Objectives Prognosis in ovarian cancer is poor. Previous conditional survival (CS) studies have found outcomes improved based on each year survived since diagnosis. Our objective was to estimate CS and compare to previous studies (≤ 2010).

Methods The Tempus EMR dataset includes patients from National Cancer Institute designated centers and community oncology centers across the U.S. This study included adult women with a primary diagnosis of ovarian, fallopian tube, or peritoneal cancer from 1982–2018; women treated with a poly-ADP ribose polymerase (PARP) inhibitor were excluded due to low numbers (final n=3,370). We calculated relative 5-year CS as the Kaplan-Meier probability of surviving an additional 5 years (from diagnosis), given no OS event in the previous x years, adjusted for expected mortality by age and sex from U.S. life tables.

Results Median age was 60 years and 69% of patients were Caucasian. Relative 5-year survival was 41.3% overall, and decreased with increasing stage (71.2% for stage I vs. 26.0% for stage IV). CS did not improve with previous survival.

Conclusions Overall, survival rates were lower than those previously reported from Surveillance, Epidemiology, and End Results (SEER) data, and prognosis in ovarian cancer remains poor. Differences in estimates may reflect differing demographics, clinical characteristics, and treatment patterns. In contrast to prior studies, CS did not improve with time already survived. Our previous research in this sample suggests that time survived progression-free, specifically, may improve prognosis. Further research in diverse datasets is needed.