Outcomes in patients with splenic marginal zone lymphoma and marginal zone lymphoma treated with rituximab with or without chemotherapy or chemotherapy alone




The optimal management of patients with splenic marginal zone lymphoma/marginal zone lymphoma (SMZL) is controversial. The objective of this retrospective study was to compare the outcomes of patients with SMZL who received treatment with rituximab, rituximab plus chemotherapy, or chemotherapy alone.


The Leukemia Service database was searched for patients with splenic lymphoma who were registered between May 1995 and October 2004. The indications for treatment were the same as those used for patients with chronic lymphocytic leukemia.


SMZL was confirmed in 70 patients. The median age was 64 years. The median number of CD20 molecules per cell was 69 × 103. Forty-three patients required systemic therapy; rituximab in 26 patients, chemotherapy plus rituximab in 6 patients, and chemotherapy alone in 11 patients. Ten additional patients underwent splenectomy, and 17 patients were in the observation group. The overall response rates were 88% with rituximab, 83% with rituximab plus chemotherapy, and 55% with chemotherapy alone; the 3-year survival rates were 95%, 100%, and 55%, respectively. The 3-year failure-free survival (FFS) rates were 86%, 100%, and 45% in the rituximab, rituximab plus chemotherapy, and chemotherapy alone groups, respectively. Rituximab treatments resulted in longer survival and FFS compared with chemotherapy. Rituximab alone resulted in disappearance of splenomegaly in 92% of patients and normalization of absolute lymphocyte counts. In univariate analysis, younger age and rituximab-based therapy were predictive of longer FFS.


Rituximab with or without chemotherapy was found to have major activity in patients with SMZL. These results may be associated with high levels of cellular CD20 antigen sites. Rituximab should be the treatment of choice, at least in older patients with SMZL who have comorbid diseases. Cancer 2006. © 2006 American Cancer Society.

Splenic marginal zone lymphoma (SMZL) with or without villous lymphocytes is a disorder that recently was recognized as a distinct pathologic entity in the World Health Organization classification.1 SMZL is a B-cell neoplasm that comprises <1% of lymphoid malignancies, although it may account for most of the otherwise unclassifiable chronic lymphoid leukemias that are negative for CD5.2 It was described first in 1987 as splenic B-cell lymphoma with circulating “villous” lymphocytes (SLVL).3 SLVL is indistinguishable histologically from SMZL, and it is believed that the 2 entities represent different phases of the same process in the spleen and the bone marrow.4, 5 The term “marginal zone B-cell lymphoma” (MZL) describes mucosa-associated lymphoid tissue lymphoma and nodal marginal zone B-cell lymphoma.6

The optimal treatment for SMZL has not been defined. In previous years, splenectomy was considered the treatment of choice for SMZL. The “watch-and-wait” approach commonly is preferred in asymptomatic patients with no immediately threatening disease. However, guidelines for deferring therapy have not been defined as they have been for chronic lymphocytic leukemia (CLL),7 and this choice usually is made for older patients who have disease with a slow natural history. Although SMZLs are considered “indolent” lymphomas, approximately 50% of deaths from SMZL are caused by progressive disease.8–13

More recently, numerous treatment strategies have become available for patients with B-cell lymphomas. The nucleoside analogs fludarabine and pentostatin and the chimeric anti-CD20 monoclonal antibody rituximab have demonstrated significant antitumor activity as single agents.14–19 Clinical trials have incorporated rituximab into combination chemotherapy regimens, and results suggest that the addition of rituximab improves response rates and prolongs survival in patients with non-Hodgkin lymphoma.20–22 The combination of fludarabine, mitoxantrone, dexamethasone, and rituximab (R-FMD) has significant activity in indolent lymphoma,23 and the combination of fludarabine, cyclophosphamide, and rituximab (FCR) also is effective in CLL.24 The objective of this retrospective study was to compare the complete remission (CR), failure-free survival (FFS), and overall survival (OS) rates in patients with SMZL who were treated with rituximab, chemotherapy plus rituximab, or chemotherapy alone.


A database that included all untreated and treated patients with chronic lymphoproliferative disorders who were referred to the Leukemia Service was searched for patients with splenic lymphoma who were registered between May 1995 and October 2004. Patients were assessed prior to their first treatment at The University of Texas M. D. Anderson Cancer Center.

Available pathology slides were reviewed by 2 hematopathologists to confirm the diagnosis of SMZL/MZL. All records were reviewed to determine clinical, laboratory, and pathologic features at presentation in addition to disease stage, treatment, and outcome. Staging evaluations at the time of presentation included complete physical examinations, bone marrow aspirates and biopsies, spleen or lymph node biopsies, chest radiography, and computed tomography scans of the chest, abdomen, and pelvis.

In the current study, at least 2 of the following criteria had to be met to make the diagnosis of SMZL: immunophenotype with a “CLL score” ≤2,25 any degree of splenomegaly, a nodular/interstitial or diffuse pattern and/or an intrasinusoidal pattern in the bone marrow biopsy sample, spleen histology typical of SMZL. and exclusion of CLL follicular lymphoma or mantle cell lymphoma in the examination of peripheral blood and/or bone marrow. Staging and treatment were determined after review of all clinical, laboratory, and pathologic data in a multidisciplinary conference. Standard or investigational treatment was administered either at our center or in the community by collaborating physicians. The evaluation of response for all patients who were included in the study was performed at M. D. Anderson Cancer Center. Signed informed consent was obtained before all procedures and before all experimental therapy, as required by the Institutional Review Board.

The indications for therapeutic intervention in patients with SLVL/SMZL were similar to those for patients with CLL/SLL. Rituximab (Rituxan; Genentech, Inc., San Francisco, CA) was given at standard doses intravenously every week for 4 weeks or 8 weeks. FCR or R-FMD was given as described previously.23, 24 The chemotherapy group was composed of patients who received other cytotoxic agents, not including rituximab or alemtuzumab therapy.

Immunophenotypic Analysis

Flow cytometric immunophenotypic studies were performed on bone marrow or peripheral blood. An immunophenotype score was generated by using the system described by Matutes et al.25 Assessment of the quantitative expression of CD20 molecules on the cell surface was performed as described previously.26

Endpoints and Statistical Methods

CR was defined as the complete disappearance of all detectable clinical and radiographic evidence of disease, the disappearance of all disease-related symptoms, and normalization of biochemical abnormalities that definitely were attributable to lymphoma for at least 1 month. Unconfirmed CR (CRu) included patients with minimal stable radiographic changes or with persistent lymphoid aggregates in the bone marrow without atypia. Partial remission (PR) was defined as a reduction ≥50% in the sum of the products of the greatest dimensions of bidimensionally measurable disease.27 Any lesser response was considered a failure. The response and endpoint assessments conformed to published International Workshop response criteria.27 Survival was measured from the start of the first treatment at M. D. Anderson or from observation for patients in the “watch-and-wait” group to death from any cause or last follow-up. Treatment failure was defined as progression, recurrence, or death. FFS was defined as the time from the start of the first treatment or observation until progression, recurrence, or death.

The Wilcoxon signed rank test was used for comparison of the paired data (before therapy vs. after therapy) for continuous variables. The Wilcoxon rank sum test was used to examine the difference between 2 groups of patients for continuous variables. Survival curves were estimated using the Kaplan–Meier method. The 2-sided log-rank test was used to test the association between variables and survival or FFS. A multivariate Cox proportional hazards regression model was fit to examine the impact of risk factors on FFS after adjusting for other factors. P values <.05 were considered statistically significant. Statistical analyses were performed using SAS software (version 8.2; SAS Institute, Cary, NC) and S-Plus software (version 2000; Insightful Corp., Seattle, WA).



In total, 93 patients were registered in the database with a diagnosis of SMZL. After review of the available pathology slides, the diagnosis was confirmed in 70 patients. Twenty-three patients were excluded from the study for the following reasons. Eighteen patients had CLL, 2 patients had mantle cell lymphoma, and 3 patients did not have slides available for review because they were returned to their referring physician.

The clinical and laboratory characteristics of the 70 patients at the time they received their first treatment at M. D. Anderson are summarized in Table 1. Forty patients (57%) were age > 60 years. There were 39 men and 31 women. Two patients (3%) had a Zubrod performance status >1. Fifteen patients (22%) had Rai stage 0 or 1 disease, 22 patients (32%) had Rai stage 2 disease, and 32 patients (46%) had Rai stage 3 or 4 disease. Forty-seven patients (67%) had splenomegaly. Sixty-seven patients had SMZL, and 3 patients had predominantly nodal MZL with concurrent bone marrow involvement.

Table 1. Presenting Characteristics of 70 Patients with Splenic Marginal Zone Lymphoma/Marginal Zone Lymphoma
  • *

    The upper limit of normal for lactate dehydrogenase at M. D. Anderson Cancer Center is 618 IU/L.

  • Forty-seven patients (67%) had an enlarged spleen.

  • Sixty-one of 64 patients (95%) had CD20 levels >20% in their bone marrow by immunophenotype.

Age, y6433–88
Lactate dehydrogenase (IU/L)*568334–1861
β2-microglobulin (mg/dL)3.62–10
Hemoglobin (g/dL)12.08–16
Platelets (× 109/L)1611–498
Absolute lymphocyte count (× 109/L)12.10.2–164
Bone marrow cellularity (%)4015–100
Bone marrow lymphocytes (%)475–82
Spleen size (cm below costal margin)40–20
Bone marrow CD20 (%)781–97
CD20 molecules/cell (× 103)6916–260

Fifty-nine patients (84%) had absolute lymphocyte counts >5 × 109/L. Twenty of 33 patients (61%) who were evaluated had monoclonal gammopathy (immunoglobulin G [IgG], 6 patients; IgA, 2 patients; and IgM, 12 patients). Table 2 shows the results of immunophenotyping in bone marrow and the CLL scores and their components. The bone marrow infiltration pattern was interstitial in 34% of patients, nodular and interstitial in 24% of patients, nodular in 20%, diffuse and interstitial in 12%, diffuse in 6%, nodular and diffuse in 2%, and nodular, interstitial, and diffuse in 2% of patients. The median number of CD20 was 69 × 103 molecules per cell (range, 16–260 × 103 molecules per cell). These results were compared with the results from patients with CLL/SLL, in which the median number of CD20 molecules per cell was 11.5 × 103 molecules per cell and ranged from 0.6 × 103 molecules per cell to 148 × 103 molecules per cell. A significantly greater number of CD20 molecules per cell was observed in patients with SMZL (P<.0001; Wilcoxon rank sum test).

Table 2. Immunophenotyping in Bone Marrow
ParameterNo. of Patients%
  • +: positive; −: negative; CLL: chronic lymphocytic leukemia.

  • *

    See Matutes et al.25

CD22− or CD22+/−8/5515
Surface immunoglobulin weak expression7/5812
CLL score*  

Seven patients (10%) had a history of hepatitis B (3 patients), hepatitis C (2 patients), or both (2 patients). One of 2 patients who had both hepatitis B and C also had documented cryoglobulinemia with a positive direct Coombs test result.


Forty-three patients required systemic therapy. Patient characteristics according to therapeutic intervention are shown in Table 3. Twenty-six patients received immunotherapy (rituximab, 25 patients; alemtuzumab, 1 patient); 6 patients received chemotherapy plus rituximab (FCR, 5 patients; R-FMD, 1 patient), and 11 patients received chemotherapy (fludarabine, 2 patients; fludarabine plus prednisone, 1 patient; fludarabine plus cyclophosphamide, 3 patients; cladribine, 3 patients; fludarabine, cyclophosphamide, and granulocyte-macrophage–colony-stimulating factor [GM-CSF], 1 patient; and fractionated cyclophosphamide, vincristine, liposomal daunorubicin, dexamethasone, and GM-CSF, 1 patient). Pretreatment characteristics were similar between the 3 groups (Table 3). Ten patients underwent splenectomy alone, and 7 additional patients who underwent splenectomy required subsequent systemic therapy. In 17 patients, a “watch-and-wait” approach was used. Patients in the splenectomy and observation groups had normal absolute lymphocyte counts, as expected, compared with high rates of patients with elevated absolute lymphocyte counts in the rituximab, chemotherapy plus rituximab, and chemotherapy groups (P = .0001) (Table 3).

Table 3. Pretreatment Characteristics in 70 Patients with Splenic Marginal Zone Lymphoma/Marginal Zone Lymphoma by Therapeutic Intervention
VariablePercentage of PatientsPPercentage of PatientsP
Rituximab (n = 26)*Chemotherapy and Rituximab (n = 6)Chemotherapy (n = 11)Splenectomy (n = 10)“Watch and Wait” (n = 17)
  • CLL: chronic lymphocytic leukemia; LDH: lactate dehydrogenase; ULN: upper limit of normal.

  • *

    Includes 25 patients treated with rituximab and 1 patient treated with alemutzumab.

  • The P values were calculated by comparing pretreatment characteristics between the 3 treatment groups (i.e., rituximab, chemotherapy and rituximab, and chemotherapy) (Fisher test).

  • The P values were calculated by comparing baseline characteristics between the 5 groups (i.e., rituximab, chemotherapy and rituximab, chemotherapy, splenectomy, and the “watch and wait” approach) (Fisher test).

  • §

    See Matutes et al.25

Age >60 y655064.824053.67
Zubrod performance status >1800.9900.82
Rai stage >2506727.268029.06
CLL score >2§4339.0900.06
No prior therapy818391.8590100.32
β2-microglobulin >4 mg/L545027.355019.12
LDH >1 × ULN (618 IU/L)425064.474035.67
Hemoglobin <11 g/dL381736.744024.75
Absolute lymphocyte count >5 × 109/L856788.4400.0001

Treatment was deemed necessary at the time of disease progression, which was defined as follows: 1) for rituximab (26 patients), symptomatic splenomegaly (7 patients), progressive organomegaly and lymphadenopathy (6 patients), symptomatic worsening anemia (5 patients), B-symptoms (3 patients), blood lymphocyte count doubling in <1 year (3 patients), thrombocytopenia (1 patient), and lymphoma-associated collagen vascular disease (1 patient); 2) for chemotherapy plus rituximab (6 patients), symptomatic splenomegaly and lymphadenopathy (4 patients), B-symptoms (1 patient), and worsening anemia and thrombocytopenia (1 patient); and 3) for chemotherapy (11 patients), symptomatic splenomegaly (5 patients), symptomatic cytopenias (2 patients), blood lymphocyte count doubling in <1 year (2 patients), autoimmune hemolytic anemia (1 patient), and B-symptoms (1 patients). Assignment to different therapies was based on therapy availability (e.g., patients who required systemic therapy prior to the rituximab era were treated with chemotherapy alone), patient preference, and physician discretion.

Response to Therapy

Overall, 26% of patients who received systemic therapy achieved CR, 9% of patients had a CRu, and 44% of patients achieved PR. Response by therapy is summarized in Table 4.

Table 4. Response in 43 Patients who Required Systemic Therapy*
ResponseNo. of Patients (%)Total (n = 43)
Rituximab (n = 26)Chemoimmunotherapy (n = 6)Chemotherapy (n = 11)
  • CR: complete remission; CRu: unconfirmed CR; PR: partial remission.

  • *

    Among the 10 patients who underwent splenectomy only, 3 patients had a CR and 3 patients had a PR.

  • Twenty-five of 26 patients received rituximab.

  • Chemotherapy plus rituximab.

CR8 (31)1 (17)2 (18)11 (26)
Cru3 (12)1 (17)04 (9)
PR12 (46)3 (50)4 (36)19 (44)
CR, CRu, and PR23 (88)5 (83)6 (55)34 (79)


The event charts in Figure 1 show the patients' status. The median follow-up of surviving patients was 2.7 years (range, 0.1–16.9 years). Fifteen patients have died. Causes of death were as follows: progressive disease in 5 patients (chemotherapy, 4 patients; alemtuzumab, 1 patient), secondary cancer in 3 patients (watch-and-wait, 2 patients; splenectomy, 1 patient); progressive multifocal encephalopathy, 1 patient (chemotherapy), severe congestive heart failure, 1 patient (rituximab), liver failure, 1 patient (splenectomy), and unknown in 4 patients (chemotherapy, 3 patients; rituximab, 1 patient). The median survival duration of the 70 patients was 8.5 years (95% confidence interval, 6.3–22.1+ years). Eighty-six percent of patients remained alive at 3 years (Fig. 2A).

Figure 1.

(A) Disease course and survival status of 51 patients who received 1 treatment modality at The University of Texas M. D. Anderson Cancer Center (MDACC): rituximab alone (Rit), immunotherapy (ImmunoRx), chemotherapy (Chemother) and rituximab (Rit + Chemoth), chemotherapy (chemoth), splenectomy, or a “watch-and-wait” approach. (B) Disease course and status are illustrated for 19 patients who required >1 treatment (Rx) modality.

Figure 2.

(A) Overall survival in 70 patients with splenic marginal zone lymphoma/marginal zone lymphoma (SMZL). (B) Failure-free survival in 70 patients with SMZL.

Survival rates were higher in the chemotherapy plus rituximab group than in the rituximab group and higher in the rituximab group than in the chemotherapy group (P = .047) (Fig. 2B). The median follow-up of surviving patients in the rituximab group was 2.5 years (range, 0.1–5.2 years), and 3 of 26 patients (12%) have died. No deaths were noted among the 6 patients in the chemotherapy plus rituximab group after a median follow-up of 3.4 years (range, 0.2–5.2 years). Eight of 11 patients (73%) in the chemotherapy group died after a median follow-up of 6.8 years (range, 6.3–16.9 years). Two of 10 patients (20%) in the splenectomy group have died after a median follow-up of 2.4 years (range, 0.8–7.6 years); at 3 years, 89% of those patients remained alive. The median follow-up of patients in the “watch-and-wait” group was 2.7 years (range, 0.2–7.1 years). Two of 17 patients (12%) in that group have died, and their 3-year survival rate was 83%.

On univariate analysis, the only factors that correlated with longer survival in the 70 patients were age younger than 60 years (P = .01) and rituximab with or without chemotherapy (P = .047). There was no correlation between survival and lactate dehydrogenase or β2–microglobulin levels, absolute lymphocyte counts, or percentage of CD20 in the bone marrow or peripheral blood.

Failure Free Survival

The median FFS duration of 70 patients was 5.7 years. Seventy-five percent of 70 patients remained failure free at 3 years (Fig. 3A). Treatment failure occurred in a total of 26 patients, including 6 of 26 patients (23%) in the rituximab group, 1 of 6 patients (17%) in the chemotherapy plus rituximab group, and 11 of 11 patients (100%) in the chemotherapy group (P = .06) (Fig. 3B).

Figure 3.

(A) Overall survival in 43 patients by therapy. (B) Failure-free survival in 43 patients by therapy. Chemo: chemotherapy; Immuno: immunotherapy.

Five of 10 patients (50%) in the splenectomy group experienced treatment failure, and their 3-year FFS rate was 80%. Three of 17 patients (18%) in the observation group developed progressive disease or died. No factor was identified that predicted longer FFS. There was no significant difference in FFS between the 26 patients who received rituximab and the 10 patients who underwent splenectomy as their first treatment at M. D. Anderson (P = .56).

Response, Survival, and FFS with Rituximab With or Without Chemotherapy

Nineteen patients received >1 therapy. Therefore, some patients received sequential chemotherapy plus rituximab and rituximab alone. To examine the role of these therapies in clinical outcome, patients who received chemotherapy plus rituximab combination therapy at any time, irrespective of other therapies, were considered. In total, 12 patients had received chemotherapy plus rituximab at 1 time. The median follow-up of surviving patients in this group was 2.9 years (range, 0.2–5.2 years), and their overall response rate was 83% (CR, 4 patients; CRu, 2 patients; and PR, 4 patients). The median survival duration has not been reached; however, 100% of patients remained alive at 3 years (Fig. 4A). One patient (8%) died at 3.5 years. The median FFS duration is 3.9 years, and 64% of patients had not experienced treatment failure at 3 years (Fig. 4A).

Figure 4.

(A) Survival and failure-free survival in patients who received chemotherapy plus rituximab at any time. (B) Survival and failure-free survival in patients who received rituximab at any time.

In total, 31 patients had received rituximab at 1 time. The overall response rate was 93.5% (CR, 10 patients; CRu, 3 patients; and PR, 16 patients). The median follow-up of surviving patients was 3.1 years (range, 0.1–6.1 years). At the time of last follow-up, 3 patients had died. One patient died of preexisting congestive heart failure at age 84 years, with stable disease 11 months after rituximab therapy; another patient died of progressive disease at age 76 years, 5.6 years after rituximab therapy and subsequent salvage therapies; and the third patient was lost to follow-up and died at age 73 years, 5 years after rituximab therapy. The median survival duration had not been reached at the time of last follow-up, and 97% of patients remained alive at 3 years (Fig. 4B). The median FFS duration was 2.9 years, and 88% of patients had not experienced treatment failure at 3 years (Fig. 4B).

Comparison of Rituximab with Splenectomy

Changes in bone marrow and blood counts among patients who received rituximab (25 patients) were compared with changes among patients who underwent splenectomy as initial therapy (17 patients, including 7 patients who received subsequent therapies) (Table 5). Rituximab resulted in the disappearance of a palpable spleen in 23 patients (92%). Rituximab also was superior to splenectomy in normalizing white blood cell counts (P<.001) and absolute lymphocyte counts (P<.001). Splenectomy resulted in higher platelet counts compared with rituximab (P<.001). Differences in hemoglobin levels and bone marrow cellularity did not reach statistical significance, but a trend toward a significantly lower proportion of lymphocytes was noted in patients who received rituximab (P = .10).

Table 5. Changes in Complete Blood Counts and Bone Marrow According to Treatment with Splenectomy or Rituximab
VariableSplenectomy (n = 17)*Rituximab (n = 25)Wilcoxon Test P
Median (Range)No. of Patients*Wilcoxan Test For Paired Data (P)ChangeMedian (Range)No. of Patients*Wilcoxan Test For Paired Data (P)Change
  • Hgb: hemoglobin; WBC: white blood cells; ALC: absolute lymphocyte count; BM cell: bone marrow cellularity; BM lymph: bone marrow lymphocytes.

  • *

    Values shown are the number of patients with available paired data.

  • The Wilcoxon signed rank test was used for comparison of the paired available data (before vs. after therapy) for continuous variables.

  • The Wilcoxon rank sum test was used to examine the difference of changes between the 2 groups of patients for continuous variables.

Hgb (g/dL)10.1 (7.5–13.4)12.3 (8.6–15.1)15.121.1 (− 2.2, 4.3)11.5 (7.1–14.7)12.7 (7.8–16)23.0011 (− 1.1, 4.6).53
WBC (× 109/L)15.2 (2.5–64.1)16.5 (9.1–68.8)15.0039.2 (− 3.2, 44.9)21.2 (1.8–181)4.4 (1.9–12.8)23<.001−15.4 (− 176, 1.7)<.001
Platelets (× 109/L)112 (1–216)321 (123–713)15<.001178 (− 58, 682)143 (52–288)180 (28–293)23.00731 (− 65, 109)<.001
ALC (× 109/L)6.5 (0.4–48.9)4.7 (0.9–58.5)14.240.6 (− 12.7,13)17.2 (0.4–164)1.19 (0.5–2.9)23<.001−14.4 (− 163, 0.4)<.001
Spleen size (cm)9 (0–24)0 (0–0)17<.001−9 (− 24, 0)6 (0–20)0 (0–3)23<.001−7 (− 20, 1).217
BM cell (%)55 (20–90)25 (5–80)9.04−25 (− 65, 10)45 (5–90)30 (10–100)17.03−10 (− 60, 20).316
BM lymph (%)56 (4–84)25.8 (12–78)9.26−10 (− 52.2, 18)48 (3–84)19 (7–74)18<.001−76 (− 62, 10).117

In patients who underwent splenectomy, the pattern of bone marrow infiltration by lymphocytes/lymphoma cells changed in 3 of 9 patients who underwent a bone marrow biopsy after splenectomy. One patient had a nodular interstitial pattern that changed into a nodular pattern; 1 patient had an interstitial diffuse pattern that changed into an interstitial nodular pattern, and 1 patient had a nodular pattern that changed into a nodular interstitial pattern.

Among the patients who received rituximab who had “paired” assessments for patterns of bone marrow infiltration by lymphocytes/lymphoma cells, 10 patients had no evidence of disease in the bone marrow after therapy. Eight of those 10 patients had available biopsies prior to rituximab therapy that demonstrated an interstitial (4 patients), nodular (2 patients), or interstitial diffuse (1 patient) pattern or lymphoid aggregates (1 patient).

Other Malignancies

Seventeen of 70 patients (24%) had another malignancy. Eight of those 17 patients (47%) had a malignancy before their diagnosis of splenic lymphoma. The distribution of cancers was as follows: basal cell or squamous cell skin carcinoma in 4 patients, breast cancer and a squamous cell carcinoma in 1 patient, vulvar squamous cell carcinoma in 1 patient, bladder cancer in 1 patient, and melanoma in 1 patient. Eight of 17 patients (47%) developed a secondary malignancy after their diagnosis of splenic lymphoma, including breast cancer in 3 patients, mesencephalic/thalamic large cell lymphoma in 1 patient, Hodgkin lymphoma in 1 patient, testicular cancer in 1 patient, prostate cancer in 1 patient, and basal cell carcinoma in 1 patient. One patient had rectal cancer prior to and also developed a secondary meningioma after the diagnosis of splenic lymphoma. There were 9 patients who developed a cancer after their diagnosis or treatment of SMZL, including 2 patients in the observation group, 3 patients in the chemotherapy group, 2 patients in the rituximab group, and 2 patients in the chemotherapy plus rituximab group.


The CR, FFS, and survival rates were found to be higher with rituximab alone and in combination with a purine analog-based therapy (FCR or R-FMD) than with chemotherapy alone. Our results appear to be in line with published data that have demonstrated similar survival rates in patients with SMZL.8–11, 13, 28, 29 However, the proportion of patients who received therapy was higher in the current study compared with others, suggesting that higher risk patients may have been referred to our department.

The high response rate with rituximab alone was encouraging, including its efficacy in patients who had adverse prognostic features, such as older age and high lactate dehydrogenase or β2-microglobulin levels. The current study data demonstrated that rituximab effectively controls SMZL, as evidenced by improvement in white blood cells, absolute lymphocyte counts, and splenomegaly, and suggested that rituximab therapy may be proposed as the treatment of choice, at least in older patients with SMZL who have comorbid diseases. The high response rate with rituximab may be associated with higher levels of surface CD20 molecules per cell on the circulating lymphocytes in patients with SMZL compared with the levels in patients with CLL.26

Three of 6 patients who received chemotherapy plus rituximab in the current study had a PR because of prolonged cytopenia, which is typical of myelosuppressive therapy. Although it is difficult to interpret these data because of the very small numbers of patients, prolonged FFS and survival were observed in the patients who received chemotherapy plus rituximab.

The results of the current study suggest that rituximab-based therapy may be more effective than chemotherapy alone. However, this cannot be a definitive conclusion given the retrospective nature of the analysis and the relatively small number of patients evaluated.

Earlier experiences with chemotherapy in patients with SMZL showed some benefit with alkylating agents, although survival rates were noted in 1 study among patients who initially were treated with chemotherapy compared with other treatments.9 Previous studies have shown that fludarabine induces responses even in the salvage setting, but the majority of patients eventually develop disease recurrence.14–16 Treatment with 2-chlorodeoxyadenosine was associated with poor outcomes and high toxicity in splenectomized patients with SLVL in 1 study,17 but it was associated with favorable outcomes in another study that used a different dose and schedule.18

Several factors have been described that predict clinical outcome in patients with SMZL.6, 8, 9, 28, 30–32 Cytogenetics, e.g., t(11;18), also reportedly is predictive of response in patients with low-grade gastric lymphoma of mucosa-associated lymphoid tissue.33 However, the only factors in the current study that predicted shorter survival among patients who required therapy were age >60 years and treatment with chemotherapy. The difference may be associated with the activity of rituximab-based therapies in this group of patients.

It is noteworthy that flow analysis identified relatively high rates of CD5 positivity in the current series, even though SMZL has been considered a CD5-negative lymphoma. This discrepancy may be explained by differences in the sensitivity of flow-cytometry and immunocytochemistry methods or by the acquisition of different cell phenotypes according to their microenvironment,5 which is compatible with the high proportion of patients with leukemic-phase disease noted in the current study.

In the current series, the incidence of other malignancies appeared to be higher than that in other lymphoproliferative disorders. However, approximately 50% of these cancers were diagnosed prior to diagnosis of or therapy for SMZL, which suggests a pathogenetic role for inactivation of tumor suppressor genes, such as p53. Indeed, loss or mutations of p53 have been associated with a poor outcome in patients with SMZL.34, 35

SMZLs are heterogeneous diseases and, thus, may respond variably to therapy. Both mutated and unmutated cases of SMZL have been reported.36 The heterogeneity of SMZL is considered to reflect the heterogeneity of their nonneoplastic counterparts.37 Some SMZLs are antigen driven, and differences in antigen influx (through blood or lymph) and antigenic nature (T-cell dependent or not) may enhance their heterogeneity, which largely depends on the anatomic site from which the lymphoma arises.37

The results of the current study combined with data demonstrating the distinct biology of SMZL suggest that carefully designed clinical trials with the objective of curing SMZL should be developed separately from trials for other lymphomas. To our knowledge, prospective studies of SMZL are nonexistent, and current therapies are based on case reports, retrospective analyses of relatively small numbers of patients, or the application of treatments that are effective in other lymphomas. Rituximab-based regimens appear to be superior to chemotherapy; therefore, these therapies should be explored further in patients with SMZL.