Benefit of rituximab maintenance is associated with Follicular Lymphoma International Prognostic Index in patients with follicular lymphoma

Rituximab maintenance (RM) prolongs the progression-free survival (PFS) of responding patients with follicular lymphoma (FL), but the maintenance efficacy in different Follicular Lymphoma International Prognostic Index (FLIPI) risk group is still confusing. We performed a retrospective analysis of the effect of RM treatments in patients with FL responding to induction therapy based on their FLIPI risk assessment carried out prior to treatment. We identified 93 patients between 2013 and 2019 who received RM every 3 months for ≥4 doses (RM group), and 60 patients who did not accept RM or received rituximab less than 4 doses (control group). After a median follow-up of 39 months, neither median overall survival (OS) nor PFS was reached for the entire population. The PFS was significantly prolonged in the RM group compared to the control group (median PFS NA vs 83.1 months, P = .00027). When the population was divided into the 3 FLIPI risk groups, the PFS differed significantly (4-year PFS rates, 97.5% vs 88.8% vs 72.3%, P = .01) according to group. There was no significant difference in PFS for FLIPI low-risk patients with RM compared to the control group (4-year PFS rates, 100% vs 93.8%, P = .23). However, the PFS of the RM group was significantly prolonged for FLIPI intermediate-risk (4-year PFS rates, 100% vs 70.3%, P = .00077) and high-risk patients (4-year PFS rates, 86.7% vs 57.1%, P = .023). These data suggest that standard RM significantly prolongs the PFS of patients assigned to intermediate- and high-risk FLIPI groups but not to low-risk FLIPI group, and pending larger-scale studies to validate.


INTRODUCTION
Follicular lymphoma (FL) is the one of common lymphomas in the western countries, 1,2 and accounts for 2.9% to 3.9% of all lymphoid neoplasms in China. 3,4 FL is an incurable, indolent disease characterized by repeated relapses. 5 With the advent of rituximab immunotherapy in 1997, significant improvements have been made in lymphoma and other cancer therapy over the past 2 decades, which has changed the natural course of the disease and its treatment regimens. Consequently, rituximab-based immunochemotherapy is now the standard choice for first-line therapy of FL, 6,7 and patients can typically expect response rate of nearly 90%, and 5-year overall survival (OS) rates approaching 90%. 8 Several studies have explored the outcomes of rituximab maintenance (RM) treatment for patients responding to induction. These studies have shown that RM prolongs the progression-free survival (PFS) of patients with untreated FL and relapsed FL [9][10][11] ; however, patients with FL have heterogeneous outcomes regardless of RM treatment. This might be due to differences in individual FL clinical characteristics. There were a few prognostic systems for patients with FL including the Follicular Lymphoma International Prognostic Index (FLIPI), FLIPI-2, m7-FLIPI, and PRIMA-PI. [12][13][14] FLIPI is the most commonly used system to stratify patients according to their clinical features and to predict survival. 15,16 How RM affects patients in different FLIPI risk groups remains unclear. Here, we aimed to evaluate the effectiveness of RM in patients belonging to the 3 FLIPI risk groups to see whether all patients in different FLIPI risks benefit from RM treatments. To do so, we performed a retrospective analysis of the effect of RM treatments in patients with FL responding to induction therapy based on the FLIPI risk assessment carried out prior to treatment.

Patients
This single-center retrospective study was conducted between January 2013 and December 2019 at the Institute of Hematology & Blood Diseases Hospital, Tianjin, China. A total of 173 adult patients with newly diagnosed FL completed induction therapy with rituximab-combination were identified (Fig. 1). All cases involved in this study were approved by the institutional ethics committees (IIT2020023-EC-1), and patient informed consent was obtained in accordance with the Declaration of Helsinki. Eligibility criteria included: (i) patients newly diagnosed with FL according to the criteria of the World Health Organization classification of tumors of hematopoietic and lymphoid tissues in 2013, 17 and (ii) histological grade 1-3A. Exclusion criteria were as follows: (i) histological grade 3B or histological transformation into diffuse large B cell lymphoma (DLBCL), (ii) incomplete FLIPI score.

Treatment regimens and response monitoring
The induction regimen was based on the physicians' choice. The most common regimen was rituximab, cyclophosphamide, vincristine, doxorubicin, and prednisone (R-CHOP). Other regimens include rituximab, cyclophosphamide, vincristine; doxorubicin, and dexamethasone (R-Hyper-CVAD), the rituximab, etoposide, doxorubicin, cyclophosphamide, vincristine, and prednisone (DA-EPOCH-R), the rituximab, fludarabine, cyclophosphamide, and mitoxantrone (R-FCM), and the bendamustine plus rituximab regimen (BR). Some patients combined rituximab and lenalidomide with chemotherapy such as R2-CHOP and R2-DA-EPOCH. Patients responding to induction therapy received rituximab treatments every 3 months as maintenance therapy, or observation based on the advice of physicians and the choice of patients.
The response was evaluated within 1 to 3 months after the end of the induction therapy through (18)F-fluorodeoxyglucose positron emission tomography (FDG-PET) or computed tomography (CT) plus bone marrow biopsy if the bone marrow was involved at diagnosis according to the 2014 Lugano Classification. 18 The response after induction therapy included complete response (CR), partial response (PR), stable disease (SD), progressive disease (PD). PFS was defined as the duration from the time of initiation of induction treatment to progression. OS was defined as the time of initiation of induction treatment to death.

Statistical analysis
Categorical variables were compared using Fisher exact tests. Continuous variables were compared by the Wilcoxon ranksum test. PFS and OS were estimated using the Kaplan-Meier method and comparisons between groups were made by the logrank test. Statistical analyses were conducted using R version 4.1.2 (https://www.r-project.org/).

Baseline characteristics
The clinical characteristics of the 173 patients are shown in Table 1. The median age at diagnosis was 47 years old (range 19-70) and 79 patients (45.7%) were male. The bone marrow was involved in most cases (116 patients, 67.1%) and most of the patients (155 patients, 90.1%) had advanced-stage disease (90.1%). The patients were assigned to 1 of 3 risk groups prior to receiving treatment according to FLIPI, 46 (26.6%) patients were in the low-risk group, 82 patients (47.4%) in the intermediate-risk group, and 45 (26.0%) patients in the high-risk group.  Fig. 3B).

Survival according to RM status
Of the 153 patients who responded to the induction therapy, 60 patients did not receive RM treatment or received fewer than 4 doses (control group). The other 93 patients received RM more than 4 times (RM group). The median number of rituximab treatments in the RM group was 7 (range 4-10). Most patients discontinued RM treatment due to financial reasons. The 2 groups were well matched in terms of baseline characteristics such as age, sex, Ann Arbor stage, lactate dehydrogenase (LDH) levels, and FLIPI score (Table 1).
PFS was significantly prolonged in the 93 patients in the RM group compared with the control group (P = .00027). Specifically, the median PFS was 83.1 months (95% CI, 68.0-NA) for the control group, and was not reached for the RM group (95% CI, NA-NA) (Fig. 4A). The 4-year PFS rate was 72.4% (95% CI, 59.2%-88.7%) for patients in the control group and 95.5% (95% CI, 90.6%-100%) for patients in the RM group (P = .0007). Despite this prolongation in the PFS, we saw no difference in OS between the 2 groups (Fig. 4B). The follow-up of 120 patients were longer than 24 months. Among them, there were 44 patients in the control group and 4 (9.1%) patients experienced progression within 24 months (POD24). And there were 76 patients in the RM group, 1 (1.3%) patient underwent POD24. There were apparently less patients experienced progression in the RM group (P = .025).

The impact of RM treatment on patients at different FLIPI risks
To determine the efficacy of RM treatment based on the pre-treatment FLIPI risk assessment, the PFS of patients in the RM and control groups after induction were analyzed according to their FLIPI score (Fig. 6)

DISCUSSION
FL is an indolent disease that tends to respond well to immunochemotherapy. However, patients with FL have heterogenous outcomes. To our knowledge, this is the first evaluation in a Chinese cohort of effects of RM treatment in newly diagnosed patients with FL who responded to induction therapy. This study showed the significance of a RM treatment strategy guided by a pre-treatment FLIPI assessment. The results indicated that patients with intermediate-and high-risk FLIPI scores benefited from RM treatment in terms of PFS while patients with low-risk FLIPI scores did not.
In our center, the ORR was 95.6% and CRR was 76.8% at end of induction therapy for all eligible patients with FL. Validating that our cohort is generally representative of responses in the rituximab era, the response rates in the current study are similar to those identified by Luminari et al in their FOLL12 study (ORR 96%, CRR 82%). 19 The induction regimens in the FOLL12 study included R-CHOP and BR regimens.
Patients with FL often experience recurrent relapses and unfortunately, the response duration and survival time shortens after each relapse. 20 Moreover, an early relapse is predictive of a high risk for death. 21 Patients who experienced progression within 24 months had poor outcomes. 22 Taking these factors  into consideration, treatment regimens must be designed that can prolong the duration of response and postpone or even prevent relapses, especially early relapse. Over the past 2 decades, we have seen that the addition of rituximab to chemotherapy has improved the response and survival rate of patients with FL as rituximab can remove residual FL tumor cells which may lead to relapse. 23,24 For this reason, rituximab is considered the first-choice maintenance therapy. In line with previous randomized controlled trial PRIMA 9,25 and a retrospective study with BR as induction regimen, 26 we found that RM treatment significantly prolonged the PFS for patients who responded to induction therapy, even though the baseline characteristics between patients at our center and PRIMA were different. Specifically, the patients at our center and eligible for our study were younger than those in PRIMA (median age, 47 vs 56 years), but this was consistent with clinical features of 1845 patients in the Chinese multicenter study. 27 In addition, more patients at our center presented with bone marrow involved (67.1% vs 55.0%) and elevated β2 microglobulin (43.5% vs 32.0%) than those in the PRIMA study. Patients at our center were representative of Chinese patients with FL except for higher BM involvement which may be due to the higher rate of usage of flow cytometer technology in bone marrow aspirate at our hospital. Despite these differences, we saw a significant reduction in the risk of progression for the patients with CR or PR, which was consistent with the results of the FOLL12 study. 19 Importantly, we found that RM significantly reduced the rate of POD24. But OS was not prolonged due to small sample. It is the first time that RM has been found to reduce the incidence of POD24. Validation is needed in the future.
In most studies, RM was scheduled every 8 weeks. However, the pharmacodynamics study of rituximab showed that CD20+ B cells in the peripheral blood remained depleted for at least 2 to 3 months. 28 Besides, RM every 2 months is more toxic than every 3 months. 29 Take these into account, we recommended RM every 3 months. Regarding the optimal duration of RM  treatments, our data confirmed that inclusion of RM on the schedule for ≥4 times effectively prolonged the 4-year PFS by up to 95.5%. Importantly, data from a previous study demonstrated that 5-year RM therapy does not improve event-free survival but increases toxicity, thus resulting in infections and lower life quality. 30 Based on these findings, we consider that the optimal duration of RM treatment is 1 to 2 years to achieve maximal efficacy and minimal toxicity; however, results of our cohort did not contain the adverse events during and after the RM treatments. The toxicity of RM should be taken into consideration in the follow-up in the future studies.
The FLIPI score was developed to risk-stratify patients and to predict survival in the pre-rituximab era. 15 According to our cohort and previous studies, 16 the FLIPI score remains a good predictor for PFS in the rituximab era (P = .0025). Our subgroup analysis showed no statistical benefit of RM treatment for patients in the FLIPI low-risk group. These findings differ from those of the PRIMA study which showed a PFS benefit of RM in all FLIPI risk groups. 25 And in FOLL12 study, the patients were assigned into 2 groups according to FLIPI score, those are low-intermediate risk and high risk. Patients in both groups benefited from standard RM treatment in terms of PFS. 19 However, patients in the FLIPI low-risk group included in this study achieved long PFS irrespective of RM treatment, with approximately 94% of patients in the FLIPI low-risk group still in remission after 4 years. As a common toxic effect of R is lymphocytopenia, which can lead to infection, we would not suggest patients in the low-risk group to undergo RM treatment based on our data. These patients will also benefit from reduced costs and duration of hospitalization by not undergoing, in our opinion, unnecessary RM treatment. Further clinical trials designed to evaluate RM treatments for patients in FLIPI low risk are needed to demonstrate the effects of RM. Patients in our center were much younger and could undertake intensified regimens, like R-EDOCH, and some patients received lenalidomide for leukemic phase, which could prolong the PFS and reduce the progression rate. Here, we saw that the PFS of patients at intermediate and high risks were significantly prolonged with RM treatment. We thus recommend that patients at intermediate and high risk undergo RM treatment. FLIPI-2 and PRIMA-PI were built in the rituximab era. They both worked well in stratifying the patients with PFS. However, β2MG test was not performed routinely at our center before 2014. As a result, FLIPI-2 and PRIMA-PI were not available for part of the patients. In addition, the diameter of nodal disease was missing for near one third of patients which is one of the FLIPI-2 factors. Consequently, FLIPI-2 and PRIMA-PI were not used to determine the effect of RM.
There were limitations. First of all, this study is a retrospective study, the inherent limitations include selection, confounding variables, recall biases, and nonuniform follow-up also affect the results. Thirty-seven patients were excluded from this study because of incomplete FLIPI component which might create a selection bias. As we concerned, the baseline and induction therapies were comparable to the patients enrolled in this study. Second, the induction therapies varied between patients which might influence the results. In real-world practice, the induction therapies were adjusted considering the physical and financial conditions in China.

CONCLUSION
In summary, our data suggest that to improve PFS and prevent relapse, RM be included on the schedule for a minimum of 1 year for patients with FL whose pre-treatment clinical and biochemical indices indicate that they fall into the FLIPI intermediate-and high-risk groups, and they respond to immunochemotherapy induction. With the duration of remission prolonged, these patients would have more opportunities to receive new target therapies in the future.