen and Peyer’s patches. While the etiology of marginal zone lymphomas is likely diverse, an association with chronic infection and inflammation exists and is especially prominent in the MALT lymphomas. A model of chronic antigen-dependent immune stimulation with a microbial pathogen acting as the antigenic source has clearly emerged from the characterization of Helicobacter pylori–associated gastric MALT lymphoma. Other microbes are implicated in the pathogenesis of other marginal zone lymphomas (Campylobacter jejuni, Borrelia burgdorferi, Chlamydia psittaci, and hepatic C virus). Similarly, MALT lymphomas occurring in the salivary and thyroid gland associated with Sjögren syndrome and Hashimoto thyroiditis, respectively, may stem from heavy exposure to autoantigens in these organs as a result of chronic autoimmune inflammation. While the initial lymphoid hyperplasia is driven by physiologic antigenic stimulation, additional oncogenic events, such as chromosomal translocations leading to constitutive activation of signaling pathways, occur during the progression of disease that ultimately result in antigen-independent lymphoproliferation.¹ Despite having a common origin in the marginal zone of the B-cell follicle, there are patently distinct clinical and molecular characteristics of marginal zone lymphomas originating at different anatomic sites. As such, marginal zone–derived lymphomas are currently categorized by the World Health Organization (WHO) into those originating in the spleen (splenic marginal zone lymphoma), extranodal mucosa-associated lymphoid tissue (MALT lymphoma), or lymph node (nodal marginal zone lymphoma, NMZL) (Figure 1; see Color Figures, page 505).

A relatively uncommon form of NHL, NMZL, accounts for less than 2% of new NHL cases.² It is a distinct B-cell lymphoma originating in the lymph node. After its initial description, this tumor was recognized to have morphologic and immunophenotypic similarities with either extranodal lymphomas of the MALT lymphoma or splenic marginal zone lymphoma (SMZL), suggesting a respective common origin. However, as a distinct entity, diagnosis of NMZL requires the exclusion of extranodal and splenic involvement. Concordantly, the World Health Organization (WHO) definition of NMZL is “a primary nodal B-cell neoplasm that morphologically resembles lymph nodes involved by marginal zone lymphomas of extranodal or splenic types, but without evidence of extranodal or splenic disease.”

In the affected lymph node, the pattern of tumor growth may be perivascular, perisinusoidal, and perifollicular, possibly with invasion of malignant cells into the follicles in a pattern known as follicular colonization. With this pattern, monocytoid B cells, recognized by their abundant pale cytoplasm, may be prominent. Alternatively, the growth pattern may be nodular with expansion of the marginal zone and destruction of the mantle zone surrounding the germinal center of the follicle.³ 

NMZL cells are mature B cells that express the B-cell markers CD20, CD79a, and surface immunoglobulin IgM, but lack expression CD5, CD23, and cyclin D1, and lack the expression germinal center markers CD10 and Bcl-6. This phenotype is similar to MALT lymphoma; however, an occasional case will express surface IgD, which is seen in SMZL.^3,4 In regards to the cell of origin, NMZL may arise from different types of B cells found in the marginal zone.⁵ Beyond morphologic and immunophenotypic descriptions, the biology of NMZL is not well studied. It is known that the t(11;18)(q21;q21) associated with MALT lymphoma is not found in NMZL, but trisomy 3 appears to occur in all three types of marginal zone lymphoma.^6,7 

The clinical features have been well summarized in previous publications.^8,9 The median age at diagnosis is 60 and the gender distribution is approximately equal. Like other nodal indolent lymphomas (follicular lymphoma and small lymphocytic lymphoma), the disease tends to present in advanced stages with widespread but non-bulky lymphadenopathy. Bone marrow involvement is seen in approximately one-third of new cases. B symptoms are reported in less than 15% and the majority of patients are asymptomatic at diagnosis. A small to moderate IgM paraprotein can be detected in approximately 10% of cases.

The clinical course is also similar to other nodal indolent lymphomas. The disease does not appear to be curable with chemotherapy, but despite a continuing pattern of relapse, there is a long median survival, approaching 10 years in some series.⁸ The prognosis appears to be less favorable than MALT lymphoma, splenic marginal zone lymphoma, and follicular lymphoma and roughly comparable to small lymphocytic lymphoma.^2,9 As yet, there are no publications providing information on median overall survival in the rituximab era. Transformation to diffuse large B-cell lymphoma can occur, although the incidence is not well studied. In one series, 20 of 124 (16%) transformed at a median time of 4.5 years from diagnosis.⁸ 

Regarding treatment, no prospective trials have been conducted specifically for NMZL to help clarify the optimal therapeutic approach. In general, one should apply the same therapeutic principles that would be applied to follicular lymphoma or small lymphocytic lymphoma. Observation alone is appropriate for asymptomatic patients with low tumor burden. Single-agent rituximab is also appropriate for low tumor burden patients unwilling to consider a ‘watch-and-wait’ strategy or for elderly patients with symptoms who are not candidates for chemotherapy. Patients requiring therapy due to symptoms, organ compromise, or high tumor burden should receive a rituximab-chemotherapy combination. The optimal chemotherapy regimen is a matter of debate and will depend somewhat on the clinical presentation, but reasonable choices include the CVP regimen, CHOP, or a purine analogue-based regimen. Fludarabine 25 mg/m² for 5 days plus rituximab 375 mg/ m² day 1 was evaluated in a phase II study of patients with marginal zone lymphoma (including MALT, nodal and splenic types).¹⁰ The overall response rate (ORR) was 83% and the complete response rate was 52%. The follow-up time was too short to make a determination regarding response duration. While response rates were high, significant toxicity, especially hematologic, was noted and prevented one-half of the patients from completing therapy.

SMZL is a distinctive form of indolent lymphoma originating in the spleen, characterized by prominent splenomegaly and variable involvement of lymph nodes, bone marrow, peripheral blood, and other organs. The WHO definition is “a B-cell neoplasm comprising small lymphocytes which surround and replace the splenic white pulp germinal centers, efface the follicle mantle and merge with a marginal zone of larger cells including scattered transformed blasts; both small and large cells infiltrate the red pulp. Splenic hilar lymph nodes and bone marrow are often involved; lymphoma cells may be found in the peripheral blood as villous lymphocytes.” It is an uncommon form of NHL, accounting for less than 1% of new cases.

The immunophenotype of SMZL is similar to NMZL and MALT lymphoma with the exception of occasional surface IgD co-expression.³ IgD expression may be associated with a subset of cases that lack somatic mutations of IgVH genes, which in itself may be an independent indicator of adverse clinical course.^11,–13 The majority of SMZL cases have abnormal karyotypes, and complex chromosomal aberrations are common, but no unique genetic alterations have been noted. The most frequent aberrations are gains of 3q and deletion of 7q22–36. The t(11;18)(q21;q21) associated with MALT lymphoma is not seen in SMZL.^14,15 Further evidence supporting SMZL as a distinct entity is a unique gene expression profile when compared to other indolent B-cell lymphomas.¹³ 

Association with hepatitis C infection has been reported, although the prevalence ranges from 36% to less than 10% and appears to by geographically influenced. The highest associations are reported in Japan and northern Italy.^16,17 One report indicated successful treatment of hepatitis C can result in regression of the SMZL.¹⁸ Seven of 9 patients with chronic hepatitis C and SMZL with villous lymphocytes experienced complete remission of both conditions after treatment with interferon α-2b. In contrast, none of 6 patients with SMZL with villous lymphocytes and negative hepatitis C infection experienced remission after treatment with interferon. An obvious implication is that selected cases may be driven by chronic antigenic stimulation, analogous to gastric MALT lymphoma and Helicobacter pylori. Although the literature is scant, it seems reasonable to test all SMZL patients for hepatitis C infection and to treat infected individuals with appropriate antiviral agents.

The most common presenting symptom is abdominal discomfort due to splenomegaly, which can be moderate to massive. Patients often have modest cytopenias that are primarily due to splenic sequestration with a smaller contribution from marrow infiltration. Usually a bone marrow biopsy is the best initial diagnostic test and will often establish the diagnosis. Occasionally the distinction with other lymphoproliferative disorders, such as hairy cell leukemia (HCL), can be challenging. Flow cytometry of circulating lymphoma cells or the marrow can be helpful as SMZL is typically CD25 negative and CD103 negative, unlike HCL. If no blood or marrow involvement is present, the diagnosis is best established by splenectomy.

The prognosis is usually good after a diagnosis of SMZL. The Intergruppo Italian Linformi (IIL) conducted a retrospective review of 309 patients and identified a 5-year cause-specific survival (CSS) of 76%.¹⁹ They identified 3 risk factors predicting inferior OS in a multivariate model; hemoglobin < 12 g/dL, albumin < 3.5 g/dL, and lactate dehydrogenase (LDH) > upper limits of normal (ULN). Three risk groups were identified using this model: low risk (0 factors), intermediate risk (1 factor), high risk (2–3 risk factors) with 5-year CSS of 88%, 73%, and 50%, respectively.

Splenectomy is usually the treatment of choice for SMZL. While not curative, it provides rapid symptom relief and often completely corrects associated cytopenias. Furthermore, it typically provides excellent disease control, often allowing patients to avoid systemic therapies. Observation alone may be an appropriate management strategy in asymptomatic individuals, and following this strategy, the median time to treatment was 3 years in one series.²⁰ Splenic irradiation, as a palliative maneuver, can be considered for patients too frail for splenectomy, or systemic chemotherapy.²¹ One may consider systemic therapy for patients with contraindications to splenectomy, or patients with a heavy burden of disease outside of the spleen. Reasonable systemic therapies include regimens appropriate for other indolent B-cell lymphomas. In the absence of comparative trials, it is difficult to know if any particular regimens should be preferred, but it appears as though alkylating agents may be less active than purine analogues, and amongst the purine analogues, 2-chlorodeoxy-adenosine may be less active than fludarabine and pento-statin.²² Interestingly, single-agent rituximab has been reported to be highly active in SMZL with an ORR of 100% and a CR rate of 71% in a small Phase II study.²³ 

MALT lymphomas are a distinctive indolent extranodal lymphoma. Paradoxically, they arise in sites normally devoid of lymphoid tissues and are often preceded by chronic inflammation of the affected sites. Overall, they have an excellent long-term prognosis. The WHO definition for MALT lymphoma is “an extranodal lymphoma comprised of morphologically heterogeneous small B-cells including marginal zone cells, cells resembling monocytoid cells, small lymphocytes, and scattered immunoblasts. There is plasma cell differentiation in a proportion of cases. The infiltrate is in the marginal zone of reactive B-cell follicles and extends into the interfollicular region. In epithelial tissues, neoplastic cells typically infiltrate the epithelium, forming lymphoepithelial lesions.”

MALT lymphomas are more likely than other lymphomas to present as localized disease. A unique and challenging issue for MALT lymphoma is staging. Unknown is whether staging should be different than it is for nodal lymphomas, since CT scans are often not optimal for the detection of occult extranodal disease. One group of investigators adopted the following strategy for all newly diagnosed MALT lymphoma patients: CT scans of the chest, abdomen and pelvis, a bone marrow evaluation, upper endoscopy with random biopsies, endoscopic ultrasound, colonoscopy with random biopsies, an endoscopic otolaryngology exam, MRI of the salivary and lacrimal glands, and biopsies of any abnormal appearing sites.²⁴ In a series of 140 consecutive patients, they found 25% of the gastric MALT lymphoma patients had multi-organ involvement and 46% of the nongastric MALT lymphoma patients had multi-organ involvement. However, there was no difference in the PFS or OS between the localized and disseminated cases, raising the question of the value of such an exhaustive staging evaluation. Perhaps a more reasonable approach would be to conduct standard lymphoma staging plus more specific studies based upon the findings in the history and physical examination (Table 1 ). The role of PET scanning for staging and monitoring MALT lymphomas has been evaluated.²⁵ While it may be useful in selected cases, in general the sensitivity is low (less than 60%) and routine use does not seem warranted.

Extranodal marginal zone lymphomas have important site-specific associations; thus, when considering MALT lymphomas, the first principle is location and it is conceptually useful to distinguish gastric MALT lymphomas from nongastric MALT lymphomas.

Gastric MALT lymphoma comprises about 30% of all MALT lymphomas. Most patients with gastric MALT lymphoma present with symptoms of dyspepsia, reflux, pain, nausea, or weight loss. At endoscopy, erythema, erosions, or ulceration may be seen. Masses are uncommon. The histologic diagnosis can be challenging due to the heterogeneity of cells within the tumor specimen, but the presence of lymphoepithelial lesions (invasion of tumor cells into individual glands) is a strong indicator of gastric MALT lymphoma. Once the diagnosis is confirmed, the next task is to establish the H pylori status. Gastric MALT lymphoma is uniquely associated with H pylori infection (present in 90% of cases) and regression of gastric MALT lymphoma is expected in about two-thirds of cases after H pylori eradication therapy.²⁶ A variety of detection methods are available, including identification of the microbes in the histologic specimen, a biopsy urease test, a urea breath test, a stool antigen test, and serology. At the time of hematology/oncology referral, usually the histologic test plus one other test have been completed. If both tests are negative, it is worthwhile to perform the serologic test. If any test is positive, it is reasonable to consider the patient an appropriate candidate for H pylori eradication therapy.

Upon establishment of the diagnosis, it is then worthwhile to perform a FISH evaluation for a t(11;18)(q21;q21). This translocation is recurrent in MALT lymphomas and produces a fusion between the AIP2 gene and the MLT/ MALT1 gene, generating a chimeric protein that promotes cell survival and proliferation via activation of the transcription factor NF-κB²⁷ (Figure 2; see Color Figures, page 506). This translocation can be detected in 30% to 40% of gastric MALT lymphoma cases and has clinical significance. Patients harboring t(11;18) are more likely to have widely disseminated disease, are more likely to be H pylori negative, and are less likely to respond to H pylori eradication. In one series, only 2 of 44 patients with t(11;18) responded to eradication therapy and both later relapsed.²⁸ The t(1;14)(p22;q32) translocation is also found in gastric MALT lymphoma, but less frequently (~4% of cases). This translocation results in deregulation of the BCL10 gene with subsequent activation of NF-κB (Figure 2; see Color Figures, page 506). Gastric MALT lymphoma with t(1;14) is also unlikely to respond to H pylori eradication therapy.²⁹ 

The optimal therapeutic approach will depend upon the H pylori status and the stage. Patients most likely to respond to H pylori eradication are those who are infected with stage IE disease and without a t(11;18). In one series, 47 of 63 (75%) of such patients responded to H pylori eradication therapy.²⁸ In the same series, only 1 of 21 (5%) of stage IIE patients responded to eradication therapy and only 2 of 44 (5%) with t(11;18) responded to eradication therapy. Eradication can be accomplished in several ways (Table 2 ). There is no consensus on how to treat H pylori eradication failures or H pylori–negative disease. Since gastric MALT lymphomas are often localized and tend to be extremely sensitive to low doses of radiation, our preferred approach is radiotherapy. In a series from Memorial Sloan-Kettering Cancer Center, this strategy yielded an overall response of 100% with excellent durability.³⁰ Another group demonstrated a 96% CR rate and 87% long-term DFS with radiotherapy as the sole treatment modality.³¹ Surgery is not warranted given the effectiveness of other modalities.³² Patients with advanced stage disease should be treated according to the same principles as for other advanced-stage indolent lymphomas (single-agent rituximab, R-CVP, R-CHOP, R-fludarabine, etc.). The choice of therapy will depend upon patient and disease factors such as tumor burden, symptoms, and comorbidities. There are no comparative trials to suggest the superiority of one approach over another. Single-agent rituximab produced objective responses in 77% of patients in a small phase II study.³³ Most chemotherapy combinations report response rates of over 90% and purine analogue-based therapy appears to have more efficacy but also more toxicity than alkylator-based therapy.³⁴ In fact, one publication indicated cases with t(11;18) tended to be poorly responsive to oral alkylator therapy.³⁵ 

Regression of established disease can be quite slow after H pylori eradication therapy, with some reports indicating a full year before complete remission was established. It is reasonable to repeat endoscopy approximately 8 to 12 weeks after completing eradication therapy (ideally with the same endoscopist). Some improvement in the appearance of the stomach wall should be apparent, although normalization is not likely at this early time point. Optimal monitoring is unclear, but a reasonable strategy is to repeat the endoscopy every 3 to 6 months until the appearance of the stomach wall has normalized and then annually for several years thereafter.³⁶ A difficult issue is whether to re-biopsy normal appearing mucosa at the end of therapy. Patients desire to know whether the disease is “gone.” The histologic determination of complete remission can be extremely difficult for the hematopathologist, due to residual microscopic chronic inflammation. Using polymerase chain reaction for detection of clonal B-cell receptor rearrangement, residual “disease” is often detected.³⁷ However, this is likely clinically insignificant. Another difficult issue is monitoring once remission is established. Biopsies of normal appearing mucosa can suggest microscopic recurrence, creating angst for the patient and the physician. However, in one series, 16 of 16 patients achieved a “second CR” with just watchful waiting, which casts doubt on the clinical significance of such a finding.³⁸ Therefore, we would recommend against biopsy of normal appearing mucosa as part of surveillance.

Nongastric MALT lymphomas comprise about 70% of MALT lymphomas. Common presenting sites include the parotid and salivary glands, conjunctiva, lungs, skin, intestinal tract, thyroid, and breast. There are scattered reports implicating infectious agents in the pathogenesis of certain nongastric MALT lymphomas (cutaneous MALT lymphoma, B burgdorfi; conjunctival MALT lymphoma, C psittaci; immunoproliferative small intestinal disease, C jejuni).³⁹ The frequency of these associations and responses to antimicrobial therapy are areas of ongoing investigation. The t(14;18)(q32;q21) translocation, which brings the MALT1 gene under the control of the IGH enhancer, is found more frequently in nongastric MALT lymphomas than gastric MALT lymphomas (Figure 2; see Color Figures, page 506).

The nongastric MALT lymphomas typically have an excellent prognosis, with most series demonstrating 5-year OS in excess of 90% and 10-year OS in excess of 80%.^40,41 Because these are often the most indolent of the indolent lymphomas, it is important to avoid over-treatment. Patients with limited stage disease can often be managed very successfully with judicious use of involved field radio-therapy. A recent presentation at an international conference reported on 166 patients with MALT lymphoma (including 22 patients with gastric MALT) treated with radio-therapy (typically 25–30 Gy).⁴² The complete response rate was 99% and the 10 year relapse-free survival rate and overall survival rate were 77% and 87% respectively, indicating the excellent outcome that can be achieved with this approach. Most of the relapses occurred within the first 5 years of follow up. Transformation to large cell lymphoma occurred in just 4 patients, suggesting a lower risk than noted in follicular lymphoma. As mentioned previously, with extensive staging advanced-stage disease will be detected in 46% of patients. For such patients, an approach similar to other indolent lymphomas is recommended. Asymptomatic patients with low tumor burden can be initially managed with a watch-and-wait strategy. Patients requiring therapy can be treated using systemic options appropriate for other indolent lymphomas.

Marginal zone lymphomas are uncommon indolent B-cell malignancies. The biology is unique and, for a subset, the pathogenesis is critically related to chronic antigenic stimulation. Given the characteristic clinical features associated with NMZL, SMZL, and the MALT lymphomas, a distinct clinical approach for each entity is required.