← Back

CheckMate 143

Effect of Nivolumab vs Bevacizumab in Patients With Recurrent Glioblastoma: The CheckMate 143 Phase 3 Randomized Clinical Trial

Share Share Copied! Compare

Year of Publication: 2020

Authors: Reardon DA, Brandes AA, Omuro A, ..., Weller M

Journal: JAMA Oncology

Citation: JAMA Oncology 2020;6(7):1003-1010

Link: https://doi.org/10.1001/jamaoncol.2020.1024

PDF: https://jamanetwork.com/journals/jamaonc...article/2766213

Clinical Question

Does single-agent nivolumab improve overall survival compared with bevacizumab in patients with first-recurrence glioblastoma after standard RT + temozolomide?

Bottom Line

In first-recurrence glioblastoma after standard chemoradiation, single-agent nivolumab did NOT improve overall survival compared with bevacizumab (median 9.8 vs 10.0 months; HR 1.04; p=0.76). This phase 3 negative result tempered enthusiasm for immune checkpoint blockade monotherapy in glioblastoma and redirected the field toward combination approaches, neoadjuvant dosing, and biomarker-selected strategies.

        Major Points
        Phase 3 open-label multicenter randomized controlled trial
369 patients with first-recurrence glioblastoma after standard RT + temozolomide, enrolled Sept 2014 to May 2015
Randomized 1:1 to nivolumab 3 mg/kg IV every 2 weeks or bevacizumab 10 mg/kg IV every 2 weeks until progression or toxicity
Median follow-up: 9.5 months at data cutoff January 2017
Conducted at 57 sites in multiple countries
Primary endpoint: overall survival
Median OS: 9.8 months (nivolumab) vs 10.0 months (bevacizumab); HR 1.04 (95% CI 0.83-1.30); p=0.76 — NOT significant
12-month OS: 42% (both arms)
Progression-free survival: worse with nivolumab (1.5 vs 3.5 months); HR 1.97
Objective response rate: 7.8% nivolumab vs 23.1% bevacizumab — bevacizumab response rate dominated by radiographic changes from anti-VEGF effect
Median duration of response: 11.1 months (nivolumab) vs 5.3 months (bevacizumab) — nivolumab responses durable when they occurred
MGMT methylated: 23% in each arm; unmethylated: ~33%; not reported in ~44%
Grade 3-4 TRAEs: 18.1% (nivolumab) vs 15.1% (bevacizumab)
Immune-related AEs with nivolumab: colitis, pneumonitis, hepatitis — manageable with standard protocols
Bevacizumab AEs: hypertension, bleeding, wound-healing impairment
Exploratory MGMT methylated without baseline steroid use subgroup suggested possible benefit with nivolumab — hypothesis-generating
First phase 3 of PD-1 monotherapy in GBM; negative result dampened enthusiasm for single-agent checkpoint blockade in GBM
Redirected field toward biomarker selection, combinations (anti-PD-1 + anti-VEGF, + RT, + vaccines), and neoadjuvant approaches (Cloughesy 2019 showed neoadjuvant pembrolizumab signal)

      

Design

Study Type: Phase 3 open-label multicenter randomized controlled trial

Randomization: 1

Blinding: Open-label

Enrollment Period: September 2014 - May 2015

Follow-up Duration: Median 9.5 months at data cutoff January 2017

Centers: 57

Sample Size: 369

Analyzed: 369

Analysis: Intention-to-treat; stratified Cox proportional hazards

Inclusion Criteria

Adults ≥18 years
Histologically confirmed glioblastoma (WHO grade IV)
First recurrence after standard RT + concomitant temozolomide
KPS ≥70
Measurable disease by RANO
Adequate organ function

Exclusion Criteria

More than 1 prior recurrence
Prior anti-VEGF or immunotherapy
Active autoimmune disease
Immunosuppression
Active CNS metastases from other cancers
Known NF1 or other hereditary tumor syndrome

Baseline Characteristics

Characteristic	Control	Active
N	185	184
Median age	~55	~55
KPS ≥90	~35%	~35%
MGMT methylated	22.7%	23.4%
MGMT unmethylated	36.2%	32.1%
Baseline steroids	common	common

Arms

Field	Control	Nivolumab
N	185	184
Intervention	Bevacizumab 10 mg/kg IV every 2 weeks until progression or toxicity	Nivolumab 3 mg/kg IV every 2 weeks until progression or toxicity
Duration	Until progression	Until progression

Outcomes

Outcome	Type	Control	Intervention	HR / OR / RR	P-value
Overall survival	Primary	Median 10.0 months (bevacizumab)	Median 9.8 months (nivolumab)	1.04	p=0.76 (not significant)
12-month overall survival	Secondary	42%	42%		Similar
Progression-free survival (median)	Secondary	3.5 mo (bevacizumab)	1.5 mo (nivolumab)	HR 1.97 favoring bevacizumab	Worse PFS with nivolumab
Objective response rate (RANO)	Secondary	23.1%	7.8%		Higher response rate with bevacizumab (mostly radiographic anti-VEGF effect)
Median duration of response	Secondary	5.3 months	11.1 months		Longer durability with nivolumab when responses occurred
Exploratory: MGMT methylated without baseline steroids	Secondary	Baseline survival	Possible survival benefit with nivolumab (hypothesis-generating)		Not statistically significant overall
Grade 3-4 TRAE	Adverse	15.1%	18.1%		Slightly higher with nivolumab
Hypertension (bevacizumab-related)	Adverse	Common	Less common		Bevacizumab-specific
Bleeding (bevacizumab-related)	Adverse	Higher rate	Lower rate		Bevacizumab-specific
Colitis / diarrhea (immune-related)	Adverse	Less common	More common with nivolumab		Nivolumab-specific
Pneumonitis (immune-related)	Adverse	Rare	Uncommon with nivolumab		Nivolumab-specific
Hepatitis (immune-related)	Adverse	Rare	Uncommon with nivolumab		Nivolumab-specific
Fatigue	Adverse	Common	Common		Similar
Treatment-related deaths	Adverse	None reported	None attributable to nivolumab		No fatal toxicity imbalance

Subgroup Analysis

Exploratory subgroup analysis suggested possible survival benefit with nivolumab in MGMT methylated patients without baseline corticosteroid use — hypothesis-generating but not confirmed. Baseline steroids appeared to blunt immunotherapy response. These observations suggested that patient selection (MGMT methylation, steroid use, tumor immune phenotype) may be critical for GBM immunotherapy benefit, but have not been validated in a prospective biomarker-stratified trial.

Criticisms

Bevacizumab has FDA accelerated approval but no proven OS benefit in rGBM — choice of active comparator without survival proof
Open-label design could affect crossover and post-progression management
Baseline corticosteroid use likely blunted nivolumab efficacy
MGMT methylation status not reported in ~44% of patients
No biomarker stratification or PD-L1 expression integration
Single-agent approach may be insufficient — combination and neoadjuvant strategies explored in subsequent trials
Short median follow-up (9.5 months) and immature long-term data

Funding

Bristol-Myers Squibb (manufacturer of nivolumab/Opdivo)

Based on: CheckMate 143 (JAMA Oncology, 2020)

Authors: Reardon DA, Brandes AA, Omuro A, ..., Weller M

Citation: JAMA Oncology 2020;6(7):1003-1010

Content summarized and formatted by NeuroTrials.ai.

CheckMate 143

(2020)

Objective

Nivolumab (anti-PD-1) vs bevacizumab (anti-VEGF) monotherapy — to evaluate whether single-agent immune checkpoint blockade improves overall survival in recurrent glioblastoma.

Study Summary

• Single-agent nivolumab did NOT improve overall survival in first-recurrence glioblastoma compared with bevacizumab.
• Median overall survival: 9.8 months (nivolumab) vs 10.0 months (bevacizumab); HR 1.04 (95% CI 0.83-1.30; p=0.76).
• Objective response rate was lower with nivolumab (7.8%) than bevacizumab (23.1%), but responses were longer-lasting with nivolumab.
• 12-month overall survival was similar: 42% (nivolumab) vs 42% (bevacizumab).
• Immune-related AEs were common with nivolumab; bevacizumab caused hypertension and bleeding.
• Tempered enthusiasm for PD-1 monotherapy in GBM; redirected research toward biomarker selection, combinations, and neoadjuvant approaches.

Intervention

Nivolumab 3 mg/kg IV every 2 weeks vs bevacizumab 10 mg/kg IV every 2 weeks until progression, unacceptable toxicity, or death.

Inclusion Criteria

Adults with histologically confirmed glioblastoma, first recurrence after standard RT + temozolomide, KPS ≥70, measurable disease.

Study Design

Arms: Nivolumab vs Bevacizumab (1:1)

Patients per Arm: Nivolumab 184; Bevacizumab 185 (N=369)

Outcome

• Primary: Median overall survival — 9.8 months (nivolumab) vs 10.0 months (bevacizumab); HR 1.04 (95% CI 0.83-1.30); p=0.76; NOT significant
• 12-month OS: 42% vs 42%
• Progression-free survival: shorter with nivolumab (1.5 vs 3.5 months); HR 1.97
• Objective response rate: 7.8% (nivolumab) vs 23.1% (bevacizumab)
• Median duration of response: 11.1 vs 5.3 months (longer with nivolumab when responses occurred)
• Grade 3-4 TRAEs: 18.1% (nivolumab) vs 15.1% (bevacizumab)
• MGMT methylation exploratory subgroup signal: possible survival benefit with nivolumab

Clinical Question

Does single-agent nivolumab improve overall survival vs bevacizumab in first-recurrence glioblastoma?

Bottom Line

Major Points

Phase 3 open-label multicenter randomized controlled trial
369 patients with first-recurrence glioblastoma after standard RT + temozolomide, enrolled Sept 2014 to May 2015
Randomized 1:1 to nivolumab 3 mg/kg IV every 2 weeks or bevacizumab 10 mg/kg IV every 2 weeks until progression or toxicity
Median follow-up: 9.5 months at data cutoff January 2017
Conducted at 57 sites in multiple countries
Primary endpoint: overall survival
Median OS: 9.8 months (nivolumab) vs 10.0 months (bevacizumab); HR 1.04 (95% CI 0.83-1.30); p=0.76 — NOT significant
12-month OS: 42% (both arms)
Progression-free survival: worse with nivolumab (1.5 vs 3.5 months); HR 1.97
Objective response rate: 7.8% nivolumab vs 23.1% bevacizumab — bevacizumab response rate dominated by radiographic changes from anti-VEGF effect
Median duration of response: 11.1 months (nivolumab) vs 5.3 months (bevacizumab) — nivolumab responses durable when they occurred
MGMT methylated: 23% in each arm; unmethylated: ~33%; not reported in ~44%
Grade 3-4 TRAEs: 18.1% (nivolumab) vs 15.1% (bevacizumab)
Immune-related AEs with nivolumab: colitis, pneumonitis, hepatitis — manageable with standard protocols
Bevacizumab AEs: hypertension, bleeding, wound-healing impairment
Exploratory MGMT methylated without baseline steroid use subgroup suggested possible benefit with nivolumab — hypothesis-generating
First phase 3 of PD-1 monotherapy in GBM; negative result dampened enthusiasm for single-agent checkpoint blockade in GBM
Redirected field toward biomarker selection, combinations (anti-PD-1 + anti-VEGF, + RT, + vaccines), and neoadjuvant approaches (Cloughesy 2019 showed neoadjuvant pembrolizumab signal)

Study Design

Study Type: Phase 3 open-label multicenter randomized controlled trial
Randomization: Yes
Blinding: Open-label
Sample Size: 369
Follow-up: Median 9.5 months at data cutoff January 2017
Centers: 57

Primary Outcome

Definition: Overall survival

Control	Intervention	HR/OR	P-value
Median 10.0 months (bevacizumab)	Median 9.8 months (nivolumab)	1.04 (0.83-1.30)	p=0.76 (not significant)

Limitations & Criticisms

Bevacizumab has FDA accelerated approval but no proven OS benefit in rGBM — choice of active comparator without survival proof
Open-label design could affect crossover and post-progression management
Baseline corticosteroid use likely blunted nivolumab efficacy
MGMT methylation status not reported in ~44% of patients
No biomarker stratification or PD-L1 expression integration
Single-agent approach may be insufficient — combination and neoadjuvant strategies explored in subsequent trials
Short median follow-up (9.5 months) and immature long-term data

Citation

JAMA Oncology 2020;6(7):1003-1010

View Original Publication →

CheckMate 143

Effect of Nivolumab vs Bevacizumab in Patients With Recurrent Glioblastoma: The CheckMate 143 Phase 3 Randomized Clinical Trial

Clinical Question

Bottom Line

Major Points

Design

Inclusion Criteria

Exclusion Criteria

Baseline Characteristics

Arms

Outcomes

Subgroup Analysis

Criticisms

Funding

Ask about CheckMate 143