Influence of censoring on conclusions of trials for women with metastatic breast cancer

Arnoud J Templeton; Olga Ace; Eitan Amir; Francisco Vera-Badillo; Alberto Ocana; Gregory R Pond; Ian F Tannock

doi:10.1016/j.ejca.2014.12.016

Influence of censoring on conclusions of trials for women with metastatic breast cancer

Eur J Cancer. 2015 Apr;51(6):721-4. doi: 10.1016/j.ejca.2014.12.016. Epub 2015 Feb 26.

Authors

Arnoud J Templeton¹, Olga Ace¹, Eitan Amir¹, Francisco Vera-Badillo¹, Alberto Ocana², Gregory R Pond³, Ian F Tannock⁴

Affiliations

¹ Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, Department of Medicine, University of Toronto, Toronto, Canada.
² Medical Oncology, University Hospital, Albacete, Spain.
³ Department of Oncology, McMaster University, Hamilton, Canada.
⁴ Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, Department of Medicine, University of Toronto, Toronto, Canada. Electronic address: ian.tannock@uhn.ca.

PMID: 25728257
DOI: 10.1016/j.ejca.2014.12.016

Abstract

Progression-free survival and time-to-progression (PFS/TTP) are used commonly as primary end-points in trials evaluating treatments for metastatic breast cancer (MBC). We reviewed the impact of censoring on interpretation of these end-points. A systematic review identified phase 3 trials in MBC published between 2001 and 2012 that reported hazard ratios (HRs) for PFS/TTP and Kaplan-Meier curves indicating numbers at risk. We calculated HRs for time-to-treatment-failure (TTF) where discontinuation of treatment for any reason is considered an event. Mean HRs for PFS/TTP, TTF, and overall survival (OS) were 0.79, 0.89 and 0.91, respectively. Unbalanced censoring of patients prior to progression was prevalent, usually with more patients censored in the experimental arms. There was moderate-to-poor correlation of HRs of PFS/TTP and TTF with HRs for OS. We suggest that TTF should be reported as supportive analysis in registration trials and extent of missing data due to censoring be considered in decisions made by regulatory agencies.

Keywords: Breast cancer; Clinical trial; End-point; Progression free survival; Time to progression; Time to treatment failure.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Breast Neoplasms / pathology
Breast Neoplasms / therapy*
Clinical Trials as Topic / methods*
Data Interpretation, Statistical*
Disease-Free Survival
Endpoint Determination / methods*
Female
Humans
Linear Models
Neoplasm Metastasis
Proportional Hazards Models