4 ms·

This sounds amazing. I know this is just one case, but can someone with more knowledge confirm that this is as game-changing as it seems?

by osrec 8y ago

TheAdamAndChe 8y ago
It is a spectacular proof of concept, but it's not exactly game-changing in any way that will revolutionize cancer care around the world in the near future. This was incredibly labor- and capital-intensive procedure that involved extracting tumor material, sequencing its genes, determining which main mutations appear in the material, finding immune cells that attack that specific mutation, replicating those immune cells, and then injecting them into the patient. Since most cancers carry unique or nearly-unique mutations, this treatment would need to be tailor-made for every single patient, meaning its use would only be limited to those with a lot of resources.
- ForHackernews 8y ago
  Gene-sequencing costs just keep getting cheaper and cheaper.[0] If this result is replicated, there will be strong incentives to streamline and automate other steps in the process as well. https://www.genome.gov/images/content/costpermb_2017.jpg https://www.genome.gov/images/content/costpermb_2017.jpg
  - nitrogen 8y ago
    Indeed; I can imagine machines performing all of the more laborious steps, with doctors just needing to verify the results and perform the procedures. Mass-produced tailor-made treatments.
  - aaavl2821 8y ago
    Gene sequencing is a small part of the costs. The first approved car-t cell therapies are priced at ~$500k, took hundreds of millions to develop and cost over $30k per batch to make at scale. These more personalized approaches that are still unoptimized in terms of manufacturing process might cost $60k plus per batch The manufacturing process is logistically and technically very difficult. You are engineering live cells, an extremely complex and delicate "process". Things that shouldn't affect product quality for simpler drugs like proteins end up totally changing the biologic function of the product There's been effort to automate this but it's been difficult. Innovation here would be incredibly valuable
    - jtchang 8y ago
      Out of curiosity why is engineering live cells at scale impossible? Is it because there is a huge probability that things will go wrong? Is it logistically and technically more difficult than manufacturing say a microprocessor? Almost everything we do these days is automated to some degree at some point. Is it mostly a question of large sums of capital to develop the tooling necessary?
      - aaavl2821 8y ago
        I don't think it's impossible, it's just really hard. I have no experience with microprocessor manufacturing so can't really speak to that. I also dont have experience in the lab but have provided operations / admin support to cell therapy manufacturing processes Part of the challenge is just that live cells are complex living things. They interact with their environment in difficult to predict ways. You can only measure so many cell characteristics at a time, certainly not enough to get a full picture of everything that's going on, which makes in-process quality testing very difficult. A manufacturer we worked with said "process is product", which means we dont know how to measure product quality that well so we have to basically reinvent the process every time it changes. I think getting the right tooling to measure quality would help but that would require some really breakthrough innovation. Another challenge is shelf life. You need to get these cells to patients quickly once they leave the lab, often requiring climate controlled air shipment if the patient is far from the lab. And you dont want to build too many labs, as transferring a process from one lab to another is very expensive and fickle.
    - dorchadas 8y ago
      Once you have the gene sequenced, could you automate the part where you pinpoint the mutation? And then couldn't you automate which cells you need to engineer in order to attack that mutation? That seems like it could take out at least some time/money, though I know the biggest issue would be in engineering the cells.
      - aaavl2821 8y ago
        Discovering which mutations are important is hard, but once you know what mutation to target, finding it is relatively straightforward and is done by software. You can actually test for 500K+ specific mutations for a few hundred dollars with a gene chip, you dont even need sequencing if you know what you're looking for. But to discover what mutations matter you need a lot of training data including expensive / hard to access tumor samples, and high quality clinical data showing how patients with different mutations fared. The process of selecting cells to engineer is pretty complicated and labor intensive. Basically you have to get a patient to a clinic, draw blood (or biopsy), process the sample for shipment, ship it to a lab (often have just a few hours to get this done), then you do a number of steps to isolate the specific cell types you want. A lot of this is manual but isn't really novel, the techniques are pretty well established and straightforward, some probably could be automated but im not sure how much, and it would be expensive to automate. I think a few large manufacturers probably invest heavily in automation but most small companies or academic labs probably cant afford it
- epistasis 8y ago
  Agreed that this process would need to be tailor made for each person, but there are a lot of efforts out there at industrializing and streamlining this process. The process itself is extremely difficult with lots of ways that it can go wrong. But as with other areas of molecular biology, there's hope that the consistency of automation may improve that. Current cancer treatments that are used in practice can easily make it in to tens of thousands per month of treatment. That provides a lot of wiggle room for doing expensive things. Once there's a reason to produce thousands or millions of individualized cell therapies, it's likely that the process could be made economically feasible.
gus_massa 8y ago
Let's be optimistic, but cancer is a very broad and difficult problem. Many treatments are overhyped by the press office of the hospital/university. The article has a few good cautious warning, in particular: > But experts caution that the treatment has only proved itself in one woman and that the clinical trials are needed to see how effective the therapy could be in other cancer patients. Too good short explanations in comic forms are: http://phdcomics.com/comics.php?f=1162 http://phdcomics.com/comics.php?f=1162 https://xkcd.com/931/ https://xkcd.com/931/
- garmaine 8y ago
  People who are down voting you probably aren’t aware the comics you link to are public service specials written by a scientist (I think?) in the first case, and the life partner of a cancer survivor in the second. They are actually very good overviews of the depth of problems involved in “curing cancer,” and the uncertainty involved in real instances of cancer recovery, not captured in TFA headline.
roberte3 8y ago
Its not game-changing. That said, its a great step forward. The T-Cell treatments or CAR-T cell treatments is one of the most promising lines of research in cancer treatment. Fred Hutch is doing a lot of research on right now. Seattle Cancer Care (which is associated with Fred Hutch) has an entire floor devoted to CAR-T trials. https://www.seattletimes.com/pacific-nw-magazine/the-hutch-is-closing-in-on-a-cure-for-cancer/ https://www.seattletimes.com/pacific-nw-magazine/the-hutch-i... The interesting bit is that this is a successful trial on a solid tumors. CAR-T has shown that its most effective on non-solid tumors or blood based cancers. A few months ago there was a Stanford press release on them doing a combination of injecting solid tumor CAR-T and alcohol, which caused a singular tumor to be attacked and removed.
snarf21 8y ago
Another good step forward. I am excited about the advances that are being made using things like poliovirus to attack lethal cancers.

This sounds amazing. I know this is just one case, but can someone with more knowledge confirm that this is as game-changing as it seems?

by osrec 8y ago

This sounds amazing. I know this is just one case, but can someone with more knowledge confirm that this is as game-changing as it seems?

TheAdamAndChe 8y ago

It is a spectacular proof of concept, but it's not exactly game-changing in any way that will revolutionize cancer care around the world in the near future. This was incredibly labor- and capital-intensive procedure that involved extracting tumor material, sequencing its genes, determining which main mutations appear in the material, finding immune cells that attack that specific mutation, replicating those immune cells, and then injecting them into the patient. Since most cancers carry unique or nearly-unique mutations, this treatment would need to be tailor-made for every single patient, meaning its use would only be limited to those with a lot of resources.

ForHackernews 8y ago
Gene-sequencing costs just keep getting cheaper and cheaper.[0] If this result is replicated, there will be strong incentives to streamline and automate other steps in the process as well. https://www.genome.gov/images/content/costpermb_2017.jpg https://www.genome.gov/images/content/costpermb_2017.jpg
- nitrogen 8y ago
  Indeed; I can imagine machines performing all of the more laborious steps, with doctors just needing to verify the results and perform the procedures. Mass-produced tailor-made treatments.
- aaavl2821 8y ago
  Gene sequencing is a small part of the costs. The first approved car-t cell therapies are priced at ~$500k, took hundreds of millions to develop and cost over $30k per batch to make at scale. These more personalized approaches that are still unoptimized in terms of manufacturing process might cost $60k plus per batch The manufacturing process is logistically and technically very difficult. You are engineering live cells, an extremely complex and delicate "process". Things that shouldn't affect product quality for simpler drugs like proteins end up totally changing the biologic function of the product There's been effort to automate this but it's been difficult. Innovation here would be incredibly valuable
  - jtchang 8y ago
    Out of curiosity why is engineering live cells at scale impossible? Is it because there is a huge probability that things will go wrong? Is it logistically and technically more difficult than manufacturing say a microprocessor? Almost everything we do these days is automated to some degree at some point. Is it mostly a question of large sums of capital to develop the tooling necessary?
    - aaavl2821 8y ago
      I don't think it's impossible, it's just really hard. I have no experience with microprocessor manufacturing so can't really speak to that. I also dont have experience in the lab but have provided operations / admin support to cell therapy manufacturing processes Part of the challenge is just that live cells are complex living things. They interact with their environment in difficult to predict ways. You can only measure so many cell characteristics at a time, certainly not enough to get a full picture of everything that's going on, which makes in-process quality testing very difficult. A manufacturer we worked with said "process is product", which means we dont know how to measure product quality that well so we have to basically reinvent the process every time it changes. I think getting the right tooling to measure quality would help but that would require some really breakthrough innovation. Another challenge is shelf life. You need to get these cells to patients quickly once they leave the lab, often requiring climate controlled air shipment if the patient is far from the lab. And you dont want to build too many labs, as transferring a process from one lab to another is very expensive and fickle.
  - dorchadas 8y ago
    Once you have the gene sequenced, could you automate the part where you pinpoint the mutation? And then couldn't you automate which cells you need to engineer in order to attack that mutation? That seems like it could take out at least some time/money, though I know the biggest issue would be in engineering the cells.
    - aaavl2821 8y ago
      Discovering which mutations are important is hard, but once you know what mutation to target, finding it is relatively straightforward and is done by software. You can actually test for 500K+ specific mutations for a few hundred dollars with a gene chip, you dont even need sequencing if you know what you're looking for. But to discover what mutations matter you need a lot of training data including expensive / hard to access tumor samples, and high quality clinical data showing how patients with different mutations fared. The process of selecting cells to engineer is pretty complicated and labor intensive. Basically you have to get a patient to a clinic, draw blood (or biopsy), process the sample for shipment, ship it to a lab (often have just a few hours to get this done), then you do a number of steps to isolate the specific cell types you want. A lot of this is manual but isn't really novel, the techniques are pretty well established and straightforward, some probably could be automated but im not sure how much, and it would be expensive to automate. I think a few large manufacturers probably invest heavily in automation but most small companies or academic labs probably cant afford it
epistasis 8y ago
Agreed that this process would need to be tailor made for each person, but there are a lot of efforts out there at industrializing and streamlining this process. The process itself is extremely difficult with lots of ways that it can go wrong. But as with other areas of molecular biology, there's hope that the consistency of automation may improve that. Current cancer treatments that are used in practice can easily make it in to tens of thousands per month of treatment. That provides a lot of wiggle room for doing expensive things. Once there's a reason to produce thousands or millions of individualized cell therapies, it's likely that the process could be made economically feasible.

gus_massa 8y ago

Let's be optimistic, but cancer is a very broad and difficult problem. Many treatments are overhyped by the press office of the hospital/university. The article has a few good cautious warning, in particular: > But experts caution that the treatment has only proved itself in one woman and that the clinical trials are needed to see how effective the therapy could be in other cancer patients. Too good short explanations in comic forms are: http://phdcomics.com/comics.php?f=1162 http://phdcomics.com/comics.php?f=1162 https://xkcd.com/931/ https://xkcd.com/931/

garmaine 8y ago
People who are down voting you probably aren’t aware the comics you link to are public service specials written by a scientist (I think?) in the first case, and the life partner of a cancer survivor in the second. They are actually very good overviews of the depth of problems involved in “curing cancer,” and the uncertainty involved in real instances of cancer recovery, not captured in TFA headline.

roberte3 8y ago

Its not game-changing. That said, its a great step forward. The T-Cell treatments or CAR-T cell treatments is one of the most promising lines of research in cancer treatment. Fred Hutch is doing a lot of research on right now. Seattle Cancer Care (which is associated with Fred Hutch) has an entire floor devoted to CAR-T trials. https://www.seattletimes.com/pacific-nw-magazine/the-hutch-is-closing-in-on-a-cure-for-cancer/ https://www.seattletimes.com/pacific-nw-magazine/the-hutch-i... The interesting bit is that this is a successful trial on a solid tumors. CAR-T has shown that its most effective on non-solid tumors or blood based cancers. A few months ago there was a Stanford press release on them doing a combination of injecting solid tumor CAR-T and alcohol, which caused a singular tumor to be attacked and removed.

snarf21 8y ago

Another good step forward. I am excited about the advances that are being made using things like poliovirus to attack lethal cancers.