Splenic marginal zone lymphoma in Sweden 2000–2020: Increasing rituximab use and better survival in the elderly

Abstract The treatment of splenic marginal zone lymphoma is debated: splenectomy (the old standard‐of‐care) is better than chemotherapy but maybe not better than rituximab‐containing treatment. We examined all 358 patients diagnosed with splenic marginal zone lymphoma in Sweden 2000–2020. The median overall survival was 11.0 years. The median age was 73 years; 61% were women. Age was the only independently prognostic clinical characteristic. Eighty‐six patients were started on wait‐and‐watch, 90 rituximab monotherapy, 47 rituximab‐chemotherapy, 88 splenectomy, 37 chemotherapy, and 10 both systemic therapy and splenectomy. Overall survival was inferior in patients treated with chemotherapy, but equal in patients treated with rituximab, rituximab‐chemotherapy and splenectomy. Patients treated with both systemic therapy and splenectomy showed good outcome, suggesting that surgery can be safely reserved for nonresponders. After adjustment for age, survival did not differ between patients started on wait‐and‐watch and those treated with splenectomy or rituximab‐containing therapy. Over time, rituximab use and survival increased in patients ≥73 years. This is, to our knowledge, the largest population‐based study of splenic marginal zone lymphoma patients treated with upfront rituximab. We conclude that wait‐and‐watch remains the most reasonable option in asymptomatic splenic marginal zone lymphoma patients. Symptomatic patients should be offered single‐agent rituximab in first line.

of the patients never require therapy [3]. There is no consensus on first-line therapy of splenic marginal zone lymphoma. Modern first-line treatments include splenectomy [3][4][5][6], rituximab monotherapy [7][8][9][10] and rituximab combined with chemotherapy [9][10][11][12][13]. Splenectomy is preferable from a diagnostic perspective and gives a rapid relief of symptoms (i.e., abdominal discomfort and cytopaenias related to splenic sequestration) [13]. However, perioperative complications may occur (i.e., infections, bleeding and thrombosis), and patients with splenic marginal zone lymphoma are often elderly and at risk for surgical complications [14]. also considered a risk factor in COVID-19, now an endemic disease, and some countries have prioritized asplenic patients for vaccinations [15,16]. In the US, rituximab has become increasingly common, especially for frail patients, while splenectomy has become rarer [17]. The current Swedish guidelines for treating splenic marginal zone lymphoma recommend either splenectomy or rituximab as first-line therapy [18].
We wanted to investigate the outcome in splenic marginal zone lymphoma by first-line treatment modalities in a large, unselected and relatively recent patient population including many rituximab-treated patients.

Swedish registries
The Swedish lymphoma register was established in 2000. The Swedish lymphoma register records all new cases of lymphoma in Sweden in patients 18 years or older, with information on subtype, clinical and demographic factors, and first-line treatment (including the option wait-and-watch), as previously described [19]. For the purpose of this study, patient and lymphoma-specific data were extracted from the

Patients
We identified all patients in the Swedish lymphoma register diagnosed with splenic marginal zone lymphoma between 1st January 2000 and 31st December 2020. Patients who were reported to have, at diagnosis, other extranodal involvement than bone marrow, peripheral blood or liver were excluded, because those features are more congruent with extranodal marginal zone lymphoma (n = 11). Patients with nodal involvement outside the abdomen were considered nodal marginal zone lymphoma, and also excluded (n = 5). Furthermore, we a priori planned to remove any cases with evidence of primary transformation

Statistical analysis
Depending

Treatments
As first-line strategy, wait-and-watch was initiated in 86 patients.
Adverse World Health Organization (WHO) performance scores were more common in patients treated with rituximab-chemotherapy and in those who were started on wait-and-watch (on the other hand, the latter showed less B symptoms). There were no differences in age or WHO performance status between the sexes (data not shown).

Survival analysis
The estimated median survival was 11.0 years ( Figure 1A). High age, poor WHO performance status and low albumin levels were significantly associated with inferior survival in univariate analysis, but calendar period of diagnosis had no impact (Table 1) in patients 25-65 years of age ( Figure 1B). All variables showing an association to overall survival with p < 0.10 were included in multivariable analysis: age, WHO performance status, albumin levels, and the presence of B symptoms.
With respect to overall survival, chemotherapy without rituximab was significantly inferior to rituximab monotherapy and rituximabchemotherapy combinations ( Figure 1C), also in multivariable analysis (Table 3; multivariable p = 0⋅001). Median survival was 5⋅6 years in those treated with chemotherapy and 11⋅7 years in those treated with rituximab-containing therapy. Chemotherapy was also inferior to splenectomy (Table 3). There appeared to be equally good outcome in the 88 patients who underwent splenectomy alone as in the 10 who received systemic therapy and splenectomy (Figure 2A; Table 3; . In a subsequent analysis, all splenectomized patients were grouped as one category (n = 98) with an unadjusted median survival of 13⋅7 years: there were no survival differences between rituximab monotherapy, rituximab-chemotherapy, and any splenectomy, with adjusted hazard ratios (HRs) between 0⋅9 and 1⋅1 and p values > 0⋅6 for all comparisons (Table 3; Figure 2B). The results were identical when the 88 patients treated with splenectomy alone were compared with those given rituximab-containing treatment (adjusted HR 1⋅0 and p = 0⋅92; Table 3). Finally, patients who were started on wait-andwatch showed a tendency to inferior outcome compared with those who received rituximab-containing therapy or splenectomy, but only in univariate analysis ( Figure 2C); there was no difference in multivariable analysis adjusted for age (HR 1⋅0; Table 3). We analyzed treatments in different age orders and found that there were no survival differences between rituximab-containing treatment and splenectomy in any age bracket ( Figure 3A), nor were there significant survival differences between wait-and-watch and rituximab-containing treatment/splenectomy ( Figure 3B). Rather, Figure 3 shows that with modern therapeutic decisions, encompassing rituximab, and splenectomy, prognosis is mostly driven by age. Figure 3C shows that in patients <73 years, overall survival was similarly excellent after both (p = 0.80) rituximab monotherapy and upfront splenectomy (at 10 years 73%). Likewise, there was no overall survival difference between patients ≥73 years treated with rituximab monotherapy and upfront splenectomy (p = 0.62), although overall survival as expected was shorter for these (median, 8.9 years). Age remained highly prognostic in all multivariable analyses, but otherwise the only nontreatment factor that retained some independence in multivariable analysis was the presence of B symptoms, which was significant in the models of rituximab-containing therapy and chemotherapy.
In the entire population, there were no changes in survival over time, which is explained by the fact that younger patients exerted strong weight in long-term survival analysis of this indolent disease, and throughout the study period patients <73 years mostly received modern lymphoma treatment (splenectomy or rituximab-containing regimens): 73% between 2000-2010 and 72% 2011-2020 (Table 4) with an identical survival between epochs (p = 0⋅91). However  Note: One patient in the register was of unknown age at diagnosis. She was upfront splenectomized. This is why the sum of patients splenectomized aged 25-72 and 73-95 years is 97, not 98.
later period among the elderly (p = 0⋅038); their median overall survival improved from 5⋅2 to 8⋅0 years ( Figure 3D). The increase of modern treatment in the elderly was driven by more use of rituximabcontaining treatment (from 10% to 56%), while splenectomy became rarer (from 36% to 10%), see Table 4.

DISCUSSION
This study presents the largest cohort of splenic marginal zone lymphoma patients treated upfront with rituximab-containing therapy (n = 137), of whom 90 received rituximab as a single agent. The followup times are sufficient for meaningful interpretation (median, 7⋅8 years). Patients who were treated with chemotherapy alone showed inferior survival compared to those who received rituximab-containing therapy or splenectomy, agreeing with earlier smaller series [9,10].
However, we saw no survival differences between patients treated with rituximab-containing treatment or splenectomy. Some authors have shown better long-term outcome after rituximab [10,20], others after splenectomy [13]. Still, the advent of rituximab allowed for a good treatment option for all patients, including the elderly, in whom rituximab use and overall survival increased.
Wait-and-watch remains a good strategy for asymptomatic patients, since there appears to be at least equal survival after adjustment for age, which is the strongest prognostic factor in this disease. The study also shows that in an unselected patient population, the median age is well above 70 years and the female-to-male ratio almost 2:1, making this a disease of elderly women. The median survival time in this study is 11⋅0 years and the median age 73. This suggests that with modern management of this disease, at least in elderly patients, the lymphoma will probably seldom affect their expected survival. Our results agree with previous reports that show equal outcome between splenectomy and rituximab-containing therapy in patients >65 years [21]. Our study also suggests that wait-and-watch remains the most sensible approach to asymptomatic patients. Indeed, in younger patients, survival is excellent with a management using wait-and-watch ( Figure 3B).
In spite of the inherent bias in a retrospective analysis of different treatments, it appears that rituximab monotherapy, rituximabchemotherapy and splenectomy are equal first-line approaches with respect to overall survival. It is also noteworthy that patients who in first line underwent both splenectomy and systemic treatment (where the first given of these two was presumably suboptimal) had good prognosis, suggesting that one does not lose anything from starting with rituximab monotherapy (preferably preceded by pneumococcus vaccinations). In the small fraction of poor rituximab responders [8], salvage splenectomy appears to be a good second-line option.
Our study has limitations. We do not know these patients' hepatitis C status. However, hepatitis C-induced splenic marginal zone lymphoma would be exceedingly rare in Sweden, because of the low national prevalence of the virus; indeed, in the previous report where the Swedish lymphoma register was used there was not a single case of active hepatitis C in 289 investigated splenic marginal zone lymphoma patients [13]. Another weakness is that there is no information on relapses or second-line therapy, and thus no information on time to next treatment or progression-free survival. But those endpoints are less interesting when rituximab monotherapy is compared with rituximab-chemotherapy or with surgery.
We conclude that wait-and-watch is the most reasonable option in unsymptomatic splenic marginal zone lymphoma patients, and those who have symptoms should be offered single-agent rituximab as a standard first-line regimen. Although splenectomy shows the same long-term survival, the complications and lifelong infectious risks after the procedure suggest that it should be reserved for patients who do not respond to rituximab.

AUTHOR CONTRIBUTIONS
HRJ and BEW planned the study. HRJ and BEW analyzed data and wrote the manuscript. KS, OL, MH, MVS, MO, IT, AJ, and BEW contributed to the registry and data interpretation. All authors participated in the final version of the manuscript.

CONFLICT OF INTEREST STATEMENT
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

FUNDING INFORMATION
The funding source did not affect any part of the research.

DATA AVAILABILITY STATEMENT
None.