- Email: [email protected]
Rituximab maintenance improves overall survival of patients with follicular lymphoma—Individual patient data meta-analysis
European Journal of Cancer xx (2017) 1e10
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Original Research
Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis Liat Vidal a,b,*, Anat Gafter-Gvili a,c, Gilles Salles d, Sami Bousseta e, Bernice Oberman f, Carmit Rubin f, Marinus H.J. van Oers g, Catherine Fortpied h, Michele Ghielmini i,j, Ruth Pettengell k, Mathias Witzens-Harig l, Peter Dreger m, Umberto Vitolo n, Maria Gomes da Silva o, Andrea Evangelista p, Hailun Li q, Laurence Freedman f, Thomas M. Habermann r, Ofer Shpilberg s a
Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel Institute of Hematology, Davidoff Cancer Center, Rabin Medical Center, Petah Tikva, Israel c Internal Medicine A, Rabin Medical Center, Petah Tikva, Israel d Hospices Civils de Lyon, Centre Hospitalier Lyon-Sud, Pierre-Benite, Universite´ Claude Bernard Lyon-1, Lyon, France e Biostatistics Department, LYSARC, Pierre-Benite, France f Sheba Medical Center, Gertner Institute for Epidemiology and Health Policy Research, Tel Hashomer, Israel g EORTC Lymphoma Group/HOVON, Academic Medical Center, Amsterdam, The Netherlands h EORTC, Brussels, Belgium i Oncology Institute of Southern Switzerland, Bellinzona, Switzerland j Swiss Group for Clinical Cancer Research, Bern, Switzerland k Department of Haematology, St. George’s University of London, London, UK l Department of Internal Medicine V, University of Heidelberg, Heidelberg, Germany m EBMT Lymphoma Working Party, Paris, France n Citta` della Salute e della Scienza Hospital and University, on behalf of FIL, Turin, Italy o CEDOC, Instituto Portugueˆs de Oncologia de Lisboa Francisco Gentil EPE Rua Prof. Lima Basto, 1099-023 Lisboa, Portugal p University-Hospital Citta` della Salute e della Scienza, Torino, Italy q Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute, Boston, MS, USA r Division of Hematology, Mayo Clinic, Rochester, MN, USA s Assuta Medical Center, Tel Aviv, Israel b
Received 25 September 2016; received in revised form 15 January 2017; accepted 24 January 2017 Available online - - -
* Corresponding author: Institute of Hematology, Davidoff Cancer Center, Rabin Medical Center, Petah Tikva 49100, Israel. E-mail address: [email protected] (L. Vidal). http://dx.doi.org/10.1016/j.ejca.2017.01.021 0959-8049/ª 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
2
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
KEYWORDS Follicular lymphoma; Maintenance; Rituximab; Meta-analysis
Abstract Background: Randomised trials of rituximab maintenance (MR) for patients with follicular lymphoma support improved progression-free survival (PFS), but the effect on overall survival has been inconclusive. To evaluate the effect of MR on overall survival according to patient and disease characteristics, and to explore certain adverse events, we performed an individual patient data (IPD) meta-analysis. Methods: All investigators of randomised controlled trials that compared MR therapy with observation or treatment only at relapse (no MR) for patients with follicular lymphoma were invited to participate in an IPD meta-analysis. We obtained baseline patient and disease characteristics and time to progression and death for each patient. All analyses took into account the trial and original randomised treatment group. We analysed data in two ways: a two-stage analysis and a multivariate model including patient and disease characteristics. Findings: Seven trials including 2315 patients were analysed. Overall survival of patients improved with MR compared with no MR (hazard ratio [HR] 0.79, 95% CI 0.66e0.96). We could not detect any patient or disease characteristics that were associated with a survival benefit with MR. In all of the models, MR had a beneficial effect on overall survival compared with observation for all types of patients, which was not shown in a particular subgroup in which the patient had already received rituximab in the induction phase and received first-line therapy. MR improved PFS compared with observation (HR 0.57, 95% CI 0.51e0.64). The risk of adverse events was higher with MR, specifically infection of any grade and grade 3e4 infections. Interpretation: Based on IPD from randomised controlled trials, MR improves overall survival consistently in all patients, regardless of patient and disease characteristics when compared with observation, and should be prescribed after a successful induction with R-CVP or R-CHOP for patients with follicular lymphoma. It is still uncertain if that holds when the patient has already received rituximab in his/hers first induction. The effect of MR after bendamustine-rituximab induction compared with rituximab at progression should be further explored. ª 2017 Elsevier Ltd. All rights reserved.
1. Introduction
2. Methods
Randomised controlled trials support the benefit of rituximab maintenance treatment (MR) for patients with follicular lymphoma. Their results have consistently shown improved progression-free survival (PFS), but evidence of improved overall survival has been inconclusive. A summary data meta-analysis of studies addressing patients with follicular lymphoma demonstrated an improved overall survival with MR among patients with relapsed or refractory lymphoma (hazard ratio [HR] for death 0.72 95% CI 0.57e0.91, 909 patients) [1]. It remains unclear whether MR has a similar effect in other specific subgroups of patients. To evaluate the effect of rituximab maintenance on overall survival according to patient, disease and treatment characteristics, and to explore specific adverse events, we performed an individual patient data (IPD) meta-analysis. IPD meta-analysis, considered the gold-standard of meta-analysis, allows analyses that are not feasible using summary data such as subgroup analysis of overall survival or analysis of specific adverse events that are not uniformly reported in the literature.
2.1. Inclusion criteria We included all randomised controlled trials that compared rituximab maintenance therapy given after induction with observation or treatment only at relapse in patients with histologically confirmed B-cell follicular lymphoma. In all the trials that contributed data for this study, patients in the control group were observed and none used treatment with rituximab at progression. We included trials regardless of publication status, date of publication and language. 2.2. Search for trials The process of systematic review including search and study selection was previously described [1]. The search was updated in June 2014. Eleven eligible trials performed between 1998 and 2009 were found [2e13] and the investigators of these trials were invited to cooperate. A collaborative group consisting of investigators from seven study groups that contributed IPD was formed.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Systematic review and meta-analysis of IPD were performed as described in the Cochrane Handbook [14]. The following baseline (before induction) variables were obtained for each patient: age, sex, performance status, presence of B symptoms, Ann Arbor stage, a measure of bulkiness (7 cm), follicular lymphoma international prognostic index (FLIPI) score, type of induction, treatment line and response to induction (complete response, partial response and less than partial response). We attempted to collect the baseline variables haemoglobin level, lactate dehydrogenase level, white blood count, number of involved areas and bone marrow involvement, but could not include them in the analysis due to the high rate of missing data. The data of each trial were checked for consistency. Queries, including missing variables were sent to the principal investigators and changes were made to the data if necessary accordingly. 2.3. Outcome measures The primary end-point was overall survival defined as the time from randomisation till death. The survival times of patients still alive were censored at the date of last follow up. Secondary end-points were PFS and toxicity. Data on time to next chemotherapy were unavailable in most of the trials. Preplanned subgroup analyses for the primary end-point by sex, age, FLIPI, type of induction therapy, treatment line and response to induction therapy were performed. For all trials, the methods of allocation concealment and sequence generation were adequate and we judged all trials to be at low risk of selection bias according to the Cochrane tool for assessing bias [14]. 2.4. Statistical analysis All analyses took into account the trial and original randomised treatment group: no comparisons of patients across trials were made. Analyses were done according to the intention to treat principle. We analysed data in two ways: (a) using a two-stage analysis and (b) using a multivariate model. In the two-stage analysis, HRs for each trial were assessed using Cox regression analysis and were then pooled using a fixed effect model. We performed two sensitivity analyses: (i) pooling results using a randomeffects model and (ii) including the results (based on summary data) of the four additional eligible trials for which we did not have IPD. In the multivariate model, we used Cox regression stratified by trial with explanatory variables that included treatment and a variety of different covariates as described below. The proportional hazards assumption for the treatment effect was checked and found to
3
hold. The interaction between treatment effect and trial was checked. The following variables were each entered in separate models together with the treatment group variable in stratified Cox regression analyses (we call it ‘bivariable analysis’ as two variablesdtreatment group and one covariatedwere entered to each model): age, gender, baseline performance status, presence of B symptoms, Ann Arbor stage, length of the largest lymph node as a measure of bulky disease, rituximab-containing induction, type of induction chemotherapy and response before maintenance. The interactions of each of these variables with the treatment group effect were also investigated. Following these analyses, all variables with a P value < 0.1 were included simultaneously in a larger multivariate stratified Cox regression model analysis. We analysed adverse events as survival data with time-dependent covariate using time to the first event, and also ignoring time as dichotomous data. In the latter case, we used the two-stage analysis approach described earlier.
3. Results 3.1. Included trials Seven trials contributed data for this collaborative project [4,6e8,11,12,15]. The database included 2317 randomised patients. Two patients had unclear survival status and date and were excluded from analysis. In all trials, the control group was observed without rituximab treatment and did not receive rituximab at progression while on study, but rituximab was allowed when treatment was indicated. The characteristics of each trial are described in the Table 1. Patients’ characteristics are described in Table 1 by study and by allocated group. The median age of all patients was 57 years (range 23e87). In all the trials, the vast majority of patients had advanced stage. In three trials, all patients received their first-line induction [8,11,15], in three trials all patients had received their second (or more) induction [7,9,12] and in one trial [4,6] 34.5% of patients received first-line induction and the remainder had received at least second-line induction. Most patients received induction that included chemotherapy. Induction chemotherapy was cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) based in 62% of the patients, cyclophosphamide, vincristine and prednisone (CVP) based in 23% of the patients and fludarabine-based chemotherapy in 15% of patients with no difference between the patients randomised to MR and no MR. Induction included rituximab for 74% of patients in the IPD analysis (Table 1). Median follow-up was 6 years (IQR 4.9e6.8).
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
4
Study characteristics and patients’ characteristics by trial.
Number of patients Induction line
Hochster
Ghielmini
Pettengell
Salles
Vitolo
Van Oers
Witzens-Harig
279
151
280
1030
202
334
41
Treatment naive
Relapsed
Treatment naive
Treatment naı¨ve
Relapsed
Induction followed by BEAM and ASCT A single infusion every 3 months for 2 years 76.8
Rituximab and CHOP, CVP or FCM A single infusion every 2 months for 2 years 36
Rituximab-FND
CHOP Rituximab
Treatment naı¨ve and relapsed Any induction
A single infusion every 2 months for 4 doses 42
A single infusion every 3 months for 2 years 84
Induction regimen before MR
CVP/FC
Treatment naı¨ve and relapsed Rituximab
MR schedule
4 weekly infusions every 6 months for 2 years 48
A single infusion every 2 months for 4 doses 114
Median follow up (months)
Female (%) Age, median (IQR) B symptoms (%) Ann Arbor stage 3 Ann Arbor stage 4 4 nodal sites (%) Haemoglobin 12 (%) FLIPI % Low % Intermediate % High Complete response to induction (%)
A single infusion every 3 months for 2 years 28
MR
Obs.
MR
Obs.
MR
Obs.
MR
Obs.
MR
Obs.
MR
Obs.
MR
Obs.
48.5 53 (46e62) 22.5 36 62 NA 87
49.6 56 (49e66) 27 35 65 NA 84
62.3 58 (51e65) 24.7 32 52 23.0 73
52.7 57 (50e64) 29.9 35 55 23.4 75
50.7 52 (45e57) 30.8 29 52 NA NA
45.0 53 (46e57) 25 21 47 NA NA
46.5 56 (47e64) 32.0 21 72 74.8 78.5
49 58 (49e64) 31.1 19 72 74.1 80.3
63.3 66 (63e70) 15.8 29 59 67.3 9.0
55.4 67 (65e71) 14.9 16 68 63.4 10.0
53.3 55 (49e62) 24.6 34 65 37.7 75.5
50.3 54 (48e59) 23.4 34 65 40.1 77.7
52.9 57 (44e64) 25.0 25 38 NA NA
29.2 56 (51e66) 25.0 21 52 NA NA
NA
NA
38.8 32.7 34.5 33.65 26.7 33.65 84/280 (30)
NA
NA
26.1 26.67 36.4 36.67 37.5 36.67 61/271 (22.5)
14/145 (9.7)
21.3 21.4 36.2 36.1 42.5 42.5 728/1030 (70.7)
36.6 29.7 44.6 55.4 18.8 14.9 160/201 (79.6)
37.7 34.7 35.3 30.0 27 35.3 96/323 (29.7)
NA
ASCT Z autologous stem cell transplantation; BEAM Z BCNU, etoposide, cytarabine and melphalan; CVP Z cyclophosphamide, vincristine and prednisone; FC Z fludarabine and cyclophosphamide; FCM Z fludarabine, cyclophosphamide and mitoxantrone; CHOP Z cyclophosphamide, doxorubicin, vincristine and prednisone; MR Z rituximab maintenance; FLIPI Z follicular lymphoma international prognostic index; IQR Z interquartile range.
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
Table 1
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
5
Table 2 Models of overall survival and subgroup analyses. Models include study, treatment group, one patient’s characteristic and the interaction of patient’s characteristics with the effect of MR. Variable All Gender Age ECOG performance status Stage
Tumour size B symptoms FLIPI
Treatment line Rituximab induction
Response to induction
Subgroup analysis Type of chemotherapy CHOP CVP Fludarabine based MR after, treatment line First induction 2 or more inductions
Male Female Age < 60 years Age 60 years 0 1 2 1 2 3 4 <7 cm 7 cm Present Low Intermediate High After first line After 2 treatment lines Without rituximab Rituximab-containing induction CR PR Other (less than PR)
Number of patients
HR
95% CI
HR of MR versus no MR
95% CI of MR versus no MR
P value interaction
2315 2315
NA 1 0.75 1 1.99 1 1.34 3.43 1 0.85 0.84 1.12 1 1.31 1.27 1 1.40 2.64 1 1.77 1 0.87
NA
0.79 0.79
0.66, 0.96 0.66, 0.95
NA
0.80
0.66, 0.96
0.77
0.64, 0.93
2315 2284
2283
2080 2046 1965
2314 2315
2315
1 1.24 2.24
0.62, 0.90
0.20
1.64, 2.41
0.59
1.1, 1.64 2.26, 5.22
0.004 <0.001 0.77
0.64, 0.94
0.38, 1.89 0.41, 1.72 0.55, 2.27 1.02, 1.70 1.02, 1.58
0.77
0.64, 0.93
0.80 0.83
0.65, 0.98 0.67, 1.02
1.03, 1.91 1.98, 3.51
0.69 0.63 0.76 0.26 0.57 0.56 0.78
0.79
0.66, 0.96
0.79
0.66, 0.95
0.95, 3.29
0.15
0.66, 1.16
0.37 0.8
0.66, 0.96
0.99, 1.55 1.55, 3.23
0.67 0.56
1194 452 233
0.78 0.77 1.97
0.59, 1.03 0.51, 1.16 0.59, 6.58
1561 753
0.92 0.70
0.70, 1.20 0.54, 0.91
CI Z confidence interval; CR Z complete response; CHOP Z cyclophosphamide, doxorubicin, vincristine, prednisone; CVP Z cyclophosphamide, vincristine, prednisone; ECOG Z Eastern Cooperative Oncology Group; FLIPI Z follicular lymphoma international prognostic index; HR Z hazards ratio; MR Z rituximab maintenance; PR Z partial response.
3.2. Overall survival
3.3. Models and subgroup analyses
2315 patients were included in the Cox regression analysis of overall survival. Of the 1145 patients in MR group, 209 patients and 246 of 1170 patients in no MR group died during follow up. The HR of death was 0.79 (95% CI 0.66e0.96) for MR compared with no MR (Table 2, Fig. 1A). Median overall survival in the MR group was 12 years (95% CI 11.5 to not yet reached) and in the observation group 11.5 years (95% CI 11 to not yet reached). Using the two-stage meta-analysis, there was no change in the results (HR 0.79, 95% CI 0.65e0.95 fixed effect model; HR 0.78, 95% CI 0.63e0.96, random-effects model). We performed a sensitivity analysis including all eligible trials based on the IPD of seven trials and in addition summary data of three trials [2,3,5]. One trial that fulfilled the inclusion criteria did not report overall survival data (Ardeshna et al.). The HR for death with MR was 0.79, 95% CI 0.66e0.94, 2786 patients, Fig. 2.
Only five trials provided information on FLIPI [7e9,11,15]. In analyses including the trial indicator, the treatment group and one other variable, female sex, age below 60 years, Eastern Cooperative Oncology Group (ECOG) performance score (PS) of 0, low FLIPI and achieving at least partial response after induction increased the duration of overall survival (p of each <0.05). Conversely, bulky disease reduced survival (p < 0.05). An effect of stage or treatment line on overall survival was not shown in the models that included treatment group. We evaluated whether the effect of MR on survival depends on patient characteristics by examining the statistical interactions between treatment group and each of the other variables. No statistically significant interactions with the effect of MR on overall survival were found (Table 2); thus, we could detect no subgroups that benefited particularly from MR. In these models, the beneficial effect on overall survival of MR
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
6
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Fig. 1. KaplaneMeier curve of overall survival, A) by intervention group; B) by treatment line and intervention group.
compared with observation remained statistically significant regardless of sex, age, ECOG PS, stage, bulky disease, B symptoms and rituximab-containing induction, even after taking into account the low risk characteristics such as a MR after first-line induction or achieving a complete response after induction. As the type of induction chemotherapy was standardised in most of the trials (only in one trial the type of chemotherapy was chosen within the trial by the treating physician), we did not include the type of chemotherapy in a Cox model, but performed subgroup analysis by that variable. The interaction with the inclusion of rituximab in induction was evaluated in the models. The effect of MR was not statistically significantly different with R-CVP/CVP or R-CHOP/CHOP (Table 2). The HR of MR effect for the 233 patients who received fludarabine-based induction was 1.97 with a wide CI. As the current therapy of follicular lymphoma includes rituximab, we performed a subgroup analysis of 1755 patients who received rituximab as part of induction. Although in the model that included both allocation group (MR versus no MR) and rituximabcontaining induction, the effect of MR remained
statistically significant; in this subgroup analysis of rituximab as part of induction, no statistically significant difference was shown (HR 0.85, 95% CI 0.67e1.07). Yet, this effect of MR was not different from its effect in the whole cohort. Repeating all models, including a model that included treatment line and treatment arm demonstrated similar results. Subgroup analysis by the induction line (MR after the first induction; MR after two or more inductions) demonstrated a statistically significant benefit of MR after second induction (HR 0.70, 95% CI 0.54e0.91), while with MR after first induction the HR was 0.92, 95% CI 0.70e1.20 (Fig. 1B).Yet, there was no statistically significant difference in the HRs between the subgroups (test for subgroup differences P Z 0.4). Median overall survival after first line was not yet reached in the MR group and the no MR group. A 5-year overall survival rate in these two groups were 89.2% (95% CI 86.8%e91.5%) and 88% (85.6%e90.4%), respectively. Estimated median overall survival in the MR group after second line was 11.2 years (95% CI 10.9e12.4) and in the observation group 10.3 years (95% CI 8.3e11.5). The 5year overall survival rates in these groups were 76% (95% CI 72%e81%) and 71% (67%e76%), respectively. Following these analyses, we performed a post hoc 2 2 analysis of four subgroups by treatment line and by the inclusion of rituximab in induction. MR following rituximab induction for relapse/refractory lymphoma statistically and significantly improved survival. The point estimates of the effect of MR following induction without rituximab for a frontline therapy and for relapse/refractory lymphoma were also similar, as shown in Table 3. Among patients who received rituximab in the induction phase and were receiving first-line therapy, rituximab did not appear to benefit in terms of overall survival (HR Z 1.049). The difference between 1.049 and 0.70 is not statistically significant, P Z 0.09; thus, one cannot be sure whether this difference is really a chance finding that has come up while looking at many different subgroups or is a true phenomenon. 3.4. Multivariable models A multivariate analysis yielded two main models: model 1 included all variables that were significant on bivariate analysis (sex, age, ECOG PS, bulky disease and FLIPI) as well as study and allocated treatment group. Sex, age, ECOG PS, FLIPI and allocated treatment group remained statistically significant, while bulky disease was not. MR statistically significantly improved overall survival (HR 0.77, 95% CI 0.64e0.93, P Z 0.007). With the addition of treatment line (2nd versus 1st) to the model, the effect of MR did not change. 3.5. Progression-free survival MR improved PFS compared with observation for patients with follicular lymphoma (HR 0.57, 95% CI
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
7
Fig. 2. Forest plot of overall survival of rituximab maintenance (MR) versus no MR based on individual patient data and summary data.
0.51e0.64, p < 0.0001, 2305 patients). Twelve patients were excluded due to missing data (five in the MR group and seven in observation group). We checked for the effect and interaction of treatment group with the following variables: study, sex, age, FLIPI, treatment line, type of chemotherapy and response to induction. Age, treatment line (relapsed/refractory), CVP chemotherapy, low FLIPI and response to induction were prognostic factors for PFS. As with overall survival, there were no statistically significant interactions of any of the variables with the effect of MR on PFS. 3.6. Adverse events We analysed the risk of any adverse event, grade 3 or 4 adverse events and specific adverse events (Table 4). The rate of any adverse event grade 1e4 in MR group was 52% and in the no MR group 40%, the corresponding rates of any grade 3 to 4 were 17% and 23%, respectively
%. The risk of adverse events increased with MR compared with observation (no significant statistical heterogeneity). The most frequent adverse event was infection and occurred in 23.6% of the observation group (4.9% grade 3e4), and in 33.6% of the MR group (7.1% grade 3e4). The risks of infection of any grade and infection of grade 3e4 increased with MR (Table 4). The HR of grade 3e4 infection was 1.27, 95% CI 1.12 to 1.45. Other adverse events were rarely reported. 4. Discussion Based on the data of 2315 patients, this meta-analysis demonstrated an overall survival benefit with MR treatment compared with no MR, expressed as the number needed to treat, 29 patients would be treated with MR to prevent one death (95% CI 17e153) within 5 years [16]. The improvement in overall survival was consistent across all types of patients and unrelated to sex, age, FLIPI risk group, rituximab-containing
Table 3 Analysis of the effect of MR by treatment line and inclusion of rituximab in induction. 2 line
First line
No rituximab induction Induction included rituximab
Events/number
HR, 95% CI
Events/number
HR, 95% CI
73/278
0.71, 0.45e1.14 p Z 0.15 1.049, 0.75e1.46 p Z 0.78
91/282
0.69, 0.46e1.04 p Z 0.075 0.70, 0.51e0.98 p Z 0.035
140/1283
151/471
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
8
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Table 4 Adverse events. Adverse event
Hazard ratio MR versus observation
95% confidence interval
Rate with MR
Rate without MR
Any, grade 1e4 Any, grade 3e4 Infection, grade 1e4 Infection, grade 3e4 Neutropenic fever
1.27 1.31 1.41
1.12, 1.45 1.08, 1.58 1.2, 1.66
52% 23% 33.6%
40% 17% 23.6%
1.48
1.04, 2.11
7.1%
4.9%
0.78
0.43, 1.43
1.8%
2.1%
induction or response to induction. It should be noted that there was no evidence of a different effect with MR with relapsed/refractory lymphoma compared with after first induction, and that in the multivariable analysis model, after adjustment for treatment line, the beneficial effect of MR on survival remained statistically significant. Overall rituximab improves survival when added to maintenance therapy. However, the caveat is that this conclusion may not hold for all circumstances, and based on the analysis of subgroups one circumstance that is particularly uncertain is when the patient has already received rituximab in the induction phase and is receiving first-line therapy. In the other circumstances explored here (either no rituximab in the induction or second line or later therapy), the advantage of rituximab maintenance seems clear. PFS was longer with MR treatment in all subgroups and models. The major toxicity was the increased risk of infection. Based on the outcomes of this meta-analysis, the benefit of MR after frontline induction may still be questionable. One explanation for the apparently smaller benefit in the subgroup analysis after first-line induction could be the longer survival (the median of which was not reached) in the frontline induction group leading to a smaller number of events and a need for larger sample size or a much longer follow-up, to reduce uncertainty. It is possible that with a longer follow-up or a larger sample size, there would be less uncertainty in the results. However, it is unlikely that such a trial would be conducted. It might be possible that certain subgroups within the first-line group may benefit even more from MR. For example, in a pooled analysis of three trials, patients who had a positive FDG-PET at the end of induction had worse PFS and overall survival [17]. These results and the debate regarding the survival benefit after first induction led to an ongoing trial that compares MR for 2 years as the standard treatment after induction to a response-adapted strategy based on end of treatment PET and minimal residual disease (NCT02063685, EudraCT 2012-003170-60). Of note, in our database, MR had a favourable effect on survival regardless of the depth of response.
The main strength of this meta-analysis lies in its very large sample size making it sufficiently powered to show a treatment effect that was not conclusively demonstrated in each of the individual trials. Analysis based on IPD allows inclusion of all randomised patients and exploration of the effect of patient and disease characteristics on the effect of MR. The trials differed with respect to inclusion criteria, type of induction therapy, the MR schedule and also the frequency of assessments. However, the similarity between the different trials, mainly the inclusion of patients with follicular lymphoma after induction therapy (either as first line or relapse) provides a solid basis for this meta-analysis. It should be emphasised that the clinical heterogeneity across the trials did not translate into statistical heterogeneity of the results. The inclusion period of most of the included trials dates back 10 years or more. The survival perspective of today’s patients might be influenced by current and future advances in the treatment of follicular lymphoma such as new antibodies and pathway inhibitors. These newer treatments might impact and interact with the effect of MR. The results of this IPD meta-analysis are consistent with the results of a summary data meta-analysis with the same clinical question in patients with follicular lymphoma [1] and confirm its findings. Six of the seven trials included in the current meta-analysis were also included in the summary data meta-analysis. The previous meta-analysis, as the current one, showed improved overall survival with MR for the whole cohort of follicular lymphoma patients. Subgroup analysis of patients after frontline induction did not support a differential effect from that of patients with relapsed/refractory lymphoma. The advantage of the IPD metaanalysis compared with the summary data metaanalysis is that it enabled us to incorporate patient characteristics and intervention (MR group) into a model to differentiate patients who would benefit from MR. Regarding the effect of MR after single-agent rituximab, we could not conduct subgroup analysis in this IPD analysis, since only one trial [4,6] was included. Yet, the previous summary data meta-analysis pooled data of three trials with patients whose induction therapy consisted of single-agent rituximab with no chemotherapy [1] showed a trend toward improved overall survival with MR compared with no MR (HR Z 0.76, 95% CI Z 0.53e1.01), but again no differential effect from that of the whole cohort. Current guidelines (ESMO 2014) [18] recommend the use of rituximab maintenance for patients with follicular lymphoma based on overall survival benefit in relapsed disease [I, A] and on PFS benefit after first-line induction [I, B]. These recommendations are based on the results of the previous summary data meta-analysis.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Thus, our recent results are in agreement with current guidelines. All the trials in the current meta-analysis did not use an active control group, and MR was compared only with patients under observation and not to rituximab at progression. This study therefore does not address the question of whether MR is superior to treatment of patients with rituximab at progression. To conclude, based on data from randomised controlled trials, MR improves overall survival when compared with observation and should be prescribed after a successful induction with R-CVP or an R-CHOP regimen. It is still uncertain if that is true in the situation that the patient has already received rituximab in his/ hers first induction. To be elucidated is the reason for the questionable effect for these patients. Optional explanations are an interaction between the type of chemotherapy and MR, a deeper response in the frontline setting or a biologic difference of relapse/refractory follicular lymphoma from a newly diagnosed lymphoma. It is unclear whether rituximab treatment at progression is equivalent to MR. Also, it is unknown whether MR has similar effectiveness after bendamustine-rituximab induction, which nowadays is one of the used induction therapies for follicular lymphoma [18,19], since none of the trials used bendamustine-rituximab as induction. These questions should be answered in future clinical trials.
Conflict of interest statement Dr. Vidal reports Celgene, Novartis, Roche: travel, and accommodations. Dr. Ghielmini has received honoraries as lecturer from the following companies: Roche, Cellgene, Gilead, Jannsen, Abbvie, Mundipharma. Dr. Salles reports grants and personal fees from Roche, during the conduct of the study; personal fees from Jannsen, personal fees from Gilead, personal fees from Celgene, personal fees from Novartis, personal fees from Amgen, personal fees from Servier, personal fees from BMS, personal fees from Merck, outside the submitted work. Dr.Dreger reports participation in advisory board: Roche. Dr. Gomes Silva reports the following: Bristol-Myers Squibb, Celgene, Gilead Sciences, Janssen-Cilag, Takeda: consultancy. Bristol-Myers Squibb, Celgene, Ferrer, JanssenCilag, Takeda: speakers’ bureau; Gilead Sciences: research funding. Celgene, Gilead Sciences, Janssen-Cilag, Roche: travel, accommodations, expenses. Dr. Vitolo reports participation in advisory board: Roche, Celgene, Janssen; and honoraria for lectures: Roche, Celgene, Takeda, Gilead, Mundipharma.
9
Dr. Shpilberg reports a grant from Roche. The other authors have no conflict of interest to declare. Funding The study was partly supported by Roche (KM209450801). Roche was not involved in any part of study design, conduction and interpretation.
Acknowledgements The authors thank the European Organization for Research and Treatment of Cancer for the permission to use the data from EORTC trial 20981 for this research and the European Society for Blood and Marrow Transplantation for using data from the Lym1 trial for this research. We would like to thank Professor Leonard Leibovici for his advice and critical review.
References [1] Vidal L, Gafter-Gvili A, Salles G, et al. Rituximab maintenance for the treatment of patients with follicular lymphoma: an updated systematic review and meta-analysis of randomized trials. J Natl Cancer Inst 2011;103:1799e806. [2] Ardeshna KM, Qian W, Smith P, et al. Rituximab versus a watchand-wait approach in patients with advanced-stage, asymptomatic, non-bulky follicular lymphoma: an open-label randomised phase 3 trial. [Internet]. Lancet Oncol 2014;15:424e35 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm.nih.gov/pubmed/ 24602760. [3] Forstpointner R, Unterhalt M, Dreyling M, et al. Maintenance therapy with rituximab leads to a significant prolongation of response duration after salvage therapy with a combination of rituximab, fludarabine, cyclophosphamide, and mitoxantrone (RFCM) in patients with recurring and refractory follicular [Internet]. Blood 2006;108:4003e8 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm.nih.gov/pubmed/16946304. [4] Ghielmini M, Schmitz S-FH, Cogliatti SB, et al. Prolonged treatment with rituximab in patients with follicular lymphoma significantly increases event-free survival and response duration compared with the standard weekly x 4 schedule. [Internet]. Blood 2004;103:4416e23. Available from: http://www.ncbi.nlm.nih.gov/ pubmed/14976046. [5] Hainsworth JD, Litchy S, Shaffer DW, et al. Maximizing therapeutic benefit of rituximab: maintenance therapy versus retreatment at progression in patients with indolent non-Hodgkin’s lymphoma e a randomized phase II trial of the Minnie Pearl Cancer Research Network. [Internet]. J Clin Oncol 2005;23: 1088e95 [cited 2016 Feb 24] Available from: http://www.ncbi. nlm.nih.gov/pubmed/15657401. [6] Martinelli G, Schmitz SFH, Utiger U, et al. Long-term follow-up of patients with follicular lymphoma receiving single-agent rituximab at two different schedules in trial SAKK 35/98. J Clin Oncol 2010;28:4480e4. [7] Pettengell R, Schmitz N, Gisselbrecht C, et al. Rituximab purging and/or maintenance in patients undergoing autologous transplantation for relapsed follicular lymphoma: a prospective randomized trial from the lymphoma working party of the European group for blood and marrow transplantation. [Internet]. J Clin Oncol 2013;31:1624e30 [cited 2016 Feb 24] Available from: http:// www.ncbi.nlm.nih.gov/pubmed/23547078.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
10
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
[8] Salles G, Seymour JF, Offner F, et al. Rituximab maintenance for 2 years in patients with high tumour burden follicular lymphoma responding to rituximab plus chemotherapy (PRIMA): a phase 3, randomised controlled trial. Lancet 2011;377:42e51. [9] Van Oers MHJ, Van Glabbeke M, Giurgea L, et al. Rituximab maintenance treatment of relapsed/resistant follicular non-Hodgkin’s lymphoma: long-term outcome of the EORTC 20981 phase III randomized intergroup study. J Clin Oncol 2010;28:2853e8. [10] Ghielmini M, Vitolo U, Kimby E, et al. ESMO Guidelines consensus conference on malignant lymphoma 2011 part 1: diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL) and chronic lymphocytic leukemia (CLL). [Internet]. Ann Oncol 2013;24:561e76 [cited 2015 Jul 12] Available from: http:// www.ncbi.nlm.nih.gov/pubmed/23175624. [11] Vitolo U, Ladetto M, Boccomini C, et al. Rituximab maintenance compared with observation after brief first-line R-FND chemoimmunotherapy with rituximab consolidation in patients age older than 60 years with advanced follicular lymphoma: a phase III randomized study by the Fondazione Italiana Lin [Internet]. J Clin Oncol 2013;31:3351e9 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm.nih.gov/pubmed/23960180. [12] Witzens-Harig M, Benner A, McClanahan F, et al. Rituximab maintenance improves survival in male patients with diffuse large B-cell lymphoma. Results of the HD2002 prospective multicentre randomized phase III trial. [Internet]. Br J Haematol 2015;171: 710e9 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm. nih.gov/pubmed/26449739. [13] Kahl BS, Hong F, Williams ME, et al. Rituximab extended schedule or re-treatment trial for low-tumor burden follicular lymphoma: eastern cooperative oncology group protocol e4402.
[14]
[15]
[16]
[17]
[18]
[19]
[Internet]. J Clin Oncol 2014;32:3096e102 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm.nih.gov/pubmed/25154829. Higgins JPT, Altman DG. Higgins 2011 Higgins JPT. Version 5.1.0 [updated March 2011]. The Cochrane Collaboration. In: Green S, editor. Cochrane Handbook for systematic reviews of interventions, 2011; 2011. Hochster H, Weller E, Gascoyne RD, et al. Maintenance rituximab after cyclophosphamide, vincristine, and prednisone prolongs progression-free survival in advanced indolent lymphoma: results of the randomized phase III ECOG1496 study. J Clin Oncol 2009;27:1607e14. Altman DG, Andersen PK. Calculating the number needed to treat for trials where the outcome is time to an event. BMJ 1999; 319:1492e5. Trotman J, Luminari S, Boussetta S, et al. Prognostic value of PET-CT after first-line therapy in patients with follicular lymphoma: a pooled analysis of central scan review in three multicentre studies. [Internet]. Lancet Haematol 2014;1:e17e27 [cited 2016 May 31] Available from: http://www.ncbi.nlm.nih.gov/ pubmed/27030064. Newly Diagnosed and Relapsed Follicular Lymphoma: ESMO Clinical Practice GuidelinesjESMO [Internet][cited 2016 Feb 24] Available from: http://www.esmo.org/Guidelines/Haematolo gical-Malignancies/Newly-Diagnosed-and-Relapsed-FollicularLymphoma. Rummel MJ, Niederle N, Maschmeyer G, et al. Bendamustine plus rituximab versus CHOP plus rituximab as first-line treatment for patients with indolent and mantle-cell lymphomas: an openlabel, multicentre, randomised, phase 3 non-inferiority trial. Lancet 2013;381:1203e10.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Original Research
Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis Liat Vidal a,b,*, Anat Gafter-Gvili a,c, Gilles Salles d, Sami Bousseta e, Bernice Oberman f, Carmit Rubin f, Marinus H.J. van Oers g, Catherine Fortpied h, Michele Ghielmini i,j, Ruth Pettengell k, Mathias Witzens-Harig l, Peter Dreger m, Umberto Vitolo n, Maria Gomes da Silva o, Andrea Evangelista p, Hailun Li q, Laurence Freedman f, Thomas M. Habermann r, Ofer Shpilberg s a
Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel Institute of Hematology, Davidoff Cancer Center, Rabin Medical Center, Petah Tikva, Israel c Internal Medicine A, Rabin Medical Center, Petah Tikva, Israel d Hospices Civils de Lyon, Centre Hospitalier Lyon-Sud, Pierre-Benite, Universite´ Claude Bernard Lyon-1, Lyon, France e Biostatistics Department, LYSARC, Pierre-Benite, France f Sheba Medical Center, Gertner Institute for Epidemiology and Health Policy Research, Tel Hashomer, Israel g EORTC Lymphoma Group/HOVON, Academic Medical Center, Amsterdam, The Netherlands h EORTC, Brussels, Belgium i Oncology Institute of Southern Switzerland, Bellinzona, Switzerland j Swiss Group for Clinical Cancer Research, Bern, Switzerland k Department of Haematology, St. George’s University of London, London, UK l Department of Internal Medicine V, University of Heidelberg, Heidelberg, Germany m EBMT Lymphoma Working Party, Paris, France n Citta` della Salute e della Scienza Hospital and University, on behalf of FIL, Turin, Italy o CEDOC, Instituto Portugueˆs de Oncologia de Lisboa Francisco Gentil EPE Rua Prof. Lima Basto, 1099-023 Lisboa, Portugal p University-Hospital Citta` della Salute e della Scienza, Torino, Italy q Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute, Boston, MS, USA r Division of Hematology, Mayo Clinic, Rochester, MN, USA s Assuta Medical Center, Tel Aviv, Israel b
Received 25 September 2016; received in revised form 15 January 2017; accepted 24 January 2017 Available online - - -
* Corresponding author: Institute of Hematology, Davidoff Cancer Center, Rabin Medical Center, Petah Tikva 49100, Israel. E-mail address: [email protected] (L. Vidal). http://dx.doi.org/10.1016/j.ejca.2017.01.021 0959-8049/ª 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
2
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
KEYWORDS Follicular lymphoma; Maintenance; Rituximab; Meta-analysis
Abstract Background: Randomised trials of rituximab maintenance (MR) for patients with follicular lymphoma support improved progression-free survival (PFS), but the effect on overall survival has been inconclusive. To evaluate the effect of MR on overall survival according to patient and disease characteristics, and to explore certain adverse events, we performed an individual patient data (IPD) meta-analysis. Methods: All investigators of randomised controlled trials that compared MR therapy with observation or treatment only at relapse (no MR) for patients with follicular lymphoma were invited to participate in an IPD meta-analysis. We obtained baseline patient and disease characteristics and time to progression and death for each patient. All analyses took into account the trial and original randomised treatment group. We analysed data in two ways: a two-stage analysis and a multivariate model including patient and disease characteristics. Findings: Seven trials including 2315 patients were analysed. Overall survival of patients improved with MR compared with no MR (hazard ratio [HR] 0.79, 95% CI 0.66e0.96). We could not detect any patient or disease characteristics that were associated with a survival benefit with MR. In all of the models, MR had a beneficial effect on overall survival compared with observation for all types of patients, which was not shown in a particular subgroup in which the patient had already received rituximab in the induction phase and received first-line therapy. MR improved PFS compared with observation (HR 0.57, 95% CI 0.51e0.64). The risk of adverse events was higher with MR, specifically infection of any grade and grade 3e4 infections. Interpretation: Based on IPD from randomised controlled trials, MR improves overall survival consistently in all patients, regardless of patient and disease characteristics when compared with observation, and should be prescribed after a successful induction with R-CVP or R-CHOP for patients with follicular lymphoma. It is still uncertain if that holds when the patient has already received rituximab in his/hers first induction. The effect of MR after bendamustine-rituximab induction compared with rituximab at progression should be further explored. ª 2017 Elsevier Ltd. All rights reserved.
1. Introduction
2. Methods
Randomised controlled trials support the benefit of rituximab maintenance treatment (MR) for patients with follicular lymphoma. Their results have consistently shown improved progression-free survival (PFS), but evidence of improved overall survival has been inconclusive. A summary data meta-analysis of studies addressing patients with follicular lymphoma demonstrated an improved overall survival with MR among patients with relapsed or refractory lymphoma (hazard ratio [HR] for death 0.72 95% CI 0.57e0.91, 909 patients) [1]. It remains unclear whether MR has a similar effect in other specific subgroups of patients. To evaluate the effect of rituximab maintenance on overall survival according to patient, disease and treatment characteristics, and to explore specific adverse events, we performed an individual patient data (IPD) meta-analysis. IPD meta-analysis, considered the gold-standard of meta-analysis, allows analyses that are not feasible using summary data such as subgroup analysis of overall survival or analysis of specific adverse events that are not uniformly reported in the literature.
2.1. Inclusion criteria We included all randomised controlled trials that compared rituximab maintenance therapy given after induction with observation or treatment only at relapse in patients with histologically confirmed B-cell follicular lymphoma. In all the trials that contributed data for this study, patients in the control group were observed and none used treatment with rituximab at progression. We included trials regardless of publication status, date of publication and language. 2.2. Search for trials The process of systematic review including search and study selection was previously described [1]. The search was updated in June 2014. Eleven eligible trials performed between 1998 and 2009 were found [2e13] and the investigators of these trials were invited to cooperate. A collaborative group consisting of investigators from seven study groups that contributed IPD was formed.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Systematic review and meta-analysis of IPD were performed as described in the Cochrane Handbook [14]. The following baseline (before induction) variables were obtained for each patient: age, sex, performance status, presence of B symptoms, Ann Arbor stage, a measure of bulkiness (7 cm), follicular lymphoma international prognostic index (FLIPI) score, type of induction, treatment line and response to induction (complete response, partial response and less than partial response). We attempted to collect the baseline variables haemoglobin level, lactate dehydrogenase level, white blood count, number of involved areas and bone marrow involvement, but could not include them in the analysis due to the high rate of missing data. The data of each trial were checked for consistency. Queries, including missing variables were sent to the principal investigators and changes were made to the data if necessary accordingly. 2.3. Outcome measures The primary end-point was overall survival defined as the time from randomisation till death. The survival times of patients still alive were censored at the date of last follow up. Secondary end-points were PFS and toxicity. Data on time to next chemotherapy were unavailable in most of the trials. Preplanned subgroup analyses for the primary end-point by sex, age, FLIPI, type of induction therapy, treatment line and response to induction therapy were performed. For all trials, the methods of allocation concealment and sequence generation were adequate and we judged all trials to be at low risk of selection bias according to the Cochrane tool for assessing bias [14]. 2.4. Statistical analysis All analyses took into account the trial and original randomised treatment group: no comparisons of patients across trials were made. Analyses were done according to the intention to treat principle. We analysed data in two ways: (a) using a two-stage analysis and (b) using a multivariate model. In the two-stage analysis, HRs for each trial were assessed using Cox regression analysis and were then pooled using a fixed effect model. We performed two sensitivity analyses: (i) pooling results using a randomeffects model and (ii) including the results (based on summary data) of the four additional eligible trials for which we did not have IPD. In the multivariate model, we used Cox regression stratified by trial with explanatory variables that included treatment and a variety of different covariates as described below. The proportional hazards assumption for the treatment effect was checked and found to
3
hold. The interaction between treatment effect and trial was checked. The following variables were each entered in separate models together with the treatment group variable in stratified Cox regression analyses (we call it ‘bivariable analysis’ as two variablesdtreatment group and one covariatedwere entered to each model): age, gender, baseline performance status, presence of B symptoms, Ann Arbor stage, length of the largest lymph node as a measure of bulky disease, rituximab-containing induction, type of induction chemotherapy and response before maintenance. The interactions of each of these variables with the treatment group effect were also investigated. Following these analyses, all variables with a P value < 0.1 were included simultaneously in a larger multivariate stratified Cox regression model analysis. We analysed adverse events as survival data with time-dependent covariate using time to the first event, and also ignoring time as dichotomous data. In the latter case, we used the two-stage analysis approach described earlier.
3. Results 3.1. Included trials Seven trials contributed data for this collaborative project [4,6e8,11,12,15]. The database included 2317 randomised patients. Two patients had unclear survival status and date and were excluded from analysis. In all trials, the control group was observed without rituximab treatment and did not receive rituximab at progression while on study, but rituximab was allowed when treatment was indicated. The characteristics of each trial are described in the Table 1. Patients’ characteristics are described in Table 1 by study and by allocated group. The median age of all patients was 57 years (range 23e87). In all the trials, the vast majority of patients had advanced stage. In three trials, all patients received their first-line induction [8,11,15], in three trials all patients had received their second (or more) induction [7,9,12] and in one trial [4,6] 34.5% of patients received first-line induction and the remainder had received at least second-line induction. Most patients received induction that included chemotherapy. Induction chemotherapy was cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) based in 62% of the patients, cyclophosphamide, vincristine and prednisone (CVP) based in 23% of the patients and fludarabine-based chemotherapy in 15% of patients with no difference between the patients randomised to MR and no MR. Induction included rituximab for 74% of patients in the IPD analysis (Table 1). Median follow-up was 6 years (IQR 4.9e6.8).
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
4
Study characteristics and patients’ characteristics by trial.
Number of patients Induction line
Hochster
Ghielmini
Pettengell
Salles
Vitolo
Van Oers
Witzens-Harig
279
151
280
1030
202
334
41
Treatment naive
Relapsed
Treatment naive
Treatment naı¨ve
Relapsed
Induction followed by BEAM and ASCT A single infusion every 3 months for 2 years 76.8
Rituximab and CHOP, CVP or FCM A single infusion every 2 months for 2 years 36
Rituximab-FND
CHOP Rituximab
Treatment naı¨ve and relapsed Any induction
A single infusion every 2 months for 4 doses 42
A single infusion every 3 months for 2 years 84
Induction regimen before MR
CVP/FC
Treatment naı¨ve and relapsed Rituximab
MR schedule
4 weekly infusions every 6 months for 2 years 48
A single infusion every 2 months for 4 doses 114
Median follow up (months)
Female (%) Age, median (IQR) B symptoms (%) Ann Arbor stage 3 Ann Arbor stage 4 4 nodal sites (%) Haemoglobin 12 (%) FLIPI % Low % Intermediate % High Complete response to induction (%)
A single infusion every 3 months for 2 years 28
MR
Obs.
MR
Obs.
MR
Obs.
MR
Obs.
MR
Obs.
MR
Obs.
MR
Obs.
48.5 53 (46e62) 22.5 36 62 NA 87
49.6 56 (49e66) 27 35 65 NA 84
62.3 58 (51e65) 24.7 32 52 23.0 73
52.7 57 (50e64) 29.9 35 55 23.4 75
50.7 52 (45e57) 30.8 29 52 NA NA
45.0 53 (46e57) 25 21 47 NA NA
46.5 56 (47e64) 32.0 21 72 74.8 78.5
49 58 (49e64) 31.1 19 72 74.1 80.3
63.3 66 (63e70) 15.8 29 59 67.3 9.0
55.4 67 (65e71) 14.9 16 68 63.4 10.0
53.3 55 (49e62) 24.6 34 65 37.7 75.5
50.3 54 (48e59) 23.4 34 65 40.1 77.7
52.9 57 (44e64) 25.0 25 38 NA NA
29.2 56 (51e66) 25.0 21 52 NA NA
NA
NA
38.8 32.7 34.5 33.65 26.7 33.65 84/280 (30)
NA
NA
26.1 26.67 36.4 36.67 37.5 36.67 61/271 (22.5)
14/145 (9.7)
21.3 21.4 36.2 36.1 42.5 42.5 728/1030 (70.7)
36.6 29.7 44.6 55.4 18.8 14.9 160/201 (79.6)
37.7 34.7 35.3 30.0 27 35.3 96/323 (29.7)
NA
ASCT Z autologous stem cell transplantation; BEAM Z BCNU, etoposide, cytarabine and melphalan; CVP Z cyclophosphamide, vincristine and prednisone; FC Z fludarabine and cyclophosphamide; FCM Z fludarabine, cyclophosphamide and mitoxantrone; CHOP Z cyclophosphamide, doxorubicin, vincristine and prednisone; MR Z rituximab maintenance; FLIPI Z follicular lymphoma international prognostic index; IQR Z interquartile range.
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
Table 1
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
5
Table 2 Models of overall survival and subgroup analyses. Models include study, treatment group, one patient’s characteristic and the interaction of patient’s characteristics with the effect of MR. Variable All Gender Age ECOG performance status Stage
Tumour size B symptoms FLIPI
Treatment line Rituximab induction
Response to induction
Subgroup analysis Type of chemotherapy CHOP CVP Fludarabine based MR after, treatment line First induction 2 or more inductions
Male Female Age < 60 years Age 60 years 0 1 2 1 2 3 4 <7 cm 7 cm Present Low Intermediate High After first line After 2 treatment lines Without rituximab Rituximab-containing induction CR PR Other (less than PR)
Number of patients
HR
95% CI
HR of MR versus no MR
95% CI of MR versus no MR
P value interaction
2315 2315
NA 1 0.75 1 1.99 1 1.34 3.43 1 0.85 0.84 1.12 1 1.31 1.27 1 1.40 2.64 1 1.77 1 0.87
NA
0.79 0.79
0.66, 0.96 0.66, 0.95
NA
0.80
0.66, 0.96
0.77
0.64, 0.93
2315 2284
2283
2080 2046 1965
2314 2315
2315
1 1.24 2.24
0.62, 0.90
0.20
1.64, 2.41
0.59
1.1, 1.64 2.26, 5.22
0.004 <0.001 0.77
0.64, 0.94
0.38, 1.89 0.41, 1.72 0.55, 2.27 1.02, 1.70 1.02, 1.58
0.77
0.64, 0.93
0.80 0.83
0.65, 0.98 0.67, 1.02
1.03, 1.91 1.98, 3.51
0.69 0.63 0.76 0.26 0.57 0.56 0.78
0.79
0.66, 0.96
0.79
0.66, 0.95
0.95, 3.29
0.15
0.66, 1.16
0.37 0.8
0.66, 0.96
0.99, 1.55 1.55, 3.23
0.67 0.56
1194 452 233
0.78 0.77 1.97
0.59, 1.03 0.51, 1.16 0.59, 6.58
1561 753
0.92 0.70
0.70, 1.20 0.54, 0.91
CI Z confidence interval; CR Z complete response; CHOP Z cyclophosphamide, doxorubicin, vincristine, prednisone; CVP Z cyclophosphamide, vincristine, prednisone; ECOG Z Eastern Cooperative Oncology Group; FLIPI Z follicular lymphoma international prognostic index; HR Z hazards ratio; MR Z rituximab maintenance; PR Z partial response.
3.2. Overall survival
3.3. Models and subgroup analyses
2315 patients were included in the Cox regression analysis of overall survival. Of the 1145 patients in MR group, 209 patients and 246 of 1170 patients in no MR group died during follow up. The HR of death was 0.79 (95% CI 0.66e0.96) for MR compared with no MR (Table 2, Fig. 1A). Median overall survival in the MR group was 12 years (95% CI 11.5 to not yet reached) and in the observation group 11.5 years (95% CI 11 to not yet reached). Using the two-stage meta-analysis, there was no change in the results (HR 0.79, 95% CI 0.65e0.95 fixed effect model; HR 0.78, 95% CI 0.63e0.96, random-effects model). We performed a sensitivity analysis including all eligible trials based on the IPD of seven trials and in addition summary data of three trials [2,3,5]. One trial that fulfilled the inclusion criteria did not report overall survival data (Ardeshna et al.). The HR for death with MR was 0.79, 95% CI 0.66e0.94, 2786 patients, Fig. 2.
Only five trials provided information on FLIPI [7e9,11,15]. In analyses including the trial indicator, the treatment group and one other variable, female sex, age below 60 years, Eastern Cooperative Oncology Group (ECOG) performance score (PS) of 0, low FLIPI and achieving at least partial response after induction increased the duration of overall survival (p of each <0.05). Conversely, bulky disease reduced survival (p < 0.05). An effect of stage or treatment line on overall survival was not shown in the models that included treatment group. We evaluated whether the effect of MR on survival depends on patient characteristics by examining the statistical interactions between treatment group and each of the other variables. No statistically significant interactions with the effect of MR on overall survival were found (Table 2); thus, we could detect no subgroups that benefited particularly from MR. In these models, the beneficial effect on overall survival of MR
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
6
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Fig. 1. KaplaneMeier curve of overall survival, A) by intervention group; B) by treatment line and intervention group.
compared with observation remained statistically significant regardless of sex, age, ECOG PS, stage, bulky disease, B symptoms and rituximab-containing induction, even after taking into account the low risk characteristics such as a MR after first-line induction or achieving a complete response after induction. As the type of induction chemotherapy was standardised in most of the trials (only in one trial the type of chemotherapy was chosen within the trial by the treating physician), we did not include the type of chemotherapy in a Cox model, but performed subgroup analysis by that variable. The interaction with the inclusion of rituximab in induction was evaluated in the models. The effect of MR was not statistically significantly different with R-CVP/CVP or R-CHOP/CHOP (Table 2). The HR of MR effect for the 233 patients who received fludarabine-based induction was 1.97 with a wide CI. As the current therapy of follicular lymphoma includes rituximab, we performed a subgroup analysis of 1755 patients who received rituximab as part of induction. Although in the model that included both allocation group (MR versus no MR) and rituximabcontaining induction, the effect of MR remained
statistically significant; in this subgroup analysis of rituximab as part of induction, no statistically significant difference was shown (HR 0.85, 95% CI 0.67e1.07). Yet, this effect of MR was not different from its effect in the whole cohort. Repeating all models, including a model that included treatment line and treatment arm demonstrated similar results. Subgroup analysis by the induction line (MR after the first induction; MR after two or more inductions) demonstrated a statistically significant benefit of MR after second induction (HR 0.70, 95% CI 0.54e0.91), while with MR after first induction the HR was 0.92, 95% CI 0.70e1.20 (Fig. 1B).Yet, there was no statistically significant difference in the HRs between the subgroups (test for subgroup differences P Z 0.4). Median overall survival after first line was not yet reached in the MR group and the no MR group. A 5-year overall survival rate in these two groups were 89.2% (95% CI 86.8%e91.5%) and 88% (85.6%e90.4%), respectively. Estimated median overall survival in the MR group after second line was 11.2 years (95% CI 10.9e12.4) and in the observation group 10.3 years (95% CI 8.3e11.5). The 5year overall survival rates in these groups were 76% (95% CI 72%e81%) and 71% (67%e76%), respectively. Following these analyses, we performed a post hoc 2 2 analysis of four subgroups by treatment line and by the inclusion of rituximab in induction. MR following rituximab induction for relapse/refractory lymphoma statistically and significantly improved survival. The point estimates of the effect of MR following induction without rituximab for a frontline therapy and for relapse/refractory lymphoma were also similar, as shown in Table 3. Among patients who received rituximab in the induction phase and were receiving first-line therapy, rituximab did not appear to benefit in terms of overall survival (HR Z 1.049). The difference between 1.049 and 0.70 is not statistically significant, P Z 0.09; thus, one cannot be sure whether this difference is really a chance finding that has come up while looking at many different subgroups or is a true phenomenon. 3.4. Multivariable models A multivariate analysis yielded two main models: model 1 included all variables that were significant on bivariate analysis (sex, age, ECOG PS, bulky disease and FLIPI) as well as study and allocated treatment group. Sex, age, ECOG PS, FLIPI and allocated treatment group remained statistically significant, while bulky disease was not. MR statistically significantly improved overall survival (HR 0.77, 95% CI 0.64e0.93, P Z 0.007). With the addition of treatment line (2nd versus 1st) to the model, the effect of MR did not change. 3.5. Progression-free survival MR improved PFS compared with observation for patients with follicular lymphoma (HR 0.57, 95% CI
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
7
Fig. 2. Forest plot of overall survival of rituximab maintenance (MR) versus no MR based on individual patient data and summary data.
0.51e0.64, p < 0.0001, 2305 patients). Twelve patients were excluded due to missing data (five in the MR group and seven in observation group). We checked for the effect and interaction of treatment group with the following variables: study, sex, age, FLIPI, treatment line, type of chemotherapy and response to induction. Age, treatment line (relapsed/refractory), CVP chemotherapy, low FLIPI and response to induction were prognostic factors for PFS. As with overall survival, there were no statistically significant interactions of any of the variables with the effect of MR on PFS. 3.6. Adverse events We analysed the risk of any adverse event, grade 3 or 4 adverse events and specific adverse events (Table 4). The rate of any adverse event grade 1e4 in MR group was 52% and in the no MR group 40%, the corresponding rates of any grade 3 to 4 were 17% and 23%, respectively
%. The risk of adverse events increased with MR compared with observation (no significant statistical heterogeneity). The most frequent adverse event was infection and occurred in 23.6% of the observation group (4.9% grade 3e4), and in 33.6% of the MR group (7.1% grade 3e4). The risks of infection of any grade and infection of grade 3e4 increased with MR (Table 4). The HR of grade 3e4 infection was 1.27, 95% CI 1.12 to 1.45. Other adverse events were rarely reported. 4. Discussion Based on the data of 2315 patients, this meta-analysis demonstrated an overall survival benefit with MR treatment compared with no MR, expressed as the number needed to treat, 29 patients would be treated with MR to prevent one death (95% CI 17e153) within 5 years [16]. The improvement in overall survival was consistent across all types of patients and unrelated to sex, age, FLIPI risk group, rituximab-containing
Table 3 Analysis of the effect of MR by treatment line and inclusion of rituximab in induction. 2 line
First line
No rituximab induction Induction included rituximab
Events/number
HR, 95% CI
Events/number
HR, 95% CI
73/278
0.71, 0.45e1.14 p Z 0.15 1.049, 0.75e1.46 p Z 0.78
91/282
0.69, 0.46e1.04 p Z 0.075 0.70, 0.51e0.98 p Z 0.035
140/1283
151/471
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
8
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Table 4 Adverse events. Adverse event
Hazard ratio MR versus observation
95% confidence interval
Rate with MR
Rate without MR
Any, grade 1e4 Any, grade 3e4 Infection, grade 1e4 Infection, grade 3e4 Neutropenic fever
1.27 1.31 1.41
1.12, 1.45 1.08, 1.58 1.2, 1.66
52% 23% 33.6%
40% 17% 23.6%
1.48
1.04, 2.11
7.1%
4.9%
0.78
0.43, 1.43
1.8%
2.1%
induction or response to induction. It should be noted that there was no evidence of a different effect with MR with relapsed/refractory lymphoma compared with after first induction, and that in the multivariable analysis model, after adjustment for treatment line, the beneficial effect of MR on survival remained statistically significant. Overall rituximab improves survival when added to maintenance therapy. However, the caveat is that this conclusion may not hold for all circumstances, and based on the analysis of subgroups one circumstance that is particularly uncertain is when the patient has already received rituximab in the induction phase and is receiving first-line therapy. In the other circumstances explored here (either no rituximab in the induction or second line or later therapy), the advantage of rituximab maintenance seems clear. PFS was longer with MR treatment in all subgroups and models. The major toxicity was the increased risk of infection. Based on the outcomes of this meta-analysis, the benefit of MR after frontline induction may still be questionable. One explanation for the apparently smaller benefit in the subgroup analysis after first-line induction could be the longer survival (the median of which was not reached) in the frontline induction group leading to a smaller number of events and a need for larger sample size or a much longer follow-up, to reduce uncertainty. It is possible that with a longer follow-up or a larger sample size, there would be less uncertainty in the results. However, it is unlikely that such a trial would be conducted. It might be possible that certain subgroups within the first-line group may benefit even more from MR. For example, in a pooled analysis of three trials, patients who had a positive FDG-PET at the end of induction had worse PFS and overall survival [17]. These results and the debate regarding the survival benefit after first induction led to an ongoing trial that compares MR for 2 years as the standard treatment after induction to a response-adapted strategy based on end of treatment PET and minimal residual disease (NCT02063685, EudraCT 2012-003170-60). Of note, in our database, MR had a favourable effect on survival regardless of the depth of response.
The main strength of this meta-analysis lies in its very large sample size making it sufficiently powered to show a treatment effect that was not conclusively demonstrated in each of the individual trials. Analysis based on IPD allows inclusion of all randomised patients and exploration of the effect of patient and disease characteristics on the effect of MR. The trials differed with respect to inclusion criteria, type of induction therapy, the MR schedule and also the frequency of assessments. However, the similarity between the different trials, mainly the inclusion of patients with follicular lymphoma after induction therapy (either as first line or relapse) provides a solid basis for this meta-analysis. It should be emphasised that the clinical heterogeneity across the trials did not translate into statistical heterogeneity of the results. The inclusion period of most of the included trials dates back 10 years or more. The survival perspective of today’s patients might be influenced by current and future advances in the treatment of follicular lymphoma such as new antibodies and pathway inhibitors. These newer treatments might impact and interact with the effect of MR. The results of this IPD meta-analysis are consistent with the results of a summary data meta-analysis with the same clinical question in patients with follicular lymphoma [1] and confirm its findings. Six of the seven trials included in the current meta-analysis were also included in the summary data meta-analysis. The previous meta-analysis, as the current one, showed improved overall survival with MR for the whole cohort of follicular lymphoma patients. Subgroup analysis of patients after frontline induction did not support a differential effect from that of patients with relapsed/refractory lymphoma. The advantage of the IPD metaanalysis compared with the summary data metaanalysis is that it enabled us to incorporate patient characteristics and intervention (MR group) into a model to differentiate patients who would benefit from MR. Regarding the effect of MR after single-agent rituximab, we could not conduct subgroup analysis in this IPD analysis, since only one trial [4,6] was included. Yet, the previous summary data meta-analysis pooled data of three trials with patients whose induction therapy consisted of single-agent rituximab with no chemotherapy [1] showed a trend toward improved overall survival with MR compared with no MR (HR Z 0.76, 95% CI Z 0.53e1.01), but again no differential effect from that of the whole cohort. Current guidelines (ESMO 2014) [18] recommend the use of rituximab maintenance for patients with follicular lymphoma based on overall survival benefit in relapsed disease [I, A] and on PFS benefit after first-line induction [I, B]. These recommendations are based on the results of the previous summary data meta-analysis.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
Thus, our recent results are in agreement with current guidelines. All the trials in the current meta-analysis did not use an active control group, and MR was compared only with patients under observation and not to rituximab at progression. This study therefore does not address the question of whether MR is superior to treatment of patients with rituximab at progression. To conclude, based on data from randomised controlled trials, MR improves overall survival when compared with observation and should be prescribed after a successful induction with R-CVP or an R-CHOP regimen. It is still uncertain if that is true in the situation that the patient has already received rituximab in his/ hers first induction. To be elucidated is the reason for the questionable effect for these patients. Optional explanations are an interaction between the type of chemotherapy and MR, a deeper response in the frontline setting or a biologic difference of relapse/refractory follicular lymphoma from a newly diagnosed lymphoma. It is unclear whether rituximab treatment at progression is equivalent to MR. Also, it is unknown whether MR has similar effectiveness after bendamustine-rituximab induction, which nowadays is one of the used induction therapies for follicular lymphoma [18,19], since none of the trials used bendamustine-rituximab as induction. These questions should be answered in future clinical trials.
Conflict of interest statement Dr. Vidal reports Celgene, Novartis, Roche: travel, and accommodations. Dr. Ghielmini has received honoraries as lecturer from the following companies: Roche, Cellgene, Gilead, Jannsen, Abbvie, Mundipharma. Dr. Salles reports grants and personal fees from Roche, during the conduct of the study; personal fees from Jannsen, personal fees from Gilead, personal fees from Celgene, personal fees from Novartis, personal fees from Amgen, personal fees from Servier, personal fees from BMS, personal fees from Merck, outside the submitted work. Dr.Dreger reports participation in advisory board: Roche. Dr. Gomes Silva reports the following: Bristol-Myers Squibb, Celgene, Gilead Sciences, Janssen-Cilag, Takeda: consultancy. Bristol-Myers Squibb, Celgene, Ferrer, JanssenCilag, Takeda: speakers’ bureau; Gilead Sciences: research funding. Celgene, Gilead Sciences, Janssen-Cilag, Roche: travel, accommodations, expenses. Dr. Vitolo reports participation in advisory board: Roche, Celgene, Janssen; and honoraria for lectures: Roche, Celgene, Takeda, Gilead, Mundipharma.
9
Dr. Shpilberg reports a grant from Roche. The other authors have no conflict of interest to declare. Funding The study was partly supported by Roche (KM209450801). Roche was not involved in any part of study design, conduction and interpretation.
Acknowledgements The authors thank the European Organization for Research and Treatment of Cancer for the permission to use the data from EORTC trial 20981 for this research and the European Society for Blood and Marrow Transplantation for using data from the Lym1 trial for this research. We would like to thank Professor Leonard Leibovici for his advice and critical review.
References [1] Vidal L, Gafter-Gvili A, Salles G, et al. Rituximab maintenance for the treatment of patients with follicular lymphoma: an updated systematic review and meta-analysis of randomized trials. J Natl Cancer Inst 2011;103:1799e806. [2] Ardeshna KM, Qian W, Smith P, et al. Rituximab versus a watchand-wait approach in patients with advanced-stage, asymptomatic, non-bulky follicular lymphoma: an open-label randomised phase 3 trial. [Internet]. Lancet Oncol 2014;15:424e35 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm.nih.gov/pubmed/ 24602760. [3] Forstpointner R, Unterhalt M, Dreyling M, et al. Maintenance therapy with rituximab leads to a significant prolongation of response duration after salvage therapy with a combination of rituximab, fludarabine, cyclophosphamide, and mitoxantrone (RFCM) in patients with recurring and refractory follicular [Internet]. Blood 2006;108:4003e8 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm.nih.gov/pubmed/16946304. [4] Ghielmini M, Schmitz S-FH, Cogliatti SB, et al. Prolonged treatment with rituximab in patients with follicular lymphoma significantly increases event-free survival and response duration compared with the standard weekly x 4 schedule. [Internet]. Blood 2004;103:4416e23. Available from: http://www.ncbi.nlm.nih.gov/ pubmed/14976046. [5] Hainsworth JD, Litchy S, Shaffer DW, et al. Maximizing therapeutic benefit of rituximab: maintenance therapy versus retreatment at progression in patients with indolent non-Hodgkin’s lymphoma e a randomized phase II trial of the Minnie Pearl Cancer Research Network. [Internet]. J Clin Oncol 2005;23: 1088e95 [cited 2016 Feb 24] Available from: http://www.ncbi. nlm.nih.gov/pubmed/15657401. [6] Martinelli G, Schmitz SFH, Utiger U, et al. Long-term follow-up of patients with follicular lymphoma receiving single-agent rituximab at two different schedules in trial SAKK 35/98. J Clin Oncol 2010;28:4480e4. [7] Pettengell R, Schmitz N, Gisselbrecht C, et al. Rituximab purging and/or maintenance in patients undergoing autologous transplantation for relapsed follicular lymphoma: a prospective randomized trial from the lymphoma working party of the European group for blood and marrow transplantation. [Internet]. J Clin Oncol 2013;31:1624e30 [cited 2016 Feb 24] Available from: http:// www.ncbi.nlm.nih.gov/pubmed/23547078.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021
10
L. Vidal et al. / European Journal of Cancer xx (2017) 1e10
[8] Salles G, Seymour JF, Offner F, et al. Rituximab maintenance for 2 years in patients with high tumour burden follicular lymphoma responding to rituximab plus chemotherapy (PRIMA): a phase 3, randomised controlled trial. Lancet 2011;377:42e51. [9] Van Oers MHJ, Van Glabbeke M, Giurgea L, et al. Rituximab maintenance treatment of relapsed/resistant follicular non-Hodgkin’s lymphoma: long-term outcome of the EORTC 20981 phase III randomized intergroup study. J Clin Oncol 2010;28:2853e8. [10] Ghielmini M, Vitolo U, Kimby E, et al. ESMO Guidelines consensus conference on malignant lymphoma 2011 part 1: diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL) and chronic lymphocytic leukemia (CLL). [Internet]. Ann Oncol 2013;24:561e76 [cited 2015 Jul 12] Available from: http:// www.ncbi.nlm.nih.gov/pubmed/23175624. [11] Vitolo U, Ladetto M, Boccomini C, et al. Rituximab maintenance compared with observation after brief first-line R-FND chemoimmunotherapy with rituximab consolidation in patients age older than 60 years with advanced follicular lymphoma: a phase III randomized study by the Fondazione Italiana Lin [Internet]. J Clin Oncol 2013;31:3351e9 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm.nih.gov/pubmed/23960180. [12] Witzens-Harig M, Benner A, McClanahan F, et al. Rituximab maintenance improves survival in male patients with diffuse large B-cell lymphoma. Results of the HD2002 prospective multicentre randomized phase III trial. [Internet]. Br J Haematol 2015;171: 710e9 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm. nih.gov/pubmed/26449739. [13] Kahl BS, Hong F, Williams ME, et al. Rituximab extended schedule or re-treatment trial for low-tumor burden follicular lymphoma: eastern cooperative oncology group protocol e4402.
[14]
[15]
[16]
[17]
[18]
[19]
[Internet]. J Clin Oncol 2014;32:3096e102 [cited 2016 Feb 24] Available from: http://www.ncbi.nlm.nih.gov/pubmed/25154829. Higgins JPT, Altman DG. Higgins 2011 Higgins JPT. Version 5.1.0 [updated March 2011]. The Cochrane Collaboration. In: Green S, editor. Cochrane Handbook for systematic reviews of interventions, 2011; 2011. Hochster H, Weller E, Gascoyne RD, et al. Maintenance rituximab after cyclophosphamide, vincristine, and prednisone prolongs progression-free survival in advanced indolent lymphoma: results of the randomized phase III ECOG1496 study. J Clin Oncol 2009;27:1607e14. Altman DG, Andersen PK. Calculating the number needed to treat for trials where the outcome is time to an event. BMJ 1999; 319:1492e5. Trotman J, Luminari S, Boussetta S, et al. Prognostic value of PET-CT after first-line therapy in patients with follicular lymphoma: a pooled analysis of central scan review in three multicentre studies. [Internet]. Lancet Haematol 2014;1:e17e27 [cited 2016 May 31] Available from: http://www.ncbi.nlm.nih.gov/ pubmed/27030064. Newly Diagnosed and Relapsed Follicular Lymphoma: ESMO Clinical Practice GuidelinesjESMO [Internet][cited 2016 Feb 24] Available from: http://www.esmo.org/Guidelines/Haematolo gical-Malignancies/Newly-Diagnosed-and-Relapsed-FollicularLymphoma. Rummel MJ, Niederle N, Maschmeyer G, et al. Bendamustine plus rituximab versus CHOP plus rituximab as first-line treatment for patients with indolent and mantle-cell lymphomas: an openlabel, multicentre, randomised, phase 3 non-inferiority trial. Lancet 2013;381:1203e10.
Please cite this article in press as: Vidal L, et al., Rituximab maintenance improves overall survival of patients with follicular lymphomadIndividual patient data meta-analysis, European Journal of Cancer (2017), http://dx.doi.org/10.1016/j.ejca.2017.01.021