The value of progression-free survival at three years as a primary endpoint for studies on radiotherapy in patients with locally advanced cervical cancer: Individual patient data from Chinese National Cancer Center and validation from 27 global randomized trials.

Abstract

5534 Background: A traditional endpoint for locally advanced cervical cancer (LACC) clinical trials is overall survival with five years of follow-up (5-year OS). The use of a shorter-term endpoint could significantly speed the translation of research findings into practice. The primary hypothesis was that PFS with three years of follow-up (PFS36) is an appropriate primary endpoint to replace 5-year OS. Methods: The primary hypothesis was developed from our individual data, was further investigated using phase III randomized controlled trials (RCTs), and then externally validated by phase II trials and retrospective studies up to 2022. Correlation analysis at the treatment-arm level was performed between 2-, 3-, 4-, and 5-year PFS rates and 5-year OS. Correlation analysis was performed using the Pearson correlation coefficient r in weighted linear regression, with weight equal to patient size. The MEDLINE, Embase, and PubMed databases, together with the Cochrane Central Register of Controlled Trials, were searched from January 1, 1999, to February 2, 2023. Articles eligible for inclusion contained complete survival data. Results: A total of 613 patients with histologically confirmed, FIGO 2009 stage IB–IVA cervical cancer who underwent radiotherapy at our institute from January 2010 to December 2013 were eligible. Individual patient data were pooled to explore the correlation between PFS and the OS trend. The recurrence rates for years 1 through 5 were 12.9%, 7.3%, 3%, 2.3%, and 1.8%, respectively. The median recurrence time was 13 months and the median time from recurrence to death was 12.2 months. Within all the recurrence, 47.3% of recurrences occurred during the first year, 71.4% in the first two years, and 85% in the first three years. Patients who did not achieve PFS36 had a 5-year OS rate of 30.3%. In contrast, a 5-year OS rate of 98.2% was observed in patients who achieved PFS36. Further data were extracted from 27 randomized phase III trials on LACC. The trials included 57 arms, with a pooled sample size of 7,692 patients. In trial-level surrogacy, PFS36 (r2, 0.778) was associated with 5-year OS. Sensitivity analysis demonstrated reasonable overall consistency. The correlation between PFS36 and OS was externally validated using independent phase II trials and retrospective data. In total, 23 studies representing 5,174 patients were included. PFS (r2, 0.719) was found to be associated with 5-year OS. Conclusions: A significant correlation was found between PFS36 and 5-year OS in clinical trials on patients with locally advanced cervical cancer. This correlation was found both within patients and across trials. These results suggest that PFS36 is an appropriate endpoint for LACC trials of radiotherapy-based regimens.