The international surgical journal with global reach

This is the Scientific Surgery Archive, which contains all randomized clinical trials in surgery that have been identified by searching the top 50 English language medical journal issues since January 1998. Compiled by Jonothan J. Earnshaw, former Editor-in-Chief, BJS

Inconsistent selection and definition of local and regional endpoints in breast cancer research. BJS 2014; 101: 1657-1665.

Published: 13th October 2014

Authors: M. Moossdorff, L. M. van Roozendaal, R.‐J. Schipper, L. J. A. Strobbe, A. C. Voogd, V. C. G. Tjan‐Heijnen et al.

Background

Results in breast cancer research are reported using study endpoints. Most are composite endpoints (such as locoregional recurrence), consisting of several components (for example local recurrence) that are in turn composed of specific events (such as skin recurrence). Inconsistent endpoint selection and definition might lead to unjustified conclusions when comparing study outcomes. This study aimed to determine which locoregional endpoints are used in breast cancer studies, and how these endpoints and their components are defined.

Method

PubMed was searched for breast cancer studies published in nine leading journals in 2011. Articles using endpoints with a local or regional component were included and definitions were compared.

Results

Twenty‐three different endpoints with a local or regional component were extracted from 44 articles. Most frequently used were disease‐free survival (25 articles), recurrence‐free survival (7), local control (4), locoregional recurrence‐free survival (3) and event‐free survival (3). Different endpoints were used for similar outcomes. Of 23 endpoints, five were not defined and 18 were defined only partially. Of these, 16 contained a local and 13 a regional component. Included events were not specified in 33 of 57 (local) and 27 of 50 (regional) cases. Definitions of local components inconsistently included carcinoma in situ and skin and chest wall recurrences. Regional components inconsistently included specific nodal sites and skin and chest wall recurrences.

Conclusion

Breast cancer studies use many different endpoints with a locoregional component. Definitions of endpoints and events are either not provided or vary between trials. To improve transparency, facilitate trial comparison and avoid unjustified conclusions, authors should report detailed definitions of all endpoints.

Full text