The impact of analysis methodology details on invasive breast cancer-free survival in randomized clinical trials
Year: 2025
Type: article
Abstract: BACKGROUND: Standardized endpoint definitions are crucial for the correct interpretation and reporting of clinical trials. In the field of adjuvant breast cancer clinical trials, the Standardized Definitions for Efficacy End Points (STEEP) criteria were introduced in 2007. In 2021, the STEEP criteria were re-assessed, and a new endpoint was proposed: invasive breast cancer-free survival (IBCFS). However, this recent endpoint also introduces complexity, and detailed considerations of the consequences must be made. MATERIALS AND METHODS: IBCFS does not include second primary non-breast cancers (SPNBCs). This complicates the statistical analysis as it is no longer perfectly straightforward. We investigate the multiple analysis options and their crucial implications from a theoretical perspective as well as from a practical point of view, using actual trial data from four large adjuvant breast cancer trials. RESULTS: SPNBCs can either be ignored, censored, or treated as competing events. Importantly, all three approaches address different clinical questions: when SPNBCs are ignored, the 'total' treatment effect is estimated. By censoring SPNBCs, a hypothetical estimand (IBCFS risk had no SPNBCs occurred) is targeted. If SPNBCs are treated as competing events, the IBCFS risk only while subjects remain free from any SPNBC is estimated. In our four large clinical trial dataset example, all three approaches yielded relatively similar results, with the largest differences being observed on the absolute risk scale. CONCLUSIONS: Full standardization of endpoints can only be achieved when the excluded components (like SPNBCs for IBCFS) of the endpoints are considered (and reported in detail!) as well. Different statistical approaches address different clinical questions. For the majority of clinical trials, reporting the total effect ('ignoring' SPNBCs) on IBCFS as well as invasive disease-free survival is strongly recommended.
Source: ESMO Open
Authors: D. Hlauschek, C. Fesl, M. Gnant
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,910
OpenAlex ID: W4411200033