3 ms·

Yeah, I always make sure to say cancer(s) so we can skip the “it’s not one disease “ digression. It adds nothing to the conversation. According to Craig Venter

Yeah, I always make sure to say cancer(s) so we can skip the “it’s not one disease “ digression. It adds nothing to the conversation. According to Craig Venter, early detection is what we need to eliminate cancer(s): https://youtu.be/iUqgTYbkHP8?t=15m37s https://youtu.be/iUqgTYbkHP8?t=15m37s

lolinder 2y ago
"It's not one disease" certainly preempts a lot of uninformed takes, but that in itself is a major contribution to the conversation we ought to be having. If your interlocutor is still thinking about cancer as a disease and not a large category of diseases with different causes, different characteristics, and therefore different solutions, there isn't much point to continuing that conversation until the situation has been clarified.
ekianjo 2y ago
early detection? there are not enough diagnostic tools that are already applied systematically right now for patients who have known cancers. this kind of idea is great on paper but hardly scales.
- LarsDu88 2y ago
  Enough diagnostic tools? Hah. Most cancer detection assays are pretty much garbage if you were to compare them to the latest advances in say deep learning image recognition models. You can actually compare the sensitivity/specificity for PSA prostate screens versus something like ResNet for image recognition. The latter is in the high 90s for both metrics. The former can be as low as 0.06 for specificity if you want to actually achieve high sensitivity (detect prostate cancer with extremely extremely high false positives)
the__alchemist 2y ago
100% re early det. Some projects to look at: GRAIL, PanSeer (blood cfDNA next-gen sequencing), and several ones looking at the fragmentome. I imagine an effective solution will take the form of something like this, but cheap and convenient enough so that everyone can get it done regularly. "cancer isn't a single disease"is often used as a thought-terminating cliche.

lolinder 2y ago

"It's not one disease" certainly preempts a lot of uninformed takes, but that in itself is a major contribution to the conversation we ought to be having. If your interlocutor is still thinking about cancer as a disease and not a large category of diseases with different causes, different characteristics, and therefore different solutions, there isn't much point to continuing that conversation until the situation has been clarified.

ekianjo 2y ago

early detection? there are not enough diagnostic tools that are already applied systematically right now for patients who have known cancers. this kind of idea is great on paper but hardly scales.

LarsDu88 2y ago
Enough diagnostic tools? Hah. Most cancer detection assays are pretty much garbage if you were to compare them to the latest advances in say deep learning image recognition models. You can actually compare the sensitivity/specificity for PSA prostate screens versus something like ResNet for image recognition. The latter is in the high 90s for both metrics. The former can be as low as 0.06 for specificity if you want to actually achieve high sensitivity (detect prostate cancer with extremely extremely high false positives)

the__alchemist 2y ago

100% re early det. Some projects to look at: GRAIL, PanSeer (blood cfDNA next-gen sequencing), and several ones looking at the fragmentome. I imagine an effective solution will take the form of something like this, but cheap and convenient enough so that everyone can get it done regularly. "cancer isn't a single disease"is often used as a thought-terminating cliche.