Background
Phase III RCT (GOG 218). 1,873 patients with newly diagnosed stage III-IV epithelial ovarian, fallopian tube, or primary peritoneal cancer after primary debulking surgery. Three-arm design: (1) carboplatin AUC 6 + paclitaxel 175 mg/m² × 6 cycles + placebo throughout; (2) carboplatin + paclitaxel × 6 cycles + bevacizumab 15 mg/kg concurrent (during chemo) + placebo maintenance; (3) carboplatin + paclitaxel × 6 cycles + bevacizumab concurrent + bevacizumab maintenance (up to 22 cycles total). Primary comparison: arms 1 vs 3 (bevacizumab throughout vs placebo throughout).
Interventions and follow up
Arm A: Carboplatin + paclitaxel (6 cycles) + placebo throughout
Arm B: Carboplatin + paclitaxel + bevacizumab concurrent only, then placebo maintenance
Arm C: Carboplatin + paclitaxel + bevacizumab concurrent + bevacizumab maintenance
Primary endpoint: PFS (arm 1 vs arm 3)
mFollow up: 17.4 months (primary); 102.9 months (updated)
Arm B: Carboplatin + paclitaxel + bevacizumab concurrent only, then placebo maintenance
Arm C: Carboplatin + paclitaxel + bevacizumab concurrent + bevacizumab maintenance
Primary endpoint: PFS (arm 1 vs arm 3)
mFollow up: 17.4 months (primary); 102.9 months (updated)
Results
PFS — bev throughout vs placebo: 14.1 vs 10.3 months, HR 0.72 (95% CI 0.63–0.82), P<.001
OS (primary): 39.7 vs 38.7 months, HR 0.89 (95% CI 0.78–1.01), P=.07 — not significant
OS (updated analysis, Tewari JCO 2019): 43.8 vs 40.6 months, HR 0.82 (95% CI 0.74–0.91) — no significant OS benefit
PFS with concurrent-only bev (arm 2 vs arm 1): 11.2 vs 10.3 months — minimal benefit
OS (primary): 39.7 vs 38.7 months, HR 0.89 (95% CI 0.78–1.01), P=.07 — not significant
OS (updated analysis, Tewari JCO 2019): 43.8 vs 40.6 months, HR 0.82 (95% CI 0.74–0.91) — no significant OS benefit
PFS with concurrent-only bev (arm 2 vs arm 1): 11.2 vs 10.3 months — minimal benefit
Adverse events
Main adverse events: Grade ≥3 hypertension: 22.9% (bev throughout) vs 0.5% (placebo). GI perforation: 2.6% (bev throughout) vs 0.7% (placebo). Wound healing complications: higher in bev arms. Thromboembolic events: 7.2% (bev throughout) vs 4.5% (placebo). Fistula: 1.6% vs 0.2%.
Conclusions
Bevacizumab added throughout first-line treatment (concurrent + maintenance) significantly prolonged PFS in newly diagnosed advanced ovarian cancer, but did not demonstrate a statistically significant OS benefit. Concurrent-only bevacizumab provided minimal benefit.
Key Limitations
Key Limitations: No OS benefit in any subgroup or updated analysis despite multiple years of follow-up — the PFS improvement does not translate to OS, raising questions about clinical meaningfulness. Bevacizumab adds substantial cost ($10,000+/month) and toxicity (hypertension, bowel perforation) without OS benefit. In modern practice, PARP inhibitor maintenance (for BRCA-mutated/HRd patients) is strongly preferred over bevacizumab alone. ESMO-MCBS score: 2 (based on PFS without OS benefit).
Clinical Context
GOG 218 established bevacizumab as the first biologic agent with first-line activity in ovarian cancer, but its role is increasingly limited to BRCA-wild-type/HRp patients where PARP inhibitors have minimal benefit, or as a backbone for PAOLA-1 combinations. NCCN lists bevacizumab + chemotherapy as a category 1 option for advanced ovarian cancer but notes the lack of OS benefit.