Antibody Drug Conjugates: A frontier in cancer treatment(sagelyhealth.com)
sagelyhealth.com
Antibody Drug Conjugates: A frontier in cancer treatment
https://www.sagelyhealth.com/blog-posts/cancers-magic-bullet-antibody-drug-conjugates
46 comments
CAR-T may have a ~50% cure rate for blood cancers:
https://ashpublications.org/blood/article/141/19/2307/494672...
My uncle underwent CAR-T therapy and the turn around was amazing. He's got a brand new lease on life. We would have lost him without it.
What disease did it address?
> whether we might see even more investment into research
The field would benefit greatly from even a fraction of the investment in the AI and crypto bubbles
The field would benefit greatly from even a fraction of the investment in the AI and crypto bubbles
AStonesThrow(3)
Pfizer acquired Seagen for $43 billion. Seagen has some very promising ADC’s in the queue.
https://www.pfizer.com/news/press-release/press-release-deta...
https://www.pfizer.com/news/press-release/press-release-deta...
My wife, now deceased from MBC, took two ADCs that had the same payload (chemo) but two different antibody signatures. The first one did great for about 3 months. The second one failed, major progression. But the conclusion was still unknown, what information did that give us? Was it the cancer rejecting the antibody or mutating to overcome the chemo effects?
Tools we need (large scale):
- how can we identify when the cancer mutates, what part of the ADC is it rejecting? This information is valuable.
- how do we assign different payloads? Can this be scaled? Do new payload/antibody delivery systems need to go through the FDA each time? How do we streamline this to add a wider net to catch cancer mutations?
Tools we need (large scale):
- how can we identify when the cancer mutates, what part of the ADC is it rejecting? This information is valuable.
- how do we assign different payloads? Can this be scaled? Do new payload/antibody delivery systems need to go through the FDA each time? How do we streamline this to add a wider net to catch cancer mutations?
You can do blood tests to catch cancer mutations, used presently. Many blood tests.
It can be a cancer mutating. It can be the cancer not mutating (occurs after drug given) but is mutated (mutation occurred already). We have a population of a billion cancer cells. Treat with drug. 999 million die, 1 million survive because they had that mutation conferring "fitness" i.e. resistance to drug. three months later, we have a billion cancer cells again, descended from the 1 million cells. Now those billion cells also get the next line of therapy. 999 million die, 1 million survive. 3 months later, those 1 million are now a billion again. And so on. Point being, I'd think of it more as a selection mechanism as Darwin taught us, not the cancer automatically generating a resistance mutation - the blind watchmaker. In this case there was probably a mutation changing or downregulating the antibody target, probably HER2 or something related.
I don't know what "rejecting the antibody" means. You would need to look up the ADME to understand physiology and think about how the cancer might modulate that.
All this talk of "oh how do we catch these mutations".. ... ... there are a few dozen companies that will tell you what mutations are there straight up, just from the blood. From the tumor, any hospital can tell you all the mutations. The problem isn't that we don't know. We can tell you the mutations. We can tell you the new mutations. So we find 100 mutations let's say. Okay. Next question, for each of the 100 mutations, does it cause resistance, yes or no? Do two of them together cause resistance? Are they from the same cell, they could be from distinct tumor lineages. You know most of the mutations are what are called "passenger mutations" right? Red herrings.
So this is the pathology of computer scientists. "If only a clever programmer took a crack at this, these biologists what with their humanities style miasmas and rote memorization topical field". Many have in fact... Bioinformatics goes as far back as the 1970s I'd say if not further. And that clever programmer did find all the mutations. Did a great job. Pretty well developed. Then you say "ah let us understand what the mutation does!" Okay. So now you're taking a subset of the broader field of genomic variation and computationally deriving how that variant influences trillions of different cells interacting with an antibody protein with a chemical bound to it. Congrats, if you're such a "clever programmer" than by solving this, you've solved life itself! Basically this notion is "this looks easy, what is this, like the 3-SAT problem, figure if someone was clever enough to take a crack at it then that would solve this whole 3-SAT issue!" completely blind to the fact that it's just as hard (and the same) as proving P=NP. So if you ARE a hardcore clever programmer, then this rabbit hole goes as deep if not deeper than P=NP, and the cancer will humble you, as this is what cancer does, to humble.
It can be a cancer mutating. It can be the cancer not mutating (occurs after drug given) but is mutated (mutation occurred already). We have a population of a billion cancer cells. Treat with drug. 999 million die, 1 million survive because they had that mutation conferring "fitness" i.e. resistance to drug. three months later, we have a billion cancer cells again, descended from the 1 million cells. Now those billion cells also get the next line of therapy. 999 million die, 1 million survive. 3 months later, those 1 million are now a billion again. And so on. Point being, I'd think of it more as a selection mechanism as Darwin taught us, not the cancer automatically generating a resistance mutation - the blind watchmaker. In this case there was probably a mutation changing or downregulating the antibody target, probably HER2 or something related.
I don't know what "rejecting the antibody" means. You would need to look up the ADME to understand physiology and think about how the cancer might modulate that.
All this talk of "oh how do we catch these mutations".. ... ... there are a few dozen companies that will tell you what mutations are there straight up, just from the blood. From the tumor, any hospital can tell you all the mutations. The problem isn't that we don't know. We can tell you the mutations. We can tell you the new mutations. So we find 100 mutations let's say. Okay. Next question, for each of the 100 mutations, does it cause resistance, yes or no? Do two of them together cause resistance? Are they from the same cell, they could be from distinct tumor lineages. You know most of the mutations are what are called "passenger mutations" right? Red herrings.
So this is the pathology of computer scientists. "If only a clever programmer took a crack at this, these biologists what with their humanities style miasmas and rote memorization topical field". Many have in fact... Bioinformatics goes as far back as the 1970s I'd say if not further. And that clever programmer did find all the mutations. Did a great job. Pretty well developed. Then you say "ah let us understand what the mutation does!" Okay. So now you're taking a subset of the broader field of genomic variation and computationally deriving how that variant influences trillions of different cells interacting with an antibody protein with a chemical bound to it. Congrats, if you're such a "clever programmer" than by solving this, you've solved life itself! Basically this notion is "this looks easy, what is this, like the 3-SAT problem, figure if someone was clever enough to take a crack at it then that would solve this whole 3-SAT issue!" completely blind to the fact that it's just as hard (and the same) as proving P=NP. So if you ARE a hardcore clever programmer, then this rabbit hole goes as deep if not deeper than P=NP, and the cancer will humble you, as this is what cancer does, to humble.
> The Food and Drug Administration approves new cancer treatments when they are shown in clinical trials to be more effective than the current approved cancer treatment for a given cancer type. The approval comes when the treatment extends lifespan without overly increasing the risks of serious side effects.
This standard just doesn’t make sense to me. Every case is different, so why “throw out” a treatment option for only being within the margin of error of efficacy on a general population? It must be incredibly frustrating for patients to know about these treatments…
This standard just doesn’t make sense to me. Every case is different, so why “throw out” a treatment option for only being within the margin of error of efficacy on a general population? It must be incredibly frustrating for patients to know about these treatments…
It's fine to ask for an approval for a part of the population (e.g. people with cancer exhibiting particular markers).
Never let a customer off the hook too early
I remember reading somewhere that doctors are experimenting with (or wanting to) some of these more modern treatments first. Right now they're all "exhaust your other options, and then try the new stuff as a weapon of last resort". The more modern treatments could be more effective if we try them first instead of waiting, but we just don't have the data.
Anyone have a source on this half-recollected fact of mine? I can't find one.
Anyone have a source on this half-recollected fact of mine? I can't find one.
Anecdotally, I can support this.
A loved one is moving through this pathway now. Its has you said: try all the established drugs first, even if they aren't directly indicated for the cancer diagnosis of the patient, then move on to the more modern and specific treatments once the established medicines have failed.
One problem with that approach is, once the patient has reached the point of several failed treatments, the cancer has possibly become advanced and the patient is worn down by the earlier interventions. In my observation, this confluence makes the newer medicines simultaneously harder for the patient to tolerate but also appear less effective than might've been the case at earlier stages of the disease.
A loved one is moving through this pathway now. Its has you said: try all the established drugs first, even if they aren't directly indicated for the cancer diagnosis of the patient, then move on to the more modern and specific treatments once the established medicines have failed.
One problem with that approach is, once the patient has reached the point of several failed treatments, the cancer has possibly become advanced and the patient is worn down by the earlier interventions. In my observation, this confluence makes the newer medicines simultaneously harder for the patient to tolerate but also appear less effective than might've been the case at earlier stages of the disease.
The issue with ADCs is you're putting a really giant homing system on a tiny warhead. How many molecules of X does it take to kill a cell? There is one really good warhead candidate, but un?fortunately it's illegal to research
What candidate are you referring to?
Nope. Don't want to be on surveillance lists, even with an anon account. Was researched in the 80s, before we knew enough about the antibody side to make a good go of it. then dropped for obvious reasons (it didn't work which makes sense in retrospect given the crude conjugation strategy plus the restrictedness of the warhead side). That's all I'll say.
Probably enough info for you to figure it out if you're smart. Don't wind up in jail.
Probably enough info for you to figure it out if you're smart. Don't wind up in jail.
Your paranoia is not justified. merely mentioned potential treatments (even ethically questionable ones) does not get you on any "watchlists".
Just say what it is you're referring to, or don't mention it at all.
Just say what it is you're referring to, or don't mention it at all.
I will do this for you. List 20 compounds surveiled by the US government for potential association with terrorist groups in a comment under your account below this comment and I will answer yes/no if it is in the group as a whole.
Merely mentioning illegal/regulated drugs does not get you on a watch list.
Answering yes/no to a list does not protect you. I will not continue to engage.
That's fine if you don't respect my boundaries I don't need your engagement
Are you referring to monoclonal antibodies? For which a Nobel Prize was awarded in 2018, and research has continued since the 80s?
No, that's the "homing system" side of this analogy. GP is talking about some sort of poison (chemo), but I don't know which one or what the paranoia is about.
Yeah there's stuff that will get you on lists especially since 9/11. But some of them have been restricted since before that, though de facto unenforceable. Today there is the ability to ingest huge volumes of data. I just don't want to deal with the pain in the ass even if I'm found innocent.
Ever more effective cocktails of drugs that target the cancers while minimising harm to everywhere else, the closer we get to attacking cancer cells from so many different directions they simply can't adapt in time and we wipe it out immediately (a cure). One can only hope.
I whether we might see even more investment into research like this as it gets closer to deliver curative, rather than just suppressive outcomes, as well