Slashdot Mirror

← Back to Stories (view on slashdot.org)

Personalized Cancer Vaccines Safely Fight, Kill Tumors In Early Human Trials (arstechnica.com)

Posted by BeauHD on from the custom-made dept.

Emily Mullin reports via MIT Technology Review: Now two personalized cancer vaccine approaches appear to have safely prevented cancer relapse in a dozen patients with late-stage skin cancer. In recent years, scientists have realized that each patient's tumor harbors a unique set of genetic characteristics, or mutations. So for cancer vaccines to be effective, they'll probably also have to be unique. Two clinical trials, detailed today in separate papers in Nature, are among the first to show that this might be possible. In one trial, eight of 13 melanoma patients who got a personalized cancer vaccine were tumor-free nearly two years after being treated. In a smaller study, four of the six patients who received a vaccine had no detectable cancer for more than two years after treatment. All patients had their tumors surgically removed before getting the vaccine. The customized vaccines are an emerging class of therapies that take advantage of neoantigens, proteins that appear on tumors and seem to be specific to each cancer patient. To make the vaccines, researchers first sequenced DNA and RNA extracted from each patient's tumor. They then used computer algorithms to analyze the mutations on each tumor and predict the best targets that code for neoantigens. Based on that data, they created a personalized vaccine containing up to 20 of these neoantigens. Each patient received several injections of the vaccine over a few months.

11 of 73 comments (clear)

  1. Re: Unfortunately by Anonymous Coward · · Score: 2, Insightful

    And 9 out of 10 slashdot AC:s turns out to be morons.

  2. The problem is cost by Anonymous Coward · · Score: 5, Informative

    This all sounds promising, but the problem with personalized treatment is the cost.

    From the article:

    "The customized vaccines are an emerging class of therapies that take advantage of neoantigens, proteins that appear on tumors and seem to be specific to each cancer patient. To make the vaccines, researchers first sequenced DNA and RNA extracted from each patient’s tumor. They then used computer algorithms to analyze the mutations on each tumor and predict the best targets that code for neoantigens. Based on that data, they created a personalized vaccine containing up to 20 of these neoantigens. Each patient received several injections of the vaccine over a few months."

    Protein discovery is really hard to do. This is a good approach, by sequencing the DNA and using that to feed into a computer model to determine what proteins are expressed by the mutations. This modeling though is extremely complex and rarely accurate, as there are too many variables that are hard to determine without extensive lab analysis to support it. This could cost around $100,000 or more, if you're lucky it could cost less, but still in the $50k to $75k range.

    Then once you have your target, you need to develop something to neutralize it. Most cancer immunotherapy are monoclonal antibodies, of which the methodology to develop them is well known in industry, but can still be expensive. There's never been a synthetic, computer modeled antibody that's as good as a mouse or rabbit monoclonal antibody, and to get one good antibody you're likely going through about 1,000 mice at around $200 to $400 a pop, so this is most likely around $200k to $400k.

    And there's no way to scale this with volume, because it's discovery not production. Every person would need a new discovery every time, so you're looking at $500,000 per patient roughly speaking. So while promising science, no healthcare system can afford this without bankrupting itself.

    Herein is the problem with cancer immunotherapy. The latest one on the market is Keytruda by Merck. One year's dosage can cost around $150,000, so while it's quite effective as a drug, it can bankrupt patients and payer systems very quickly. Somehow the cost aspects of these drugs need to be addressed, because there's simply no way they be effective on the broader market without some ability to lower the cost.

    1. Re:The problem is cost by lisaparratt · · Score: 5, Insightful

      Isn't that premature optimisation? R&D costs money - it's surely a lot cheaper to cost down one working solution than to try to cost reduce every attempt.

    2. Re:The problem is cost by CrimsonAvenger · · Score: 2

      This all sounds promising, but the problem with personalized treatment is the cost.

      Yeah, those new-fangled "computers" (anyone old enough to remember when "computer" was a job description?) cost so damn much there'll never be a real demand for more than a few dozen in the world, much less the country...

      --

      "I do not agree with what you say, but I will defend to the death your right to say it"
    3. Re:The problem is cost by angel'o'sphere · · Score: 2

      And don't get me started on the stupidity of OHIP (Ontario's Health system), I've watched them spend close to $300,000 twice on friends or relatives who had zero chance of recovery.
      That would happen in most western health care systems. And rightfully so. You never really know the chance of a small wonder.
      The only treatments in Germany which can be denied, not technically, but by playing the system are transplantations.
      People are on a list, basically how urgent it is. But there is also a factor involved, how likely and how long the receiver will survive. So if a 35 year old healthy father of a family and a 65 year old alcoholics are the two possible candidates, then the 65 is "technically" on top of the list, as he needs it more urgent, BUT: he will not get it because he is an alcoholics.

      Cancer treatment is always done with maximum use of every suitable weapon available.

      --
      Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
    4. Re: The problem is cost by Kjella · · Score: 3, Insightful

      Not that long ago "sequencing the DNA" would have been a prohibitingly expensive step. Show that this works for broad classes of cancer patients and we'll find cheaper ways than guessing with lab rats. Lab work can also often be automated to a fraction of the cost in volume. For cancer in young people $500k is not bad if it keeps them cancer free, we spend huge amounts on medicines that only prolong the inevitable.

      --
      Live today, because you never know what tomorrow brings
    5. Re:The problem is cost by wren337 · · Score: 3, Insightful

      I also expect that we will find some number of relatively common mutations, over time. Cancer is a coding mistake that leads to uncontrolled growth. The fact that we've named certain of these errors, that we know how they progress and how to treat them, tells me that there are common coding mistakes that trigger cancerous growth. If we got this down to 100 or 1000 or 10000 different vaccines that covered 75% of cancers, and you could pick the right vaccine(s) with a DNA test, we'd be kicking ass.

    6. Re:The problem is cost by tobiah · · Score: 2

      Good summary of established technology, but cost is a red herring:
      1) Moores Law may not be as fast for manufacturing, but it still holds that things get cheaper on repetition. Cars, TVs and computers all used to be luxury items for the well-off.
      2) I've done a fair amount of coding for DNA/RNA analysis, and the resulting assays cost a few hundred dollars to run. A profitable $1000 test could be developed that IDs the cancer and describes the treatment in a few years time. An alternative to monoclonal antibodies is antisense drugs, which directly redirect/reprogram DNA of specific organs of the body. Antisense can be remarkably quick to produce, as the delivery package (retroviruses,etc.) are generic and there are various commercially available machines that mass-produce the genetic sequence of your choice. Isis Pharmaceuticals is one that's been working on antisense drugs for nearly 30 years, ten years ago I'd watch their R&D spit out a dozen possible drugs per day. Not hard to write D/RNA, not hard to read it, just need to bring those worlds together, along with a big database and proper analysis tools.
      D/RNA looks just like software code, with start/end markers, execute statements, comments, etc. A rapid read-write interface is only a matter of time.

      --
      "The ability to delude yourself may be an important survival tool" - Jane Wagner -
  3. Early Human Trials by famebait · · Score: 2

    Now we only need to try it out on modern humans.

    --
    sudo ergo sum
  4. I know one of the patients by CharlieG · · Score: 5, Informative

    He thought he was dead for sure - stage IV melanoma. He's cancer free. Amazing stuff

    --
    -- 73 de KG2V For the Children - RKBA! "You are what you do when it counts" - the Masso
  5. Re:Control group by Pasquina · · Score: 5, Informative

    Early stage (Phase I) trials usually don't have control groups because the goal is to test for toxicity of the therapy. Later stage (Phase II, III) look to compare efficacy against the "standard of care." In this study, no one was admitted to the study but did not receive the therapy - that will happen in the next study.

    Phase I trials are traditionally done in health volunteers, but these days, cancer trials are frequently performed in late stage cancer patients because they are desperate and have no other (Western medical) options. These patients had exceedingly little chance of spontaneous recovery, so you can assume the "control" group would have close to 0% survival.

    The fact that they got such a huge response is amazing and highly promising.