Comment by Noaidi | Hacker News Reader

Comment by Noaidi | Hacker News Reader

I for one would like to see the data rather than a press release. Is this even serious? A press release with no link to outside twitter?

I feel like everything companies, and presidents, say now is directly geared to triggering AI trades.

The only thing they are looking at with this drug is people living one more month over existing drugs.

THIS IS NOTHING.

"The median PFS (progression-free survival) for KEYTRUDA was 5.5 months (2-week group) and 4.1 months (3-week group) compared to 2.8 months for ipilimumab (HR 0.58, P<0.00001 for the KEYTRUDA groups vs. ipilimumab,"

https://www.merck.com/news/keytruda-pembrolizumab-mercks-ant...

The median PFS (progression-free survival) for KEYTRUDA was 5.5 months (2-week group) and 4.1 months (3-week group) compared to 2.8 months for ipilimumab (HR 0.58, P<0.00001 for the KEYTRUDA groups vs. ipilimumab, 95% CI, 0.46-0.72 for 2-week group and 0.47-0.72 for 3-week group, respectively). The estimated 6-month PFS rates for the KEYTRUDA and ipilimumab arms were 47.3 percent, 46.4 percent and 26.5 percent, respectively. One-year OS for KEYTRUDA was 74.1 percent (2-week group) and 68.4 percent (3-week group) compared to 58.2 percent for ipilimumab (HR 0.63 [95% CI, 0.47-0.83, P=0.00052] for the 2-week group and HR 0.69 [95% CI, 0.52-0.90, P=0.00358] for the 3-week group). At the time of analysis, median overall survival was not reached in any treatment group.

1. That's not what progression-free survival means.

2. I don't think AI was mentioned once in this press release.

3. This drug was approved for trial in 2014, so if it had something to do with deep learning, that would actually be a massive announcement.

4. The paragraph immediately following the one you clipped talks about Overall Response Rate (ORR) and throughout the release they discuss Overall Survival (OS).

5. This is not "NOTHING" in statistical terms, but even more so to people with melanoma.

Interjecting with a fact. Moderna described the neoantigen immunogenicity prediction algorithm they used to design the vaccines as “a deterministic machine learning (ML) algorithm,” so it may be a small neural network or more traditional algorithm, probably not deep learning.

Sorry--I was unclear. I was specifically responding to the OP's "CEOs these days want to join the AI hype wave" claim. My point was that the "hype wave" OP is referencing is really about deep learning, not what we think of as traditional statistical learning. Even the dedicated medical AI startups of that era, like BenevolentAI, were still doing traditional ML in 2014. So if Moderna really was trying to join the "AI hype train" and announce a deep learning result with this treatment, they'd be announcing that they successfully trained a deep neural network for drug discovery 12 years ago.

My interpretation of "triggering AI trades" is about tricking AI into buying stock, not tricking people into buying AI stocks.

1. That's not what progression-free survival means.

I was not pointing to PFS as OS.

2. I don't think AI was mentioned once in this press release.

I did not say it was. I said it was written FOR AI trading algorithms to read and invest on.

4. The paragraph immediately following the one you clipped talks about Overall Response Rate (ORR) and throughout the release they discuss Overall Survival (OS).

Someone noted this was an earlier study. The newest study does not even have data out and yet here we are, investing on a press release.

5. This is not "NOTHING" in statistical terms, but even more so to people with melanoma.

You are talking to someone whose father died of pancreatic cancer. "Pancreatic cancer has a higher mortality rate compared to melanoma, with an estimated 52,740 deaths from pancreatic cancer in 2026, while melanoma has a significantly lower death rate. Additionally, the 5-year relative survival rate for pancreatic cancer is only 13.7%, compared to higher survival rates for melanoma."

What they did for my father was worse than nothing.

Yes every cancer is different, responds to different therapy, and with different survival rates. However, cancer treatment, as a whole, works. Oncologists do not prescribe incredibly intense treatments like chemotherapy out of nowhere. Chemotherapy is often not curative. Instead, it's used to both prolong life and improve quality of life.

Dying from late-stage cancer is often painful and miserable. It might be shocking, but getting just a few more months from chemo can provide a greater quality of life. Keyword can. Every oncologist is different with different recommendations, and they explain the risks very concretely. There is always an option to refuse treatment, but bear in mind it's not just about how quick you die. It's also about how much that dying sucks.

They told my father Chemo woudl add six months on his six month estimated lifespan. He died in two months cause the chemo blocked his liver. My sister in law was a chemo nurse and begged him not to get it. He died on loads of morphine. His last words were "It's all a scam!"

Sorry, you have no idea. Talk to a chemo nurse.

Huh? How would chemo "block his liver"? Do you mean that he had a liver metastasis that the chemo didn't help with?

Chemotherapy can damage many organs including the liver. See "Organ Damage" on Wikipedia: https://en.wikipedia.org/wiki/Chemotherapy

A friend's sister beat colon cancer over ten years ago but now is having liver problems that took a while to pin down. They warned her she'd probably die in a couple years from the cancer without Chemo, but would have this risk long term and it recently showed up.

Talk to a chemo nurse? I've gotten two different chemo regimens. Yes, I know chemo has a risk because all medicine has a risk. I'm sorry for what happened to your father, but that doesn't fucking mean chemo is a scam. It saves lives, I should know.

No it isn’t. They did a test with great specificity and got undeniable P values. Now they can refine the action and start saving people. This is a huge, huge win.

you're also looking at a different, older study. Keytruda is a little strange in that it has a very long tail in survivorship, works best in certain cases, and median PFS isn't a really great metric.

Reply on news.ycombinator.com