ICON7
A phase 3 trial of bevacizumab in ovarian cancer
Phase 3N = 1,528
- Population
- Women with newly diagnosed ovarian cancer, high-risk early-stage (FIGO I-IIa, grade 3 or clear cell) or stage IIb-IV, ECOG performance status 0-2
- Design
- International phase 3, open-label randomised trial in 11 countries. 1528 women with newly diagnosed ovarian cancer (high-risk early-stage disease, FIGO stage I-IIa grade 3 or clear cell, or stage IIb-IV) were randomly assigned 1:1 (764 per arm) to six 3-weekly cycles of carboplatin (AUC 5 or 6) and paclitaxel 175 mg/m2, or the same chemotherapy plus bevacizumab 7.5 mg/kg every 3 weeks concurrently, continued for up to 12 further 3-weekly cycles. Primary endpoint progression-free survival; the trial was also powered for overall survival. Non-proportional hazards were detected, so restricted mean survival time was used as the effect estimate.
- Primary endpoint
- PFS
- Arms
- Standard chemotherapy (control, n = 764): Carboplatin AUC 5 or 6 plus paclitaxel 175 mg/m2 every 3 weeks for six cycles
- Chemotherapy plus bevacizumab (experimental, n = 764): The same chemotherapy plus bevacizumab 7.5 mg/kg every 3 weeks, continued for up to 12 further cycles
- People
- Perren TJ, ICON7 Investigators
PFSHR 0.81 (95% CI 0.70–0.94)Δ median OS+7.8 months
Show the numbers
| Endpoint | Analysis | Chemotherapy plus bevacizumab | Standard chemotherapy | Effect |
|---|---|---|---|---|
| PFSRandomised women (N=1528); analysis population not stated in the abstract | primaryPerren 2011 | — | — | HR 0.81 (95% CI 0.70–0.94)P=0.004Restricted mean PFS at 36 months was 20.3 months with standard therapy and 21.8 months with bevacizumab (restricted means, not medians). Hazards were non-proportional (P<0.001), with a maximum effect at 12 months, coinciding with the end of planned bevacizumab treatment, and diminishing by 24 months. In updated analyses the restricted mean PFS at 42 months was 22.4 vs 24.1 months (P=0.04 by log-rank test). |
| OSPatients at high risk of progression (subgroup) | exploratoryPerren 2011 | median 36.6 months | median 28.8 months | Interim overall survival at the time of the 2011 report; subgroup of patients at high risk for progression (30% of those randomised per the abstract). Restricted mean PFS at 42 months in this subgroup was 14.5 vs 18.1 months. This 2011 'high risk for progression' subgroup (30% of those randomised, per the abstract) is not the same analysis as the 502-patient (33%) poor-prognosis subgroup of the 2015 overall survival paper, whose risk groups were refined slightly before database lock (stage IV, inoperable stage III or suboptimally debulked stage III; PMC4648090). |
| Endpoint | Analysis | Chemotherapy plus bevacizumab | Standard chemotherapy | Effect |
|---|---|---|---|---|
| PFSRandomised women (N=1528); analysis population not stated in the abstract | primaryPerren 2011 | — | — | HR 0.81 (95% CI 0.70–0.94)P=0.004Restricted mean PFS at 36 months was 20.3 months with standard therapy and 21.8 months with bevacizumab (restricted means, not medians). Hazards were non-proportional (P<0.001), with a maximum effect at 12 months, coinciding with the end of planned bevacizumab treatment, and diminishing by 24 months. In updated analyses the restricted mean PFS at 42 months was 22.4 vs 24.1 months (P=0.04 by log-rank test). |
| OSPatients at high risk of progression (subgroup) | exploratoryPerren 2011 | median 36.6 months | median 28.8 months | Interim overall survival at the time of the 2011 report; subgroup of patients at high risk for progression (30% of those randomised per the abstract). Restricted mean PFS at 42 months in this subgroup was 14.5 vs 18.1 months. This 2011 'high risk for progression' subgroup (30% of those randomised, per the abstract) is not the same analysis as the 502-patient (33%) poor-prognosis subgroup of the 2015 overall survival paper, whose risk groups were refined slightly before database lock (stage IV, inoperable stage III or suboptimally debulked stage III; PMC4648090). |
Data table
| Endpoint | HR | 95% CI |
|---|---|---|
| PFS | 0.81 | 0.7–0.94 |
- Chemotherapy plus bevacizumab (experimental)
- Standard chemotherapy (control)
Median-only curves assume a constant hazard (exponential) and fit plateaus, as seen with some immunotherapies, poorly.
Data table (modelled values)
| Months | Chemotherapy plus bevacizumab | Standard chemotherapy |
|---|---|---|
| 0 | 100.0% | 100.0% |
| 23 | 64.7% | 57.5% |
| 46 | 41.8% | 33.1% |
| 69 | 27.1% | 19.0% |
| 92 | 17.5% | 10.9% |