Reply to “The St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2017: the point of view of an International Panel of Experts in Radiation …

G Curigliano,HJ Burstein,EP Winer,M Gnant,P Dubsky,S Loibl,M Colleoni,MM Regan,M Piccart-Gebhart,HJ Senn,B Thürlimann,F André,J Baselga,J Bergh,H Bonnefoi,SY Brucker,F Cardoso,L Carey,E Ciruelos,J Cuzick,C Denkert,A Di Leo,B Ejlertsen,P Francis,V Galimberti,J Garber,B Gulluoglu,P Goodwin,N Harbeck,D F Hayes,CS Huang,J Huober,H Khaled,J Jassem,Z Jiang,Per Karlsson,M Morrow,R Orecchia,K C Osborne,O Pagani,A H Partridge,K Pritchard,J Ro,E JT Rutgers,F Sedlmayer,V Semiglazov,Z Shao,I Smith,M Toi,A Tutt,G Viale,T Watanabe,T J Whelan,B Xu
IF: 51.769
2017-01-01
Annals of Oncology
Abstract:We read with great interest the article by McCoach et al.[1] investigating the association of depth of response (DepOR) with survival in patients with metastatic non-small-cell lung cancer. The authors found that good response (such as maximal tumor shrinkage of 76%–100%) was associated with longer progression-free survival (PFS) and overall survival (OS) compared with non-responders, estimated by a very low hazard ratio (HR) of 0.03. We believe that these results are prone to survival bias. Already in 1983 and re-emphasized in 2008, Anderson et al.[2] discussed the statistical challenges of analyzing survival by tumor response. Anderson et al.[2] pointed out that a bias occurs when patients are categorized into good and bad responder groups at baseline and then survival in these groups is compared. This approach is biased because in order to obtain a good response individuals have to survive long …
What problem does this paper attempt to address?