I just retired after 19 years in the pharma industry. Animal testing is essential for new drug safety assessment because off-target damage to tissues cannot be detected using any method other than entire animals since drug byproducts from digestion or absorption do not exist in the lab -- in vitro or in silico. Measuring drug effects on the body requires exposure to ALL of an animal's tissues to see which will react. Drug efficacy too can't be measured usefully without full body animal and human studies.
Whomever proposed this obviously did not talk to _anyone_ in the pharma industry. The strategy is completely and dangerously clueless. I'm sure Derek Lowe will very soon explain this in more detail in his blog at Science.
Fact is, pharmas already pre-screen immune response in vitro using various stat and ML algorithms, a no-brainer that's been done for years. Organoids are useful to detect only the obvious toxicities in the most obvious tissues (liver, kidney, neural). Also a no-brainer that's been done for years mostly in animal studies.
The acid test is what FDA criteria will change now in trusting nonclinical data and novel biomarker thresholds in novel criteria like simulations (much less AI black boxes). Attempts to address this also has been in development in the industry for decades, to invent new standards in a precompetitive space -- to be shared among drug developers so they can be freely employed by all. I don't see AI moving the needle, especially how inexplicable AI continues to be (as opposed to classical statistical ML methods, which CAN explain themselves).
I just retired after 19 years in the pharma industry. Animal testing is essential for new drug safety assessment because off-target damage to tissues cannot be detected using any method other than entire animals since drug byproducts from digestion or absorption do not exist in the lab -- in vitro or in silico. Measuring drug effects on the body requires exposure to ALL of an animal's tissues to see which will react. Drug efficacy too can't be measured usefully without full body animal and human studies.
Whomever proposed this obviously did not talk to _anyone_ in the pharma industry. The strategy is completely and dangerously clueless. I'm sure Derek Lowe will very soon explain this in more detail in his blog at Science.
Washington has gone insane.
https://www.fda.gov/news-events/press-announcements/fda-anno...
Fact is, pharmas already pre-screen immune response in vitro using various stat and ML algorithms, a no-brainer that's been done for years. Organoids are useful to detect only the obvious toxicities in the most obvious tissues (liver, kidney, neural). Also a no-brainer that's been done for years mostly in animal studies.
The acid test is what FDA criteria will change now in trusting nonclinical data and novel biomarker thresholds in novel criteria like simulations (much less AI black boxes). Attempts to address this also has been in development in the industry for decades, to invent new standards in a precompetitive space -- to be shared among drug developers so they can be freely employed by all. I don't see AI moving the needle, especially how inexplicable AI continues to be (as opposed to classical statistical ML methods, which CAN explain themselves).
Caveat patiens.