Table 1.

Characteristics and PFS data of included randomized controlled trials.

Trial name	Stage	Grade	Trial design	Maintenance	Data source	Arms	Patient number	PFS HR (95% CI)
FOLL05 [7, 8]	II–IV	1, 2, 3a	R-CHOP vs R-CVP vs R-FM Primary endpoint: time to treatment failure	Not allowed	Federico et al. [7]	R-CHOP	165	0.73 (0.54–0.98)a p = 0.037
						R-CVP	168
PRIMA [9, 10]	III–IV	1, 2, 3a	Rituximab maintenance vs observation after responding to R-CHOP, R-CVP or R-FM (R-FM accounted for <3% of patients) Primary endpoint: PFS from randomization (6 months after induction)	Randomized	Bachy et al. [10]	Rituximab maintenance	505	0.61 (0.52–0.73)e p < 0.0001
						Observation	513
						R-CHOP-R	382	0.57 (0.47–0.70)
						R-CHOP	386
						R-CVP-R	109	0.75 (0.53–1.07)
						R-CVP	113
Stil NHL1 [11]	III–IV	1, 2	R-Benda vs R-CHOP for iNHL and MCL Primary endpoint: PFS	Not allowed	Rummel et al. [11]	R-Benda	139	0.61 (0.42–0.87)b p = 0.0072
						R-CHOP	140
BRIGHT [12–14]	II–IV	1, 2	R-Benda vs R-CHOP/R-CVP for iNHL and MCL Primary endpoint: CR rate (noninferiority design)	Allowed (at the discretion of the investigator)	Flinn et al. [14]c	R-Benda(-R)	187	0.70 (0.49–1.01) p = 0.0582
						R-CHOP(-R)/ R-CVP(-R)	186
					Kahl et al. [13]d	R-Benda-R	81	0.50 (0.26–0.94) p = 0.0295
						R-Benda	63
						R-CHOP-R/ R-CVP-R	83	0.66 (0.38–1.16) p = 0.1443
						R-CHOP/ R-CVP	61
Stil NHL7 (MAINTAIN) [15]	II–IV	1, 2	Rituximab maintenance (R4) vs observation (R) after responding to R-Benda and 2 years of rituximab maintenance Primary endpoint: PFS from randomization	Randomized	Rummel et al. [15]	R-Benda-R4	178	0.73 (0.44–1.21) p = 0.1125
						R-Benda-R	172
			Rituximab maintenance (up to 2 years) vs observation after responding to R-Benda (R4 patients censored after 2 years of maintenance) Primary endpoint: PFS	Indirect comparison combining NHL1 and NHL7 (a secondary endpoint of NHL7)	Rummel et al. [15]	R-Benda-R	595	0.68 (0.47–0.87) p = 0.0074
						R-Benda	139
GALLIUM [16, 17]	III–IV	1, 2, 3a	G-Chemo-G vs R-Chemo-R Primary endpoint: PFS	Required (if responding to induction; >90% received maintenance)	Hiddemann et al. [17]	G-Chemo-G	601	0.68 (0.54–0.87)e p = 0.0016
						R-Chemo-R	601
						G-Benda-G	345	0.63 (0.46–0.88) p = 0.0062
						R-Benda-R	341
						G-CHOP-G	196	0.72 (0.48–1.10) p = 0.13
						R-CHOP-R	203
						G-CVP-G	60	0.79 (0.42–1.47) p = 0.46
						R-CVP-R	57
RELEVANCE [18]	II–IV	1, 2, 3a	R-Len-R vs R-Chemo-R Primary endpoint: CR/CRu rate (at 120 weeks) and PFS	Required (if responding to induction; >85% received maintenance)	Morschhauser et al. [18]	R-Len-R	513	1.10 (0.85–1.43)e p = 0.48
						R-Chemo-R	517
						R-Len-R	100	1.75 (0.88–3.49)
						R-Benda-R	117
						R-Len-R	383	1.02 (0.76–1.37)
						R-CHOP-R	373
						R-Len-R	100	1.75 (0.88–3.49)
						R-CVP-R	27

PFS progression-free survival, HR hazard ratio, CI confidence interval, iNHL indolent non-Hodgkin lymphoma, MCL mantle cell lymphoma, FL follicular lymphoma, LPL lymphoplasmacytic lymphoma, MZL marginal zone lymphoma, CR complete response, CRu unconfirmed complete response.

^aPFS data adjusted by Follicular Lymphoma International Prognostic Index 2 (FLIPI2).

^bPFS data for FL patients only.

^cData for the entire iNHL cohort (predominantly FL). R-Benda arm, FL n = 154, other iNHL (LPL, MZL) n = 33; R-CHOP/R-CVP arm, FL n = 160, other iNHL n = 26.

^dData for FL patients only. Post hoc analysis of (non-randomized) rituximab maintenance vs observation in patients who achieved an objective response.

^eThese data were not used in the network meta-analysis.