Correlation between real-world progression-free survival and time to next treatment and death in advanced ovarian cancer patients receiving non-platinum therapy in second or later lines.

Abstract

e17578 Background: Advanced ovarian cancer (aOC) is typically treated with platinum-based first-line therapy, but in second or later lines (2L+) some patients receive non-platinum regimens due to platinum resistance, toxicity, comorbidities, or patient/clinician preference. Although progression-free survival (PFS) is a commonly used efficacy endpoint in clinical trials, its assessment in real-world data is often limited by incomplete or delayed documentation of progression, prompting use of time to next treatment or death (TTNTD) as a pragmatic alternative. Because treatment changes can be driven by factors beyond progression, it is important to evaluate the correlation between TTNTD and real-world PFS (rwPFS) among aOC patients receiving non-platinum therapies in 2L+. Methods: This retrospective observational study used the Flatiron Health electronic health record–derived, de-identified database. Adults aged ≥18 years with stage III–IV aOC diagnosed on or after May 1, 2020, who received platinum-based 1L therapy, initiated second-line (2L) treatment and received a non-platinum regimen in 2L+ were included, with follow-up until Feb 28, 2025. TTNTD and rwPFS were both measured from 2L initiation and estimated using Kaplan–Meier methods. Patient-level association between TTNTD and rwPFS was assessed using inverse-probability-of-censoring weighted Kendall tau and Spearman correlation coefficients. Results: Among 306 patients included in the analysis, mean age at diagnosis was 67.8 years; 50.7% had stage III and 48.4% stage IV disease; 39.9% were platinum sensitive (PSOC), 60.1% were platinum resistant (PROC). TTNTD events in 286 (94.7%) and rwPFS events occurred in 273 patients (90.4%). Median TTNTD was 5.9 months and median rwPFS was 6.0 months, with no significant difference (log-rank p>0.05). Kendall’s tau between TTNTD and rwPFS was 0.62 (95% confidence interval [CI] 0.55–0.68) and Spearman correlation was 0.76 (95% CI 0.66–0.83), both were higher in PROC (0.65 and 0.83) than in PSOC (0.59 and 0.69). Among 270 patients with both events, TTNTD event occurred before rwPFS event in 42 (15.6%), on the same day in 30 (11.1%), and after rwPFS event in 198 (73.3%). The median difference between TTNTD event and rwPFS event was 14.5 days. Conclusions: TTNTD and rwPFS from 2L therapy initiation appeared as moderately correlated in the aOC patients receiving non-platinum regimens in 2L+ with relatively short timing offset, suggesting that TTNTD may serve as a practical alternative when progression dates are incomplete or absent in real-world data. Nonetheless, sensitivity analyses and context-specific interpretation are warranted when using TTNTD as a proxy for rwPFS.