Mantle Cell Lymphoma of Mucosa-Associated Lymphoid Tissue: A European Mantle Cell Lymphoma Network Study.

While classical nodal mantle cell lymphoma (cMCL) is often associated with involvement of multiple extranodal sites, isolated extranodal disease (ED) at the time of diagnosis is a rare event; data on the outcome of these forms are lacking. On behalf of the European MCL Network, we conducted a retrospective analysis on the clinical characteristics and outcomes of MCL presenting with isolated or predominant ED (MALT MCL). We collected data on 127 patients with MALT MCL diagnosed from 1998 to 2015: 78 patients (61%) were male with a median age of 65 years. The involved sites include: upper airways + Waldeyer ring (40; 32%), gastrointestinal tract (32; 25%), ocular adnexa (17; 13%), oral cavity and salivary glands (17; 13%) and others (13; 1%); 7 patients showed multiple extranodal sites. The median follow-up was 80 months (range: 6-182), 5-year progression-free survival (PFS) was 45% (95% CI: 35-54) and 5-year overall survival (OS) was 71% (95% CI: 62-79). In an explorative setting, we compared MALT MCL with a group of 128 cMCL patients: MALT MCL patients showed a significantly longer PFS and OS compared with nodal cMCL; with a median PFS of 4.5 years vs 2.8 years (p = 0.001) and median OS of 9.8 years vs 6.9 years (p = 0.018), respectively. Patients with MALT MCL at diagnosis showed a more favorable prognosis and indolent course than classical nodal type. This clinical variant of MCL should be acknowledged to avoid possible over-treatment.

Introduction

In addition to classical nodal mantle cell lymphoma (cMCL), an aggressive disease requiring high-intensity chemotherapy,[1,2] the 2016 update of the WHO classification of lymphoid neoplasms recognizes the less common leukemic non-nodal variant (nnMCL), characterized by lymphocytosis and splenomegaly without nodal disease, showing an indolent clinical course.[3-5]

In particular, the leukemic nnMCL is represented by cells that have experienced follicular germinal center and carry IgVH somatic hypermutation with a discriminant gene-expression profiling;[7-8] this variant is frequently associated with a >7 years survival.[1] Correct identification of this variant has a potential clinical impact, since this subset of MCL can benefit from a watch-and-wait or a more conservative approach.[6]

Extranodal involvement in classical nodal MCL is common, and gastrointestinal tract is the most involved site, often found endoscopically in asymptomatic patients or as multiple lymphomatoid polyposis.[9,10] Infiltration of breast, lung, skin, soft tissue, salivary gland and orbit are also seen. Involvement of more than 2 extranodal sites occurs in 30% to 50% of patients. However, isolated extranodal disease (ED) is rarely detected in cMCL at diagnosis and only a few reports are available: therefore, in everyday clinical practice it is not possible to define the optimal treatment strategy.[11,12]

On behalf of the European MCL Network, we conducted a multicenter study, which collected MCL cases characterized by isolated or predominant extranodal disease (which we defined MALT MCL).

Patients and methods

Study design

This was a retrospective multicenter study conducted on behalf of the European MCL Network. Consecutive MCL subjects characterized by isolated or predominant ED were enrolled, from 1998 to 2015, by all participating centers. We captured baseline clinical, laboratory and pathology data, initial therapy, and active follow-up of all patients for relapse/progression and death in order to describe a possible variant of MCL with a peculiar clinical presentation in MALT sites.

Afterward, we compared MALT MCL with a group of classical nodal MCL patients consecutively diagnosed in the same period in 2 major Italian centers in order to describe, in an explorative intention, the outcomes and the distribution of prognostic factors between the 2 groups.

The study was approved by the Ethics Committee/institutional Review Boards/data protection agencies of participating sites. Patients provided written consent to participate when appropriate.

Selection criteria

All patients in the MALT MCL cohort were aged >18 years and presented at diagnosis with only extranodal site (ES) involvement or predominant extranodal disease with minimal locoregional lymphadenopathy defined as CT scan longest axis < 2 cm. nnMCL characterized by splenomegaly and lymphocytosis without nodal involvement were excluded. Patients with involvement of tonsils and Waldeyer ring sites were also included in the extranodal group. Diagnosis of MCL was established according to the WHO classification criteria and confirmed by immunohistochemistry for cyclin D1 detection and/or FISH for translocation (11;14).

Staging procedures were not standardized but varied depending on different centers, they included chest and abdomen imaging investigations (computed tomography [CT]) and bone marrow biopsy in all patients; ultrasound [US] scans, digestive tract endoscopic investigations and ear, nose and throat (ENT) evaluation in selected cases.

In the comparative cohort, we included patients with classical nodal involvement with or without concomitant ED.

Statistical methods

Study endpoints included progression-free survival (PFS), early and late progression of disease (POD) and overall survival (OS), duration of response, time to next treatment and OS from relapse/progression (OS-2) [as defined in Supplementary material, Supplemental Digital Content].

Data were summarized by descriptive statistics. Differences between groups were evaluated by the Chi-square test or t test (Fisher exact test and Wilcoxon test, when appropriate). Survival (OS and PFS) was estimated using the Kaplan-Meier method and differences between groups were evaluated by the log-rank test (or the trend test when appropriate).

Both in the univariable and multivariable analyses, the effect of clinical or demographical factors on survival were evaluated using the Cox proportional hazards model. In order to identify variables impacting OS and PFS in MCL, the following parameters were evaluated: gender, simplified prognostic index for advanced stage MCL (sMIPI),[13] Ki-67 index,[14] cytology,[15] Ann Arbor stage, bone marrow involvement, leukemic disease, splenomegaly, type of treatment (ASCT) and clinical variant of disease (MALT vs cMCL). ASCT was considered as a time-dependent variable.

A p-value <0.05 was considered statistically significant (2 sides). All statistics were performed using SAS version 9.4.

Results

Patients characteristics and treatment of MALT MCL

We collected data from 127 patients with MALT MCL in 14 European centers from 1998 to 2015.

Median age at diagnosis was 65 years (range: 36–85), patients were predominantly male (61%) with a good performance status (Eastern Cooperative Oncology Group 0 in 94 patients, 74%) and localized disease (Ann Arbor stage I/II in 73 patients, 57%) (Table 1). Twenty-seven patients showed a bone marrow involvement in addition to ES, and 2 patients had also leukemic disease. Histological and molecular features of MALT MCL patients are summarized in Table S1 (Supplemental Digital Content).

Table 1

Features of 127 patients with MALT mantle cell lymphoma and 128 patients with classical nodal mantle cell lymphoma

Seventy-four patients (58%) with MALT MCL had minimal locoregional lymphadenopathy at diagnosis. Involved ES included: upper airways and Waldeyer ring (40 patients, 31%), gastrointestinal tract (32 patients, 25%), ocular adnexa (17 patients, 13%) oral cavity and salivary glands (17 patients, 13%). Other less frequent sites were skin (n = 3), thyroid (n = 3), breast (n = 1), liver (n = 1), testicle (n = 1), bone (n = 1), paranasal sinus (n = 1), kidney (n = 1), larynx (n = 1) and in 7 patients multiple MALT sites were involved at diagnosis.

The majority of patients fell into the low and intermediate sMIPI categories (84%).

Most patients (94%) treated with chemotherapy regimen received concomitant rituximab; 41 patients received induction therapy containing high-dose cytarabine followed by ASCT in 30 cases (24%). Other induction therapies included R-CHOP-like (n = 44) and R-bendamustine (n = 6). Data on therapy were not available for three patients. A minority received rituximab single agent (n = 4) or radiotherapy (n = 27), and a watch-and-wait policy was chosen in only 2 patients.

Twenty-five patients (25/27; 92%) treated with radiotherapy alone had a limited stage of disease (stage 1). They experienced 5-year PFS and OS of 53% and 70%, respectively.

Patients who underwent ASCT were mainly in advanced stage (23/30; 77%) compared with those treated with more conservative regimens (31/97; 32%). Only 2 patients underwent allogenic stem cell transplantation after a relapse to ASCT.

Outcome and prognosis of MALT MCL

Overall response rate to the primary treatment was 97%, and median duration of response and time to next treatment were 31.1 months (range: 1.3–169.4) and 22.4 months (range: 0.6–85.5), respectively.

We observed 44 deaths (median follow-up 80 months; 6–182), of which 34% were due to disease progression. Other causes of death included: other neoplasms (8 patients, 28%), treatment-related toxicity (8 pts, 28%) and no known cause (13 patients, 45%).

Thirty-three patients relapsed, 82% of patients had isolated extranodal involvement and only four patients (12%) had systemic nodal disease. A total of 19% of patients (6/31) who received rituximab single-agent or RT alone relapsed.

During follow-up, 16 patients developed secondary malignancies (3 acute myeloid leukemia, 1 myelodysplastic syndrome and 12 solid neoplasia) corresponding to a cumulative incidence of 8.7% at 5 years. Median time from diagnosis to the second neoplasm was 17.6 months (6–106).

PFS and OS at 5 years was 45% (95% CI: 35–54) and 71% (95% CI: 62–79), respectively.

In the univariable analysis (Table 2), age >65 years, high sMIPI, ECOG ≥1 and bone marrow involvement showed a negative impact on PFS. In the multivariable model, only sMIPI confirmed its statistically significant effect. Factors affecting OS in univariable analysis (Table 2) were age >65 years, high sMIPI, ECOG ≥1 and Ki-67 ≥30%. In the multivariable analysis, sMIPI and Ki-67 confirmed their statistically significant impact on OS. The effect of transplant on PFS (hazard ratio [HR]: 0.59, 95% CI: 0.29–1.19; p = 0.139) and OS (HR: 0.58, 95% CI: 0.24–1.38; p = 0.215) was not significant even when adjusted for the other factors.

Table 2

Univariable analysis for progression-free survival and overall survival in 127 patients with MALT mantle cell lymphoma

In MALT MCL, we did not observe a different outcome according to cytology (5-year PFS of 46% vs 44% for blastoid and classic cytology, respectively) and type of ES.

Early POD occurred in 18/43 (42%) patients. OS-2 for patients with early and late POD was 35.4% and 69.1% at 5 years, respectively (p = 0.002). Patients with late POD had an extremely favorable prognosis (median OS-2 not reached).

Comparison with nodal MCL

MALT MCL patients were compared to 128 patients with cMCL (Table 1). Median age was 65 years in both cohorts. We observed a slightly higher female prevalence in the MALT MCL cohort (39% vs 26%; p = 0.029). The unfavorable features as high sMIPI, Ki-67 ≥30% and blastoid/pleomorphic cytology showed a homogenous distribution in the 2 groups. Patients with MALT MCL had lower LDH values (p = 0.002) at diagnosis and a more often limited disease (stage I/II) than patients with cMCL (57% vs 9%, respectively; p < 0.001).

Regarding secondary ES in cMCL, 57 patients (44%) presented with at least one extranodal involvement at diagnosis in addition to bone marrow (gastrointestinal tract in 21 patients, Waldeyer's ring in 12 patients and multiple ES in 14 patients).

Regarding treatment, cMCL were treated with more intensive regimens compared with the MALT type; 41% and 24% of patients underwent ASCT (p = 0.002), respectively. No patient in the nodal cMCL group was treated with radiotherapy alone.

Median follow-up was 85.4 months (range: 6–218) and was not different from that of MALT MCL. The incidence of secondary malignancies was similar between the 2 groups.

MALT MCL patients showed a longer OS and PFS compared with cMCL: median OS 9.8 years vs 6.9 years (5-year OS: 71% vs 63%; HR: 0.63, 95% CI: 0.43–0.92; p = 0.017) and median PFS 4.5 years vs 2.8 years, respectively (5-year PFS: 45% vs 28%; HR: 0.54, 95% CI: 0.40–0.75; p < 0.001) (Fig. 1). In addition, when comparing only MALT patients with Waldeyer ring involvement to cMCL we observed that PFS was longer for MALT MCL than nodal type (p = 0.01).

Figure 1

Comparison of the overall survival of 127 patients with MALT mantle cell lymphoma and of 128 patients with classical nodal mantle cell lymphoma (A) and comparison of the progression-free survival of 127 patients with MALT mantle cell lymphoma and of 128 patients with nodal classic mantle cell lymphoma (B).

Univariable analysis for PFS and OS of the cMCL cohort is summarized in Table S2 (Supplemental Digital Content).

When restricting the analysis to patients with unfavorable cytology (blastoid/pleomorphic variant), PFS was significantly longer in MALT MCL compared with cMCL (Fig. S1, Supplemental Digital Content).

To evaluate the prognostic impact of proposed novel clinical variant of the disease, a model was constructed in the whole population considering the clinical variant as a variable (MALT vs cMCL). In the whole MCL population, in the univariable analysis, factors affected PFS were age >65 years, stage III/IV, bone marrow involvement, ECOG ≥1, Ki-67 ≥30%, blastoid/pleomorphic cytology and no ASCT. In the univariable model for OS, all investigating factors for PFS remained significant except stage III/IV (Table S3, Supplemental Digital Content). Clinical disease variant resulted as an independent prognostic factor for PFS (HR: 0.57; 95% CI: 0.37–0.86; p = 0.007) adjusted for all factors which confirmed their effect (Table 3).

Table 3

Multivariable analysis for progression-free survival and overall survival in the whole mantle cell lymphoma population (127 patients with MALT mantle cell lymphoma and in 128 patients with nodal classic mantle cell lymphoma)

The introduction of Ki-67 proliferation index into the multivariable model reduced the number of observations (from 241 to 165) and events (from 151 to 105 for PFS and from 99 to 67 for OS) due to the large number of missing data. Since Ki-67 constitutes a known strong prognostic factor that cannot be excluded from the analysis, we built the model comprising it (Table 3).

In the multivariable model for OS without Ki-67 index, MALT presentation was a factor with a prognostic impact (HR: 0.65; 95% CI: 0.43–0.98; p = 0.042) adjusted for MIPI and transplant status.

Discussion

Two major clinical and biological variants of MCL are described: in addition to the most common classical nodal MCL, a non-nodal variant characterized by splenomegaly and leukemic involvement with an indolent clinical course is widely described.[16]

In this retrospective series, we described a peculiar presentation of MCL characterized by ED in the absence or with minimal nodal involvement, which we can define as MALT-oma like MCL due to its similarity to MALT lymphomas.

The starting hypothesis underlying the project comes from the everyday clinical observation that patients with extranodal MCL seemed to present a more indolent behavior compared with nodal type. Therefore, we settled with a study within the European Mantle Cell network centers to confirm this clinical observation. The main objective of the study is purely descriptive of this clinical variant; the hypothesis of a difference in terms of outcome between the 2 groups (extranodal and nodal) has an exploratory intention only.

The frequent involvement of gastrointestinal tract in classic nodal MCL is well-known. In one study, 26% of patients with MCL presented with gastrointestinal symptoms at diagnosis; however, MCL infiltration was present histologically in the lower gastrointestinal tract in 88% and in the upper gastrointestinal tract in 43% of patients.[9] In our series of MALT MCL, 12 patients presented with upper gastrointestinal tract involvement while 13 had involvement of the lower tract and seven patients showed a simultaneous involvement of upper and lower gastrointestinal tract.

In our comparative cMCL series, the main secondary ES were GI and Waldeyer ring while typical MALT sites, such as salivary glands, ocular adnexa, and skin, were more frequently involved in MALT MCL than in classical nodal cases.

We decided to include isolated Waldeyer ring cases, although Waldeyer ring was listed as a nodal site in recent recommendations.[17] However, in most studies the Waldeyer ring is historically considered as an ES.[18] Interestingly, in our cohort, there was no difference in outcome within MALT MCL cohort according to the type of ES; on the other hand, the difference was still significant between the cMCL and MCL of the Waldeyer ring.

The majority of MALT MCL cases presented with minimal locoregional lymphadenopathy, but as for MALT marginal zone lymphomas,[19] the primary site of lymphoma involvement was defined as the clinically dominant ES, which requires diagnostic approach and to which primary treatment must often be directed. On the other hand, no patients showed distant adenopathy and during the course of disease, 82% of relapses were in ES while only four patients relapsed with systemic nodal disease. Noteworthy, the presence of locoregional lymphoadenopathy did not affect the prognosis.

We also want to underline that patients in the MALT MCL group with advanced stage (54 patients) or leukemic phase (2 patients) were included in the analysis because they showed a predominantly extranodal disease in addition to bone marrow involvement or leukemic disease. Majority of cases were stage IV for multiple extranodal sites or diffuse extranodal involvement.

In addition, the clinical features at presentation seem to be different in MALT MCL: female prevalence is higher in respect to cMCL, LDH is less frequently elevated and limited disease at initial staging is more common.

Patients with MALT MCL showed prolonged PFS and OS compared with the classic nodal variant. In addition, the extranodal variant of MCL was found to be an independent prognostic factor for PFS, defining a better outcome. In a multivariable model for OS, MALT MCL variant was associated with a reduction of the risk but failed to reach significance; in the model without Ki-67 index, MALT localization was a factor with a prognostic impact adjusted for MIPI and transplant status.

Regarding the impact of transplant on outcome, in MALT MCL we did not observe a statistically significant effect of transplantation, differing from nodal MCL. However, this should not be interpreted as a different impact of ASCT in the 2 groups since as the effect of ASCT is protective in both; the absence of significance is probably due a smaller sample size, and a larger series would be needed to settle this issue. Nevertheless, MALT patients, although undergoing ASCT less frequently, still had a better prognosis and could not benefit from a more intensive therapy.

Strengths of our study include the systematic collection and analysis of a large series of patients with a rare MCL clinical variant treated in the rituximab era, the homogeneous distribution of the baseline characteristics at diagnosis (cytology, Ki-67 and MIPI), the consecutive collection during the same time period as the controls and the comparable follow-up. Limitations include the possible lack of some clinical prognostic data, not standardized staging procedures between the centers and the absence of a centralized pathological review to identify biological markers able to recognize this clinical entity. In particular, the impact on outcome of cytology could not be better investigated due to the low prevalence of blastoid cytology in MALT and nodal MCL (14 and 18 patients, respectively). In addition, we do not have data in this setting on activity of new drugs, such as ibrutinib[20] and lenalidomide.[21]

In conclusion, this study identified a novel clinical variant of MCL, predominantly extranodal, with a more favorable prognosis and a more indolent course than the classical nodal type. The possible diagnosis of this peculiar presentation of MCL in MALT sites should be recognized, a possible overtreatment with intensive approaches may be omitted and an initial watch and wait strategy could be chosen in asymptomatic patients. Additional research is needed in this form of MCL in order to attribute its cellular origin and evaluate any molecular analogies with classic MCL, nnMCL and with marginal zone lymphomas.