I actually don't mind if it's true, but your comment was the first time I got a tingle that this was AI generated but I still could tell it likely wasn't. Maybe used it to rewrite parts of your comment? Anyways i appreciate it and agree with you. I myself am gonna go check in the cancer cell line encyclopedia, which IME is the single cleanest large dataset curated in biology in decades.

For others who might be curious, this study is genuinely good (evidenced somewhat by it being published in Science, as a study that is not in humans but still showing actual cancer curing efficacy not just some pathway finding) is that they use difficult but necessary model systems that emulate real tumor environments - these are spontaneous tumors that show up in mice that have full immune systems. Literally 99.9% of cancer study papers don't use such systems and in my opin9on get fully invalidated for most interpretation.

Anyways I'm curious if the microenvironment will just evolve quickly to be EPO independent, as it typically seems to do something like that in real long term tumor environments you encounter in people compared to mouse models.