Hacker News Reader

Comment by nyagaga | Hacker News Reader

Employee in a CAR-T company here.

More than proof of concept if they went with such a statement to the press. Also, ongoing Phase 1/2 means that by 5-10 years this therapy may become commercial, the initial target being refractory/relapsing tumours as stated in the announcement. It may take some time for the therapeutic target population to be expanded, but this is not a minimal achievement.

My view is that this being an autologous therapy (the patient's very own leucocytes are being engineered: the entire manufacturing cycle [I speculate 30-50 days] is done for just a single patient), this will be very expensive and probably only covered on an insurance basis in certain countries only. Still better than nothing, and even science has to make money. The real breakthrough will be arriving to heterologous therapies, where healthy donors leucocytes can be engineered and administered to any patient.

View on HN

In the next five years, how many people will die of a cancer that this could have prevented? Why do we accept this, and aren’t willing to bring it to market faster?

Liability. This is a perverse incentive structure, but it's not an easy problem.

If people die due to your inaction, you are not liable.

If people die due to your actions, you very much are.

Even a minor issue can end up being harmful. If you have a cancer treatment that's safe and effective, but you haven't fully vetted it and it turns out to be less effective than another available option, then your drug would cause harm.

It's safer to test these things, otherwise you get sued, go to jail, your company folds, and nobody gets cancer treatments anymore.

I think it should be allowed that, for people with illnesses that end up inevitably fatal in a high percent range, they be allowed to get early access to experimental medication once it's passed some sort of verification.

Right now, only those very rich have that chance, they pay the pharma companies (and a host of outright quacks)... I'd love to see cancer medication being accessible for everyone, no matter the amount of money on their bank account.

This is my point. We should try to change this. Streamlining clinical trials would be a first step.

If you have an effective way to streamline trials, I think people would be very very interested in it, regardless of the liability argument (of which I am very very skeptical). Already trials are usually pushed to be as fast as possible, because the patent clock runs out on treatments, so a faster answer is usually the best answer to achieve for investors.

Because people can be killed by unexpected side effects of treatments that are incautiously rolled out too soon. People can also die because the money went into treatments that turned out to be ineffective when the resources should have been spent on searching for things that actually worked.

This is only one side of the question. There are thousands upon thousands of potential therapies, all with varying degrees of evidence that they work, with varying degrees of match between the available evidence and the true reality of the potential of that therapy.

We need some process to distinguish between what works and what doesn't work.

The question isn't merely "how quickly can we get this one thing to market" it's "how can we prove that it works as quickly and cheaply as possible, without biasing us to treatments that don't actually work, and with giving all options their proper due."

For a while in machine learning there was a field called "active learning" (might still be ongoing?) that was all about how to choose the most informative questions to ask to learn as quickly as possible.

We face that same question with cancer therapies. Every patien that is on one trial can not be on a trial for a different therapy, so that's one limited resource for how quickly we can learn. But the other limited resource is of course how much we want to spend on validating therapies versus deploying existing, working therapies to a broader group of humans.

I think you make good points. On the other hand, the FCC needs 9 months to review trial data. I think those are the points that could be sped up.

I would hope that in 15-20 years autologous therapy technologies will have matured and cost per-patient will be driven down significantly.

Reply on news.ycombinator.com