Response
Year: 2015
Type: article
Abstract: In addition to the three fundamental issues we discussed in our commentary ( 1 ) on the paper by Tomasetti and Vogelstein ( 2 ), Tarone points out another problem. The regression they carried out included four osteosarcoma subtypes (legs, pelvis, arms, head), and also the aggregate of the four (osteosarcoma). Clearly that overlap is inappropriate. After omitting the four subtype-based points and retaining the aggregate, Tarone found the R 2 reduced from 0.65 to 0.49. Tarone then points out that the smaller R 2 he computed is consistent with estimates that about half of cancer deaths in the UK are due to lifestyle and environment. Unfortunately, this observation recapitulates an error committed by Tomasetti and Vogelstein ( 2 ) by again implying that a measurable fraction of cancer is due to stochastic errors in stem cell replication and that the remainder must therefore be due to lifestyle and environment. As we argued in our commentary, one cannot infer causality from an R 2 and cannot quantify the contribution of other factors by subtracting R 2 from 1.
DOI: 10.1093/jnci/djv228
Authors: Clarice R. Weinberg, Dmitri Zaykin
Language: English
Subfield: Oncology
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W4233352529