Abstract 168P
Background
Although disease free survival (DFS) is not a validated surrogate endpoint for overall survival (OS) in most breast cancer subtypes, regulatory approvals based on DFS are common. The objective of this study was to assess the changes in the type of events contributing to DFS over time.
Methods
We searched MEDLINE to identify adjuvant studies in breast cancer published between 2000 and 2020 where the primary or secondary endpoint was DFS, and where the primary event contributing to DFS was reported clearly. We performed univariable and multivariable linear regression to examine distant DFS events and the breast cancer (BC) related DFS events as a proportion of total sample size, and their association with estrogen receptor (ER) expression, menopausal status, nodal status, follow-up (FU) duration and start year. Data were reported quantitatively using the Burnand criteria.
Results
We identified 1,204 articles, and our final analysis included 84 unique studies and 88 cohorts. There were 212,191 patients evaluable for DFS and 41,604 DFS events, of which 23,205 were distant DFS events. The distant DFS event rate/100 patients randomised/year has declined modestly over time (Coef -0.027, p=0.096). The average number of BC related events/100 patients randomised/year has also declined over time (Coef -0.06 p=0.009), while the average number of non-BC related events has remained relatively constant (Coef -0.01, p=0.11). In multivariable analysis, start year, ER expression and longer median FU time were statistically and quantitatively associated with distant DFS events as a percentage of the total population randomised. ER expression, start year and median FU time were also statistically and quantitatively associated with BC-related DFS events as a proportion of all patients randomised.
Conclusions
Distant and BC related events as a proportion of all patients randomised have declined over time, after adjusting for baseline trial variables. If DFS events are increasingly driven by non-distant recurrence or non-BC related events, there will likely be limited surrogacy between DFS and OS.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
A. Desnoyers: Financial Interests, Personal, Invited Speaker: Pfizer; Financial Interests, Personal, Invited Speaker: Viatris. E. Amir: Financial Interests, Personal, Invited Speaker: Novartis; Financial Interests, Personal, Expert Testimony: Genetech/Roche; Financial Interests, Personal, Advisory Role: Sandoz; Financial Interests, Personal, Advisory Role: Exact Sciences. All other authors have declared no conflicts of interest.