Extrapolation Methods Could Help Inform Ovarian Cancer Treatment, Management

The SOLO1 phase 3 clinical trial demonstrated that maintenance olaparib provided a substantial progression-free survival (PFS) benefit among patients with newly diagnosed advanced ovarian cancer who had BRCA1 or BRCA2 mutations, with a 70% lower risk of disease progression or death with olaparib compared with placebo. However, understanding the long-term survival rate of patients with newly diagnosed advanced ovarian cancer is critical to the clinical and economic appraisal of new treatments, disease management, and patient counseling, according to Dominic Muston, MSc, Director of Economic and Data Sciences at Merck.

Since overall survival (OS) data from SOLO1 is still immature, Mr Muston and his team sought to explore the external validity of 4 different methods used to extrapolate OS data for patients in the placebo arm of the SOLO1 trial. He presented the findings at the Society of Gynecologic Oncology (SGO) Annual Meeting on Women’s Cancer.

Four methods were evaluated by comparing each method’s 10-year OS projections for the placebo arm of SOLO1 to observed OS data for patients with stage III or IV ovarian cancer in the Surveillance, Epidemiology, and End Results (SEER) cancer registry (BRCA1/2 status was not available in the registry).

All methods used observed data for the first 24 months, which was the duration of maintenance treatment in the SOLO1 study. After 24 months, method 1 used a Weibull parametric model (best-fitting distribution) to extrapolate OS data from the placebo arm; method 2 used the same distribution from the olaparib arm data, but with an assumed treatment effect for placebo; method 3 projected OS by applying an acceleration factor to PFS data; and method 4 instead applied a constant hazard ratio to the hazard of disease progression. The acceleration factor and the hazard ratio used in methods 3 and 4 were estimated from SEER-Medicare data.

The investigators observed largely divergent results among the study methods. At year 10, the estimated OS rate for patients in the placebo arm was 60.3% for method 1, 18.4% for method 2, 45.0% for method 3, and 44.0% for method 4. The estimated projection from method 2 was the closest to the reported 10-year OS from SEER (15.0%).

“Compared to real-world data, method 2 provided the most externally valid projection of long-term overall survival in patients with advanced ovarian cancer, and this may inform future modeling for cost-effectiveness, service demand, and patient counseling,” reported Mr Muston.

The investigators maintain that future validation against extended follow-up of trial populations, with and without olaparib or subsequent PARP inhibitor treatment, will provide additional evidence to inform long-term survival for patients with ovarian cancer and BRCA1 or BRCA2 mutations.


  • SGO 2020 Annual Meeting on Women’s Cancer. Abstract 286.

Related Articles

Subscribe Today!

To sign up for our free e-Newsletter or print publications, please enter your contact information below.

I'd like to receive:

OPM e-Newsletter
OPM print publication