Comparing median overall survival times is often the TL;DR takeaway when glancing at a trial evaluating oncologic therapies. Here is an excellent editorial explaining how skimming an abstract can easily obscure the phenomenon of one intervention “skimming the median.” This is where Kaplan-Meier curves separate–clearly albeit minimally–before each plateauing while straddling the (arbitrary) 50% survival benchmark. The figure says it all. Ultimately, an intervention may help 10% of people, but, when that is realized where 40-something% achieve long-term survival without the intervention and 50-something% with the intervention, there can be substantial (read years-long) differences in “median” survival times that drastically exaggerate the benefit seen across a population. | Fojo & Bates, JAMA Oncol 2025
