Guideline for the diagnosis and management of marginal zone lymphomas: A British Society of Haematology Guideline

The objective of this guideline is to provide healthcare professionals with clear guidance on the diagnosis and management of patients with marginal zone lymphoma (MZL). These guidelines were compiled according to the BSH process: https://b-s-h.org.uk/media/16732/bsh-guidance-development-process-dec-5-18.pdf. The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) nomenclature was used to evaluate levels of evidence and to assess the strength of recommendations. The GRADE criteria can be found at http://www.gradeworkinggroup.org. Recommendations are based on a review of the literature using Medline/Pubmed. Search terms included: marginal zone, MZL, extranodal MZL, MALT, nodal, splenic, treatment, randomised, clinical trial, radioimmunotherapy, hepatitis C, Helicobacter pylori. The search was limited to English language publications and conference abstracts from 1 January 1998 to 20 September 2022. Titles/abstracts obtained were curated and manually reviewed by the writing group, which conducted additional searches using subsection heading terms. The manuscript was reviewed by the BSH Guidelines Haemato-oncology Task Force, the BSH Guidelines Executive Committee and the haemato-oncology sounding board of the BSH. The MZLs are a group of clinically indolent mature B-cell lymphomas derived from memory B cells of the ‘marginal’ zones of secondary lymphoid tissues. Marginal zone B cells are at the centre of inflammation, autoimmunity and malignant transformation through the coordination of innate and adoptive immunity. MZLs, especially extranodal marginal zone lymphomas (EMZL), frequently arise in the context of chronic infection or autoimmune disease, and, while the various subtypes share many biological, diagnostic and clinical features, they manifest subtype-specific features, resulting in a multisystem presentation. The 5th World Health Organization (WHO) classification of tumours of haematopoietic and lymphoid tissues recognises distinct MZL subtypes according to the microenvironment of involved tissue1—EMZL of mucosa-associated lymphoid (MALT) tissue, splenic marginal zone lymphoma (SMZL) and nodal marginal zone lymphoma (NMZL).2, 3 Paediatric marginal zone lymphoma and primary cutaneous MZL, originally included under NMZL and EMZL/MALT, respectively, are now classified as separate entities. MZL is the third most common lymphoma,4 comprising up to 15% of non-Hodgkin Lymphoma (NHL) in the Western World. Over 60% are EMZL/MALT (which can arise from any site following chronic antigenic stimulation), 20% are SMZL and <10% are NMZL. Incidence increases with age, suggesting cumulative exposure to risk factors. Age-adjusted incidence has increased by 1.1% per year5 with a current UK incidence of 2.62 per 100 000 and a male-to-female ratio of 1.6.4 Diagnosis requires a representative tissue biopsy, bone marrow or peripheral blood sample, depending on the subtype. Diagnostic material should be reviewed by an expert haematopathologist6 in the context of clinical and laboratory features and classified according to the 5th WHO classification of haematolymphoid tumours and/or the International Consensus Classification of Lymphoid Neoplasms.2, 3 Which classification has been used should be stated in the report. There are currently no widely specific diagnostic markers for MZL. Immunohistochemical (IHC) evaluation of tissue relies upon excluding other low-grade B-cell NHL entities. Commonly used IHC markers are listed in Table 1. EMZL/MALT recapitulates Peyer's patch-type lymphoid tissue2 and presents at a variety of extra-nodal sites. Tumours are composed of morphologically heterogeneous small B cells, including marginal zone (centrocyte-like) cells, monocytoid cells, small lymphocytes and centroblast-like cells. There is plasmacytic differentiation in some cases. Neoplastic cells reside in the marginal zones of reactive B-cell follicles, extend into the interfollicular region and colonise the follicles. In epithelial tumours, lymphoepithelial lesions are seen. These are aggregates of ≥3 neoplastic cells with distortion or destruction of the epithelium, often with eosinophilic degeneration of epithelial cells. These are not essential for diagnosis and are not specific for EMZL.7 Helicobacter pylori infection is strongly implicated in the pathogenesis of gastric EMZL/MALT.8 Autoimmunity-associated chronic inflammation (Sjögren syndrome and Hashimoto thyroiditis) may precede EMZL/MALT-affecting salivary glands and thyroid respectively. Primary cutaneous marginal zone lymphoproliferations are now recognised as a distinct entity because of their indolent behaviour and disease-specific survival approaching 100% without the need for aggressive therapies.3 The term ‘primary cutaneous marginal zone LPD’ is proposed in the International Consensus Classification of mature lymphoid neoplasms. The 5th WHO classification of haematolymphoid tumours categorises the same entity as a ‘Primary cutaneous marginal zone lymphoma’.2 Approximately 75% are class-switched (predominantly IgG+), with up to 40% expressing IgG4. Abundant reactive T cells and peripherally located clustered plasma cells are often seen. Cases that are IgM+ (non-class-switched) and show monocytoid B cells warrant exclusion of non-cutaneous primary disease.3 MZLs rarely involve the central nervous system (CNS), either primarily or secondarily. Most CNS MZL is dural EMZL/MALT lymphoma, with lesions often radiologically indistinguishable meningioma9-13; however, MRI may distinguish from meningioma by a more prominent dural tail.14 Histological confirmation has important implications for prognosis and management, as dural EMZL/MALT lymphoma has a better prognosis than high-grade primary central nervous system lymphomas.15, 16 IgG4 expression in light-chain-restricted clonal plasma cells is a feature of MZL involving the CNS. Other EMZL/MALT lymphomas may also develop in the context of IgG4-related disease, and an aetiological association has been suggested.17-21 Site-specific aetiology and reported genetic alterations22 are summarised in Table 2. A diagnosis can usually be established through a combination of morphology and flow cytometry on peripheral blood or aspirated bone marrow, marrow trephine biopsy histology and IHC, but in a minority of cases, a definitive diagnosis may require splenectomy.7 Splenic histology reveals a B-cell neoplasm composed of small lymphocytes that surround and replace the splenic white pulp germinal centres, efface the follicle mantle and merge with a peripheral (marginal) zone of larger cells, including scattered transformed blasts; both small and larger cells infiltrate the red pulp. SMZL shares features with other splenic lymphomas and, without splenectomy, may be indistinguishable from splenic diffuse red pulp lymphoma (Table 3). CD20+ CD103− CD25−/+ CD27+ CD11c+/− CD123− DBA44+ Annexin A1− Cyclin D1− IgD+ CD20 bright+ CD103−/+ CD25− CD27+ CD11c−/+ CD123− DBA44+ Annexin A1− Cyclin D1− IgG+ (rare IgD/IgM+ cases) CD20 bright+ CD103+ CD25+ CD27− CD11c+ CD123+ DBA44+ Annexin A1+ Cyclin D1+ (weak) IgG+ CD20 bright+ CD103+ CD25- CD27+ CD11c+ CD123− DBA44+ Annexin A1- Cyclin D1− IgG+ NMZL is a primary nodal B-cell neoplasm morphologically resembling lymph nodes involvement by EMZL/MALT or SMZL but without evidence of extranodal or splenic disease. Peripheral blood involvement may occur, and bone marrow involvement is seen in one third of cases.2, 41 Table 1 describes IHC that is useful in establishing the diagnosis. Where plasma cell differentiation is prominent, distinction from lymphoplasmacytic lymphoma (LPL) may be difficult. Demonstration of remnants of follicular dendritic cell meshworks favours a diagnosis of NMZL.2 MYD88 gene mutations are usually detected in LPL and rarely in NMZL. A mixture of cell types is seen in NMZL, but the presence of >20% large B cells is concerning for high-grade transformation (HGT).2 However, a diagnosis of diffuse large B-cell lymphoma should only be made if clearly delineated sheets of large B cells are identified. In some cases, an abundant PD1+ follicular helper T-cell infiltrate has been described and may pose a diagnostic challenge in distinction from T-cell lymphoma; this can also be seen in EMZL/MALT.42 Trisomy of chromosomes 3 and 18, gains of 2p and 6p and loss of 1p and 6q are common in NMZL, with frequent somatic variants of KMT2D, PTPRD, NOTCH2 and KLF2.2 Paediatric nodal MZL, now a separate entity in the 5th WHO classification,2 is an indolent disease with a favourable prognosis. It occurs predominantly in boys (M:F, 20:1) presenting with asymptomatic, localised disease involving lymph nodes of the head and neck. Involved nodes display progressive transformation of germinal centres.43 Differential diagnosis includes atypical marginal zone hyperplasia and marginal zone hyperplasia associated with Haemophilus influenzae, wherein the marginal zone cells are IgD-positive.44 Genetic studies are advised in this setting, in view of the differential diagnoses. Population-based long-term data on the transformation of MZL is limited; however, in some studies, a cumulative incidence of transformation to aggressive large B-cell lymphoma of 4.7% at 10 years is described.45 In this study, the highest risk of transformation was observed in patients with SMZL (14%); a range of 4%–15% is described in other studies.46, 47 Risk factors for transformation are discussed in Section “Transformed disease”. Histologically, the common form of transformation is to a diffuse large B-cell lymphoma (DLBCL) of non-germinal centre immunophenotype; a diagnosis of DLBCL should only be made if clearly delineated sheets of large B cells are identified.2 A variable proportion of large B cells are present within the neoplastic population in all histological subtypes of MZL. Rarer cases of transformation to classical Hodgkin Lymphoma and plasma cell leukaemia have been described.48, 49 A clonal relationship between the MZL and transformed disease can be demonstrated in many cases, but apparent transformations may in some cases represent clonally unrelated second lymphomas. All patients require a history and full physical examination, blood tests, radiological imaging for staging and baseline measurement of disease (summarised in Table 4). Gastric EMZL/MALT lymphoma accounts for 35% of MZL and ~50% of all EMZL. Rarer subtypes of gastrointestinal MALT lymphoma include small intestinal MALT or its variant immunoproliferative small intestinal disease (IPSID) and colonic MALT lymphoma.50 Patients with gastric EMZL/MALT lymphoma should undergo oesophago-gastro-duodenoscopy (OGD) with biopsies and careful documentation of lesions.51 Mapping biopsies are not essential, but high-quality photography and a detailed description of the site of lesions are advised for comparison. Repeat OGD is recommended in cases of diagnostic uncertainty. Fluorescence in situ hybridisation (FISH) for t(11:18)(q21;q21) and fusion of BIRC3 (formerly API2) and MALT1 is recommended in all cases, as its presence is associated with more advanced disease and a lower rate of response to H. pylori eradication.52 Helicobacter pylori is strongly implicated in the pathogenesis of gastric EMZL/MALT lymphoma. Despite the apparently decreasing incidence of H. pylori-positive gastric EMZL/MALT lymphomas,53 most cases are still thought to be H. pylori positive, particularly in older patients. H. pylori testing should be performed in every patient, with proton pump inhibitor (PPI) and antibiotics discontinued at least 2 weeks before.54 H. pylori infection is primarily evaluated by tissue IHC. In addition, faecal antigen testing (or a carbon-13 urea breath test; CLO) and Helicobacter serology are recommended. Serology is useful with low-level infection and if histology/CLO is negative.55 H. heilmannii in gastric EMZL/MALT and Campylobacter jejuni in IPSID are less frequent causative organisms.56, 57 Bone marrow involvement is rare—just 4.3% in one study58—and bone marrow biopsy is not essential except in patients with cytopenia. The role of positron emission tomography (PET) is not well defined in MZL; PET is, however, increasingly used in MZL for staging of both nodal and extranodal disease59 and subsequent response assessment if fluorodeoxyglucose (FDG) is avid at baseline.7, 60, 61 Since uptake in extranodal sites can vary by extranodal location and lesional size, it is anticipated that baseline scans may be used more often in the future to determine whether PET is the most appropriate modality for response assessment. Gastric EMZL/MALT lymphomas show variable FDG avidity, with reported rates of 50%–60%.59 A retrospective study reported inferior overall survival (OS) and a higher incidence of HGT in gastric EMZL/MALT lymphoma when standardised uptake value was ≥10.62 PET may therefore be considered when HGT is suspected. There is a lack of consensus on the best staging classification for gastrointestinal MALT lymphoma. The 1994 Lugano classification63 is most widely used in the UK but is based primarily on imaging, while the modified Ann-Arbor staging system64, 65 takes into account depth of infiltration and distant lymph node involvement. The more recent Paris staging involves a TNM system using the same principles.66 Table 5 compares the three staging systems. Stage I2E Stage II II1 = Regional LNs II2 = Distant intraabdominal LNs IIE = invasion of adjacent structures Non-gastric EMZL/MALT lymphomas involving the ocular adnexa, skin, lung, salivary gland, thyroid and breast have site-specific genetic profiles that may affect prognosis, in addition to the potential for organ-specific clinical considerations.68 Concomitant autoimmune disease is reported to be more frequent in non-gastric than gastric EMZL/MALT lymphomas, but the prognostic impact is unknown.69 Diagnosis and staging should be tailored to the site involved and include testing for underlying infectious or autoimmune causes (Table 6). Sjögren syndrome (lacrimal gland MZL) C. psittaci 3 months70 Reported range 3–36 months7, 71 Skin (9%) Lymphocytic interstitial pneumonia Achromobacter xylosoxidans SMZL involves the spleen, splenic and hilar nodes, bone marrow and frequently the peripheral blood as villous lymphocytes,23 and can present as isolated lymphocytosis. Cytopenias occur in 25% and are related to hypersplenism and less commonly to auto-antibodies or marrow infiltration.2, 75 Aside from splenic hilar nodal enlargement, lymphadenopathy and other organ involvement are rare at diagnosis.76 Work-up investigations include testing for associated autoimmune phenomena, which occur in ~20%, and testing for hepatitis C. Proposed investigations are summarised in Table 4. The majority of patients with NMZL present with disseminated, albeit often non-bulky nodal disease,77 without splenic or extranodal involvement. Bone marrow involvement is evident in one third of cases.41 Peripheral blood involvement is very rare.78 All patients should undergo computed tomography (CT) or PET/CT staging, which is also helpful to exclude nodal dissemination of EMZL/MALT, which occurs in one third of EMZL/MALT cases.78 HGT occurs at ~1% per year and conveys an inferior OS.79 The strong aetiopathogenic link with H. pylori makes eradication therapy with a PPI and two antibiotics (triple therapy) the mainstay of first-line therapy in gastric EMZL/MALT, irrespective of disease stage and H. pylori status. The NICE recommendation (CG184)80 is a 7-day, twice-daily course of PPI given with amoxycillin and either clarithromycin or metronidazole. In patients allergic to penicillin or with previous exposure to clarithromycin, quadruple therapy for 7 days with twice-daily PPI, bismuth, metronidazole and tetracycline is recommended. European guidelines (Maastricht VI/Florence consensus) recommend 14-day clarithromycin-based triple therapies on the basis of an increased cure rate without significantly increased side effects compared to 7-day regimens.54, 81 The importance of adhering to treatment to improve successful eradication should be explained to patients. Antibiotic therapy should be considered even when evidence of H. pylori is lacking, to cover false negative cases and other Helicobacter species.82 A meta-analysis of published studies reported an overall pooled CR rate of 29.3% in H. pylori-negative cases.53 Response to the first-line H. pylori eradication should be assessed with a stool antigen test or urea breath test ≥6 weeks after starting eradication and ≥2 weeks after stopping PPI.51 Late responses up to a year after treatment are reported, and ~62% achieve a CR 12 months after H. pylori eradication.83 Eradication is less effective in the presence of t(11;18)(q21;q21), when disease extends through the gastric serosa, or where there is perigastric lymph node involvement.84 After successful eradication therapy, a positive serology result may persist for up to 2 years, consistent with the previous infection; follow-up serology is therefore unhelpful to confirm H. pylori eradication. Second-line therapy is recommended for patients with a persistent positive result. Non-responding patients may be resistant to clarithromycin, and samples should be sent for culture and antibiotic sensitivity. In both H. pylori-positive and -negative cases, repeat OGD with multiple biopsies and comparison with previous biopsies is recommended at 3–6 months to assess response. Management thereafter is guided by OGD findings, symptoms and adverse signs such as deep invasion, overt progression, bulk or impending organ damage (Figure 1). Persistent clonality following H. pylori eradication is an acceptable outcome and can be monitored,85 as well as presence of microscopic disease in the absence of symptoms and endoscopic findings.86 The Groupe d'Etude des Lymphomes de l'Adulte (GELA) scoring system87, 88 may be helpful to document histological response when diagnostic biopsies are available for comparison (Table 7). Radiotherapy (RT) is indicated for stage I/IIE gastric EMZL/MALT lymphoma that has failed to respond to or relapses after H. pylori eradication and in the presence of overt progression, deep invasion, lymphadenopathy or presence of t(11;18).89-92 This is highly successful, with reported histological complete responses of 95%–100% on endoscopic follow-up.93-95 Involved-site radiotherapy (ISRT) should include the entire stomach and, if involved, adjacent lymph nodes. Treatment should be given with an empty stomach and planned with an advanced technique such as intensity modulation to minimise radiation dose to the heart, liver and kidneys.95-98 While most studies have included patients treated with ≥30 Gy, a recent series shows that 24 Gy is likely to be adequate97 and unlikely to cause significant acute toxicity. With appropriate RT planning, the dose delivered to critical structures can be kept well below organ tolerance, reducing the risk of long-term complications. Systemic therapy, together with H. pylori eradication, is indicated for first-line management of patients with advanced stage gastric EMZL/MALT if they are symptomatic, have deeply invasive disease, tumour bulk or impending organ damage. Systemic therapy treatment is discussed in Section “Management of relapsed MZL”. The prognosis of gastric EMZL/MALT lymphoma is excellent, with 5-year survival exceeding 90% and 75%–80% 10-year survival. The EMZL/MALT international prognostic index (IPI) is prognostic in both gastric and non-gastric EMZL/MALT. It defines three prognostic categories based on age >70 years, Ann Arbor stage III/IV and elevated lactate dehyrogenase (LDH) with 0, 1 or ≥2 risk factors predicting 5-year event-free survival of 70%, 56% and 29% respectively.99 In a study of 1408 gastric EMZL/MALT patients by Zullo et al., the annual recurrence rate was 2.2%. Due to a small increased risk of gastric cancer in patients with gastric EMZL/MALT lymphoma,100, 101 long-term follow-up with clinical examination, blood counts and OGD is recommended for surveillance of second cancers, including patients achieving complete lymphoma remission. Although the optimal follow-up interval is not defined, follow-up every 12–18 months under gastroenterology services is recommended, in line with EGILS and ESMO guidelines.51, 102 Unlike primary gastric EMZL/MALT lymphoma, less is known about the value of antibiotic therapy in non-gastric EMZL/MALT lymphoma. In patients with localised ocular adnexal EMZL/MALT where a causative organism has been identified, antibiotic therapy can be considered (Table 6). Doxycycline in C. psittaci-negative cases can cause disease regression in some cases and can be considered in first-line management of ocular adnexal EMZL/MALT,71-73 with a reported overall response rate (ORR) of 45%–65%.74 In other non-gastric EMZL/MALT lymphomas where a causative agent has been identified, for example Borrelia burgdorferi in cutaneous EMZL/MALT lymphoma, evidence is limited to case reports, and routine use of antibiotics is not recommended. The median time-to-response after antibiotics is 6 months, but it may be up to 36 months. Patients with hepatitis C-associated EMZL/MALT lymphoma should receive initial treatment with anti-viral therapy, which may lead to regression of disease and improved outcomes.103 Patients with non-gastric EMZL/MALT lymphoma who have completed treatment can be evaluated every 3 months for the first 2 years and every 6 months thereafter.7 Less than 5% of patients develop extra-cutaneous disease104 and bone marrow involvement is extremely rare.105 For this reason, routine bone marrow examination is not recommended for clinical staging.106 Where B. burgdorferi infection is endemic, PCR on a skin biopsy should be performed. Initial treatment of localised disease, in the absence of B. burgdorferi, is either involved lesion RT for T1 disease (solitary skin involvement107) or surgical excision.108 24 Gy is recommended for curative treatment of primary cutaneous marginal zone lymphoma based on randomised trial data.109 For T2 disease (regional skin involvement), with a small number of lesions, treatment for each individual lesion may also be beneficial.108 Where multiple lesions are present, a ‘watch and wait’ approach may be adopted, with treatment of symptomatic lesions with low doses of RT (4 Gy), rituximab or single agent or combination chemotherapy.106 RT offers excellent treatment for localised disease where regression is not achieved with antibiotic therapy. Moderate doses achieve excellent outcomes in EMZL/MALT, with 24 Gy now considered the standard of care based on the results of a large, randomised phase III trial.109, 110 Response rates exceed 90%, in-field recurrences are rare, and most patients are cured.111-113 A recent published phase II trial of involved field RT for localised non-gastric EMZL/MALT lymphoma reported 5-year progression-free survival (PFS) and OS of 79% and 95% respectively.114 Low-dose RT (4 Gy in two fractions) is also an option for Sjogren's-associated parotid EMZL/MALT lymphoma115, 116 and ocular EMZL/MALT lymphoma when there is concern about the risk of cataract or dry eye. Randomised data comparing RT to surgery, immunochemotherapy or surveillance are lacking, but several retrospective studies have demonstrated that patients treated with first-line RT have better outcomes, including OS.91, 111 A randomised study investigating the addition of adjuvant chemotherapy to RT for parotid EMZL did not show benefit.117 Advanced stage EMZL/MALT lymphoma is incurable, and for asymptomatic patients, a ‘watch and wait’ approach is appropriate. EMZL, when it disseminates, favours other extranodal sites rather than lymph nodes (Figure 2). RT can provide rapid and effective palliation for patients with advanced EMZL/MALT and symptomatic local sites. Very low doses (e.g. 4 Gy) produces good local control with minimal to no toxicity96, 110, 118 and can be repeated. Systemic therapy is indicated for patients with symptomatic, advanced disease, after failure of antibiotics or RT, or where there is evidence of HGT. The randomised IELSG-19 trial comparing rituximab, chlorambucil and rituximab–chlorambucil is the largest study in EMZL/MALT lymphoma, involving 454 patients. Rituximab–chlorambucil was more effective than monotherapy (EFS not reached vs. 5.1 years for chlorambucil and 5.6 years for rituximab),119 but did not improve OS. Treatment with rituximab plus chlorambucil is reasonable for elderly and frail patients; monotherapy of each can also be offered if combination therapy is not well tolerated. The phase II MALT2008-01 trial reported efficacy for bendamustine–rituximab (BR) in gastric and non-gastric EMZL/MALT. ORR was >85% and the estimated 5-year OS was 85.6%,120-122 with added scope to limit treatment to 4 cycles in patients achieving complete remission (CR). In clinical practice, the choice of BR must balance efficacy against higher toxicity, especially in older and/or comorbid patients. Two phase II trials from the same group evaluated RCVP (rituximab–cyclophosphamide–vincristine–prednisolone) in EMZL/MALT and NMZL. Reported 3-year PFS was 59% for RCVP without maintenance123 and 81% for RCVP with maintenance rituximab (375 mg/m2 every 8 weeks for up to 12 cycles).122 A subanalysis of previously untreated MZL patients enrolled in the phase III GALLIUM trial showed no improvement in outcomes for obinutuzumab compared to rituximab combinations, and obinutuzumab was also more toxic than rituximab.124 Obinutuzumab-chemotherapy is therefore not recommended for the treatment of MZL. Table 8 summarises frontline studies in EMZL/MALT. All subtypes R-chemo PFS 75% (3 years) BR 28 92, 20 71, 24 (ns) BR 37 93, 40 Median 57 months 91, 30 (ns) Median 47 months (ns) R 138 78, 56 EFS 50% (5 years) EFS 68% (5 years) PFS 72% (5 years) SMZL BR 61 Median 92 months 86% (6 years) Not reached (p = 0.008) 92% (6 years) (ns) NMZL BR 61 Median 92 months 86% (6 years) Not reached (p = 0.008) 92% (6 years) (ns) 15 12 Due to its rarity, therapy for CNS MZL is not standardised. ISRT or ISRT plus whole brain radiotherapy (WBRT) to 24–30 Gy is most common, especially for dural tumours15, 134 and achieves excellent long-term disease control10, 11, 14, 15, 134-136 even in the presence of concomitant leptomeningeal disease.135 Doses of 30–36 Gy produce high CR rates and low neurotoxicity,14, 135 with outcomes similar to chemotherapy alone and combined chemoradiotherapy.15 However, as this is an indolent lymphoma, lower doses (e.g. 24 Gy) may also be very effective and significantly less toxic.109, 110 The decision to use ISRT or WBRT depends on the site and number of lesions. Treatment of dural EMZL/MALT with systemic high-dose methotrexate (HD-MTX) and non-HD-MTX regimens are reported11 but combined chemoradiotherapy carries a high risk of neurotoxicity137 and is neither necessary nor recommended. Chemotherapy may be preferred for rare patients presenting with secondary CNS MZL. The literature includes cases treated with BR,138, 139 RCHOP, followed by maintenance rituximab14 and R-ibrutinib.140 The efficacy of rituximab monotherapy is uncertain.139 Patients with CNS MZL need long-term surveillance due to an increased risk of late disease recurrence.135 SMZL patients with hepatitis C infection should receive upfront anti-viral therapy.103, 141 Many patients are asymptomatic and can be monitored at 3–6 monthly intervals. Perrone et al. reported a median time-to-first-treatment of 58.5 months, and at 10 years, 30% of patients remained untreated.142 The main criteria for treatment include constitutional symptoms, symptomatic or progressive splenomegaly, bulky lymphadenopathy, autoimmune phenomena and progressive cytopenias,76 as shown in Table 9. There are no published randomised trials of treatment for SMZL. Systemic therapy options for symptomatic patients include rituximab monotherapy, immunochemotherapy and splenectomy, based on a few small prospective phase II trials.132 Rituximab monotherapy, 375 mg/m2 weekly for 6 weeks followed by 2 monthly maintenance for 1 year, is well tolerated and has achieved durable responses in retrospective data series. In the largest retrospective study of 106 patients, outcomes postinduction were 92% ORR and 44% CR. Ten-year PFS and OS rates were 64% and 85% respectively.133 As results are comparable to the largest retrospective series of 100 SMZL patients undergoing splenectomy (median PFS 8 years and 10-year OS 67%),143 rituximab has largely surpassed splenectomy in modern practice. Splenectomy may still have a role in managing fit patients with a large spleen and minimal bone marrow disease or lymphadenopathy.144 A splenectomy should be performed after appropriate vaccinations. Immunochemotherapy combining rituximab with CVP, CHOP and bendamustine has shown ORR and OS similar to rituximab monotherapy.145-147 In the BRISMA/IELSG36 involving 56 SMZL patients, ORR was 91% and CR 73%, with a 3-year PFS of 90% and a 3-year OS of 96%. Although highly effective, 25% of patients had a serious adverse event.132, 148 A large, retrospective observational study of 317 older, symptomatic patients showed no significant OS benefit of BR versus rituximab monotherapy.149, 150 Taken together, the evidence suggests that BR should be reserved for fitter/younger patients. The MAINTAIN randomised phase III trial demonstrated a PFS advantage for maintenance rituximab when given to SMZL and NMZL patients responding to induction BR or RCVP; PFS 92.2 months in the observation arm vs. not reached in the maintenance arm (HR 0.35, p = 0.008). No difference in OS at 6 years was observed.131 For patients unfit for systemic therapy or splenectomy, RT can provide rapid, effective palliation leading to a significant reduction of splenic size and pain, even with very low dose schedules of 4–10 Gy. A higher 24 Gy dose may provide more durable control but c