Background McArthur et al1 recently reported the results of vemurafenib in BRAF mutation positive melanoma (BRIM-3 study) versus dacarbazine. Difference between medians in overall survival (OS) was 3.9 months (13.6 vs. 9.7, respectively). However, given the shape of the curves, difference in median survival (DMS) may not provide a good estimate of the survival benefit.
Purpose The aim of this study was to reanalsze the survival benefit of vemurafenib in melanoma from the OS curves using an area under the curve (AUC) based method.
Material and methods Kaplan-Meier OS curves were extracted from McArthur et al ’s article. Graphical AUC methods were applied to vemurafenib versus dacarbazine curves and compared with DMS reported in the study. According to a previously published method,2 AUC was assessed. A vertical cutting line at the hand side of the graph was made based on the number of patients at risk. It was agreed that this cutting limit was defined with at least 10 patients at risk in each group or 30 in total. The AUC method quantifies the difference between areas, and the results are expressed in time units. Photoshop-CS6 was used for graphical AUC calculation.
Results AUC based reanalysis of OS curves included 63% patients with 18 months of follow-up, giving 44 and 24 patients at risk in the vemurafenib and dacarbazine groups, respectively. For OS, the AUC method showed a benefit of 2.77 months in favour of vemurafenib (9.45 vs 6.68). There was a gap of 1.13 months between the two methods.
Conclusion AUC based analysis showed a shorter survival benefit than the difference in median survival. This is probably related to the shape of the curves, which diverged at the medium zone of the graph. This may have implications on cost effectiveness of treatment in a scenario of BRAF mutation positive melanoma.
References and/or Acknowledgements
McArthur GA, et al. Safety and efficacy of vemurafenib in BRAF(V600E) and BRAF(V600K) mutation-positive melanoma (BRIM-3): extended follow-up of a phase 3, randomised, open-label study. Lancet Oncol 2014;15:323-32
Alegre-DelRey EJ, et al. Area-based measures for assessing survival benefit in Kaplan-Meier’s curves. ESMO Congress, Amsterdam 2013. URL: http://eccamsterdam2013.ecco-org.eu/scientific-programme/searchable-programme.aspx#anchorScpr (accessed 14/10/2014)
References and/or AcknowledgementsNo conflict of interest.