BRAINBITES
WEEK OF OCT 16–22, 2026 · ISSUE 31
BRAINBITES — Your weekly health, fitness, and longevity digest. Five minutes that cover the highest-signal podcasts and science before they hit the mainstream.
This week: Siddhartha Mukherjee told Peter Attia what fifteen years have changed in cancer medicine, and why the next target may be repair rather than disease. We paired it with the pancreatic cancer drug that sits in his treatment list, then added the first large randomised trial of a blood test for many cancers at once. Plus a fasting trial in Cell Metabolism, a sleep apnoea drug in the NEJM, and a testing rule from Huberman Lab Essentials.
From: The Peter Attia Drive · NEJM · Nature Medicine · Cell Metabolism · Huberman Lab Essentials
🔢 13.2 months — median survival on a new pill for pancreatic cancer, against 6.7 on chemotherapy
🎙️ Siddhartha Mukherjee on inflammation, repair and what cancer medicine has learned
🔬 Deep Dive: the first big trial of a blood test for every cancer
📰 News Bits: inflammation and fasting · a sleep apnoea drug · one supplement at a time
GM.
13.2 months. That is the median survival of people with treated, advanced pancreatic cancer on a new daily pill, against 6.7 months on chemotherapy. Which raises the question this issue answers: if treatment can move the numbers that far, what can finding cancer earlier be expected to add?
⏱ ~5 min read
🧠 FIND IT, TREAT IT
Cancer medicine is moving on both fronts, at very different speeds.
Issue 27 reported a Nature Medicine study tying faster biological ageing to more early-onset cancer. This week looks at the far end of the story: a drug changing what happens after a diagnosis, and the first large trial of a blood test built to catch cancers early.
🔢 THE VITAL SIGN · ONCOLOGY / PANCREATIC CANCER
13.2 months
Median survival on daraxonrasib, a daily pill, in people whose metastatic pancreatic cancer had already been treated once. On chemotherapy it was 6.7 months.
In RASolute 302, O'Reilly and colleagues randomly assigned 500 such patients to daraxonrasib (248) or to the chemotherapy their doctor chose (252). Median overall survival was 13.2 months against 6.7 (hazard ratio 0.40), and progression-free survival roughly doubled too, 7.2 against 3.6. Far fewer stopped for side effects, 1.2% against 11.2%, though the label warns of skin and mouth toxicity.
What it means: RAS mutations are present in more than 90% of pancreatic cancers, and this pill is aimed straight at that driver. The FDA approved it on 26 August 2026. Limits: the trial was open-label, enrolled only people already treated once, and the maker funded it. Thirteen months is a median, not a cure.
🎙️ THE EPISODE · ONCOLOGY / LONGEVITY
Inflammation loads the gun. A mutation pulls the trigger.
Siddhartha Mukherjee on what fifteen years have changed in the science of cancer.
Siddhartha Mukherjee, M.D., Ph.D., oncologist and author · The Peter Attia Drive #412 · 19 October 2026
Mukherjee returns to Attia's table with a revised The Emperor of All Maladies, fifteen years on. The detailed notes are members-only, so we stuck to what the free notes and chapter list state outright.
Inflammation is the gun; the mutation is the trigger.
Radiation, benzene and tobacco are classic carcinogens, but Mukherjee also points to chronic inflammation as a loaded gun waiting for a mutation. No blood test yet reads that state the way cholesterol reads heart risk. He expects a "high-cholesterol equivalent" for cancer, so prevention trials can start with the highest-risk people.
Deaths are falling, and the causes run late.
US cancer mortality is falling, though breast cancer's gains have levelled off after about twenty years. Trends often trail exposures by five to ten years, and early-onset colorectal cancer, plus a rise in oesophageal cancer probably tied to alcohol and tobacco together, are still only partly explained.
Maybe the target is repair, not disease.
With age, tissues lose the ability to repair and renew themselves, which Mukherjee calls regenerative failure. He and Attia debate whether longevity medicine should keep chasing diseases one at a time or protect that capacity directly.
🎙️ Start at 1:37:45 (audio) for the debate over preventing disease versus restoring repair.
The practical takeaway
Risk assessment now reaches beyond BRCA1 and BRCA2, to other genes and polygenic scores. If cancer runs in your family, ask whether broader testing is worth it.
💡 DEEP DIVE · SCREENING / THE EVIDENCE
The test for everything meets its trial
The show notes record a careful point from Mukherjee: blood tests can pick up signals of some early cancers, but screening everyone would likely bring false alarms, extra scans, biopsies and testing cascades without fewer deaths. The first large randomised trial of that idea has now reported.
NHS-Galleri randomised 142,250 people aged 50 to 77 in England to a multi-cancer blood test added to usual care, or to blood stored untested, for up to three annual rounds. The primary goal, fewer stage III or IV cancers across 12 prespecified types, was not met (rate ratio 1.03, 95% CI 0.92 to 1.14). Stage IV cancers, a key secondary endpoint, ran lower in the tested group (0.86, 0.74 to 1.00), but that is a lead, not a verdict.
A companion Nature Medicine paper, descriptive only, shows a good alarm: about 1% tested positive each round, 46% to 58% of positives were cancer, and specificity was 99.5% to 99.6%. The price is sensitivity: roughly 27% to 37% of all cancers diagnosed in each screening episode were caught.
What it means: a positive result deserves attention, and the stage shift may sharpen with follow-up. Whether earlier finds mean fewer deaths is unknown, because no mortality results are out. Grail, which makes the test, funded the trial.
📰 NEWS BITS
Inflammation did not move until the weight did.
In a 16-week feeding trial, 53 adults with metabolic risk factors ate a fully provided Mediterranean diet or their usual Western one. For eight weeks, C-reactive protein, an inflammation marker, did not budge. After twice-weekly fasting days were added, it fell in the Mediterranean group, alongside weight loss and better insulin sensitivity. The authors credit weight loss, not diet quality alone. Small, and the changes were stacked, not compared head to head.
— Cell Metabolism, Warmbrunn et al., 21 October 2026
A weight-loss drug, tried on sleep apnoea.
In two phase 3 trials, 691 adults with moderate-to-severe sleep apnoea and excess weight took weekly retatrutide, an investigational triple-hormone drug, or placebo for 80 weeks. Without diabetes, the 9 mg dose cut breathing interruptions by 34 an hour against 10 on placebo, from a baseline near 59. Weight fell 12% to 22% across doses, and the headline numbers cannot separate weight loss from the drug. Lilly funded it.
— The New England Journal of Medicine, Grunstein et al., 20 October 2026
Test supplements one at a time.
In a solo Essentials recap, Andrew Huberman puts behaviour and nutrition first, then covers supplements for sleep, hormones and focus. The rule worth borrowing is procedural: add one ingredient at a time, so you can tell what did what. The only paper in the show notes is a 2022 meta-analysis of five randomised trials of tongkat ali and testosterone, whose authors say more research is needed before clinical use.
— Huberman Lab Essentials, Dr. Andrew Huberman, 22 October 2026
📊 CHART OF THE WEEK
Both clocks roughly doubled. They are still measured in months.
P.S. Hit reply and tell us: if a blood test for dozens of cancers were offered at your next check-up, would you take it? We will share the most thoughtful answers.
What did you think of today's BrainBite?
🧠🧠🧠🧠🧠 I came for the pill and stayed for the confidence intervals.
🧠🧠🧠 The blood test trial is the part I'm repeating at dinner.
🧠 I still can't pronounce daraxonrasib. The rest landed.
We listen to every health podcast so you don't have to. Weekly. No noise.
A good test finds the cancer. A good trial finds out whether finding it helped. See you next week.
