Peripheral blood involvement in extra‐nodal marginal zone lymphoma

To the Editor,

Marginal zone lymphomas (MZLs) are a group of indolent B‐cell lymphomas originating from the memory B lymphocytes of the marginal zone of secondary lymphoid follicles. MZLs are characterized by clinical, molecular and genetic heterogeneity which separates them into three subtypes: extra‐nodal (EMZL), nodal (NMZL) and splenic (SMZL).

EMZL is the most common subtype, comprising approximately 61% of cases. ¹ EMZL arises in mucosa‐associated extra‐nodal sites such as the stomach, ocular adnexa, lungs, thyroid and salivary glands and develops in an inflammatory environment of chronic infections or autoimmune conditions. ² Accurate staging of EMZL is important, since it guides therapeutic decisions. Stage I disease is potentially curable with radiation therapy (RT), whereas patients with advanced stages require systemic therapy or combined modalities. Distant relapses are observed in 8%–19% of patients with stage I EMZL treated with RT and are associated with a shorter median survival compared to the matched general population, raising the possibility that suboptimal staging or undiagnosed dissemination may explain these events. ¹ , ³ , ⁴ , ⁵ , ⁶ , ⁷

Circulating peripheral blood (PB) MZL cells are commonly detected in SMZL and monoclonal B‐cell lymphocytosis (MBL), non‐chronic lymphocytic leukemia (CLL) with MZL‐like immunophenotype (MBL, non‐CLL, MZL‐like). The latter is a premalignant condition of clonal B‐cell lymphocytosis of marginal zone origin, defined as a persistent (>6 months) lymphocytosis of >3.0 × 10⁹/L. ⁸ , ⁹ Long‐term follow‐up of 102 patients demonstrated that 85 (83.3%) remained stable but 17 (16.7%) developed lymphoma, mostly SMZL and only one EMZL. ⁹ Overall, PB involvement in EMZL is rarely reported and almost exclusively in patients with disseminated stage IV disease. ¹⁰

Herein, we aimed to characterize the incidence of PB involvement in EMZL, especially in patients with limited stage disease at diagnosis. We hypothesized that PB involvement in EMZL could occur more frequently than previously reported, including in early‐stage patients.

We retrospectively reviewed the medical records of patients with MZL that were diagnosed and treated at the University of Miami Health System between the period of 2015 and 2025. The diagnoses of MZL in all the cases were retrospectively confirmed by experts in haematopathology and specifically in diagnoses of lymphoma. The diagnoses of lymphoplasmacytic lymphoma were ruled out by compilation of clinical presentation, laboratory findings, analysis of MYD88 mutations and pathological findings. PB flow cytometry (PBFC) was performed at the time of initial diagnosis in all the MZL patients seen in clinic as described in Supporting Information S1. This retrospective study was approved by the institutional review board which follows the tenets of the Declaration of Helsinki.

Collected data included demographic information and clinical and pathological characteristics. Staging of MZL was performed by imaging studies including computer tomography scans, magnetic resonance imaging, positron emission tomography, bone marrow (BM) biopsy and aspiration and specific studies as indicated by clinical presentation. In this study, we followed Lugano Classification staging recommendations which do not categorize blood involvement as stage IV disease. In clinical practice, many lymphoma physicians consider the presence of lymphoma cells in the blood as indicative of stage IV disease. However, in our patients, MZL cells were primarily detected by PBFC without clear morphological evidence on peripheral smears. Therefore, these cases were defined as stage I if BM involvement was absent. Fisher's exact test was used to calculate the p‐values of all variables except age and cell count comparisons that were evaluated by Wilcoxon rank‐sum test.

A total of 183 MZL patients with available results of PBFC at diagnosis were identified, including 127 patients (69%) with EMZL, 38 (21%) with SMZL and 18 (10%) with NMZL. PBFC demonstrated the presence of surface light chain restricted CD5⁻, CD10⁻ B cells with similar immunophenotype to cells detected in diagnostic biopsies in 74% (n = 28) of SMZL, 33% (n = 6) of NMZL and notably 7% (n = 9) of EMZL patients (Figure 1A). The average number of circulating MZL cells was significantly lower in EMZL 0.33 × 10⁹/L (range: 0.01–1.30 × 10⁹/L) compared to SMZL 14.1 × 10⁹/L (0.02–210 × 10⁹/L) (p = 0.000388), whereas no significant difference was observed in comparison to NMZL 0.74 × 10⁹/L (0.32–1.14 × 10⁹/L) (p = 0.105) (Figure 1B).

FIGURE 1.

Flow cytometry in marginal zone lymphoma (MZL). (A) Flow cytometry positivity of the peripheral blood by MZL subtype. (B) Cell counts by MZL subtype.

To further identify the factors associated with PBFC positivity in EMZL, we assessed the baseline demographic and clinical characteristics in the EMZL cohort and compared the patients with positive versus negative PBFC (Table 1). Patients with EMZL had a median age of 63 years (range 19–93), with a slight excess of female gender (58%), Hispanic ethnicity (52%) and White race (90%), most presenting with normal lactate dehydrogenase (80%). These variables did not statistically differ between patients with positive versus negative PBFC. Notably, there was a trend of higher prevalence of monoclonal gammopathy in patients with positive PBFC (44%) versus negative PBFC (18%) (p = 0.076). Monoclonal IgM paraprotein detected in small quantities was more prevalent in the positive PBFC subgroup (33% vs. 8%, p = 0.051). IgM levels were below 0.5 g/dL in 10 patients, and only one patient with positive PBFC had IgM levels of 1.79 g/dL. MYD88 L265P mutation was assessed in three of these patients and was negative. All patients with paraprotein had classical EMZL clinical presentations. Additionally, histological BM involvement was significantly more common in EMZL patients with positive (67%) versus negative (7%) PBFC (p = 0.00086). Concordantly, EMZL patients with positive PBFC presented more frequently with advanced Stage III–IV disease (78% vs. 21%, p = 0.0009). Conversely, stage I disease was significantly less common in patients with positive PBFC (22%) compared to the negative PBFC group (69%) (p = 0.007). Overall, two of 84 patients (2.4%) with clinical stage I EMZL had a positive PBFC. Both patients had EMZL of the lung. Classical clinical EMZL presentation without BM involvement in the presence of positive PBFC in a patient with stage I presentations suggests the presence of lymphoma cells in PB and cannot be attributed to MBL, non‐CLL, MZL‐like. Lung involvement was more frequently observed in patients with positive PBFC (44%) than those with negative PBFC (7%) (p = 0.0046). Of the seven patients with EMZL and positive PBFC who underwent treatment (2 stage I, 1 stage III and 4 stage IV), no progression was noted during a median follow‐up of 14.8 months (range 2–36.3).

TABLE 1.

Clinical characteristics of extra‐nodal marginal zone lymphomas patients.

Group	Positive FC (n = 9) (%)	Negative FC (n = 118) (%)	p‐value
Age, median, year	66	61	0.5075
Female gender	5 (56)	69 (58)	1
Hispanic	3 (33)	63 (53)	0.310
Non‐Hispanic	6 (67)	55 (47)	0.310
White race	8 (89)	106 (90)	1
Autoimmune disease	1 (11)	23 (19)	1
Cancer history	3 (33)	28 (24)	0.687
Monoclonal gammopathy	4/9 (44)	17/96 (18)	0.076
IgM	3/9 (33)	8/96 (8)	0.0512
IgG	1/9 (11)	10/96 (10)	1
Elevated LDH (%)	1/8 (13)	24/114 (21)	1
Anaemia, Hgb <12 (%)	1/9 (11)	8/116 (7)	0.5014
ECOG 0	9 (100)	111 (94)	1
Bone marrow involvement (%)	4/6 (67)	8/112 (7)	0.00086
Multiple extra‐nodal sites ≥2	1 (11)	20 (17)	1
Stage I	2 (22)	82 (69)	0.0071
Stage IV	6 (67)	25 (21)	0.0067
Site involvement
Lung	4 (44)	8 (7)	0.00461
Breast	2 (22)	5 (4)	0.0781
Orbital	1 (11)	39 (33)	0.2708
Stomach	1 (11)	29 (25)	0.6847
Skin	1 (11)	6 (5)	0.41

Note: Bold indicates statistically significant difefrenece.

Abbreviation: ECOG, Eastern Cooperative Oncology Group.

Here, we present the largest single‐centre study assessing the status of PBFC involvement in EMZL. Our analysis found that positive PBFC in EMZL is associated with advanced stage, pulmonary EMZL, BM involvement and a trend for monoclonal gammopathy with IgM paraprotein. None of the patients had splenomegaly. We also show that the number of circulating MZL cells was lower in EMZL than in SMZL. Finally, we report that localized EMZL can present with positive PBFC at the time of diagnosis, and this should be further assessed in future studies.

PB involvement by circulating lymphoma cells is well documented in indolent B‐cell lymphomas, including MZL, and varies according to the histological subtype, with SMZL being the most frequently involved. ¹¹ The novel finding of localized EMZL with positive PBFC warrants further investigation and raises the question of whether the positive PBFC should influence the disease staging and the treatment decisions. According to the current National Comprehensive Cancer Network guidelines, RT is recommended for early‐stage diseases, whereas more advanced diseases require systemic therapy when certain criteria are met. ¹² In the largest analysis to date evaluating prognostic factors and treatment‐related outcomes in stage I EMZL, we previously used data from the Surveillance, Epidemiology and End Results database and demonstrated that RT resulted in survival outcomes comparable to those of the general US population. ³ Nevertheless, the risk of distant lymphoma relapse persists, albeit it varies by primary location.

Our group previously has shown that patients with primary ocular adnexal mucosa‐associated lymphoid tissue (MALT) lymphomas treated with RT face a continuous risk of distant relapses including the central nervous system (3%) and also transformation to aggressive lymphoma (4%), with an estimated cumulative relapse (local and distant) or progression of 5.1% at 1 year and 31% at 10 years. ¹³ Similar results were recapitulated by a large cohort of 244 patients treated with curative intent RT for early stage EMZL, where 24% of patients developed relapses, mostly in distant sites, unrelated to the RT field. ¹⁴ In contrast, long‐term outcome data of patients with gastric MALT treated with RT, have demonstrated 88% freedom from treatment failure leading to a cure for most of the patients. ⁷ These studies raise the possibility of occult dissemination or involvement of PB at diagnosis that may lead to later distant relapses and vary by location. Indeed, herein we show that PB lymphoma cells may be detected at diagnosis in EMZL, with potentially increased incidence in localized EMZL of the lung.

Future studies should explore whether late relapses occurring in limited EMZL after RT are associated with positive PBFC at the time of the initial treatment. Until these data are available, in our practice, we treat patients with early‐stage EMZL with positive PBFC with systemic therapies because of the identified source for later relapses and unnecessary exposure to late RT side effects or elect low‐dose radiotherapy (2Gy × 2) for local disease control. Limitations of our study include the single‐centre experience and the small number of patients with EMZL and positive PBFC along with the short follow‐up time. Nevertheless, our findings challenge the traditional understanding that PB involvement is rare in EMZL, the recently proposed definition of disseminated MZL, ¹⁵ and highlight the importance of including PBFC in the staging of patients presenting with localized disease.

AUTHOR CONTRIBUTIONS

Izidore S. Lossos designed and performed the research, analysed the data, was involved in data acquisition and wrote the manuscript with input and approval of the final version from all co‐authors. Juliette L. Martin performed the research, analysed the data and was involved in data acquisition. Jennifer R. Chapman performed the research and wrote the manuscript. Georgios N. Pongas designed and performed the research, analysed the data, was involved in data acquisition and wrote the manuscript with input and approval of the final version from all co‐authors. All authors approved this manuscript for publication.

FUNDING INFORMATION

ISL is supported by the National Cancer Institute grant 1RO1CA233945‐01, Institute for Follicular Lymphoma Innovation, Dwoskin and Anthony Rizzo Families Foundations and University of Miami Sylvester Comprehensive Cancer Center (SCCC). This work was supported by the National Institutes of Health, National Cancer Institute grant NIH U01 CA195568. GNP research is supported by the philanthropy of Stephen Schwartz and the University of Miami Sylvester Comprehensive Cancer Center. This work was funded by the Sylvester Comprehensive Cancer Center National Cancer Institute core grant P30CA240139.

CONFLICT OF INTEREST STATEMENT

Izidore S. Lossos reports consultancy with SOBI, and research support from ADC Therapeutics and Genentech. Juliette L. Martin and Jennifer R. Chapman declares no conflict of interest. Georgios N. Pongas is a shareholder of Crispr Therapeutics and an equity holder of Mevox Ltd. and reports consultancy with ADC Therapeutics and Beigene.

ETHICS STATEMENT

This study was approved by the institutional review board (IRB) at UM/SCCC and data collection was conducted in concordance with all institutional guidelines and procedures.

Supporting information

Data S1.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.