412 ‒ The evolving science of cancer: emerging approaches to prevention and early detection, & more

KEYPOINTS

• 01

  Fifteen years after The Emperor of All Maladies, the picture splits into prevention, early detection and treatment—with real progress in each, but big gaps in prevention.

  Mukherjee says overall US cancer mortality is falling, so 'cancer is exploding' is nonsense. But colorectal cancer in young people is rising, breast cancer's mortality gains have plateaued, and esophageal cancer is climbing. Endometrial cancer's rise is most likely obesity; the colorectal trend is 'a complete mystery.' Trends lag exposures by roughly 5–10 years.

• 02

  The 'bad luck' model of cancer is early, not wrong: the 5% hereditary figure is too low, and random mutations need a second push to become tumours.

  Attia recalls Vogelstein's 2015/2017 Science papers: roughly 5% of mutations hereditary, ~30% environmental, ~two-thirds random. Mukherjee argues single-gene counts miss much of hereditary breast cancer—known single genes cover only about a third to a quarter of it—with the rest from hundreds of small-effect variants. 'Bert is not wrong. He is just early.'

• 03

  Think of hereditary risk as 'shove' mutations versus 'nudge' mutations.

  BRCA1/2 are shoves: they push you off the cliff at roughly 9–10× baseline risk of early breast cancer. Polygenic nudges each add a little, combining with environment until you tip over. Because the categories overlap, quantifying where someone sits could route them into prevention programmes or trials.

• 04

  Healthy tissue is full of dormant 'sleeper' cells carrying cancer mutations; inflammation can be what wakes them.

  Biopsy every lung cell in two non-smokers and you'd likely find thousands of cells with cancer-hallmark mutations, citing work including Charlie Swanton's lab. The real difference between people may be a particular inflammatory response converting one sleeper into a tumour. Mukherjee stresses 'inflammation' is still a loose word, like 'energy' before Newton.

• 05

  Chronic 'insulinemic tone' sustains cancers and their resistance to therapy—shown for treatment, hypothesised for causation.

  His Nature paper with Lou Cantley found in animals that insulin and its downstream cascade, not obesity itself, drove resistance to therapy. He chooses words carefully: extending this to cancer initiation is an extrapolation and a 'fussy' but doable experiment nobody has run. Attia cites Mitch Lazar's two-by-two showing cardiometabolic disease tracked hyperinsulinemia rather than adiposity.

• 06

  Osteoarthritis is not simply an obesity or wear-and-tear disease; it reflects inflammation suppressing cartilage stem cells.

  Mukherjee's lab discovered the first osteoarthritic stem cells. Cartilage is constantly made and destroyed; disease appears when genesis falls while destruction holds or rises. Lean women get it in non-weight-bearing hands and necks. In young mice, injecting an immune activator such as LPS into a joint stops stem cells making cartilage and almost exactly mimics osteoarthritis. Why stem cells respond this way is unknown—'teleology is not biology.'

• 07

  Tumours carry many recognisable neoantigens, but dozens of 'don't eat me' checkpoint flags keep T cells away.

  Attia cites Steve Rosenberg: ~80% of solid tumours have neoantigens. Mukherjee argues the ~100 point mutations don't explain this; Tyler Jacks' lab shows dysregulated gene decoding produces cryptic proteins as the real source. Normal cells display many 'I'm normal' flags, which is why we don't have devastating autoimmunity; cancers that survive evolve the same flags.

• 08

  CAR-T works in blood cancers because it kills an entire lineage we can live without—so solid-tumour CAR-T needs an expendable tissue target.

  CD19 CAR-T kills all B cells, not just malignant ones; humans tolerate that. The same logic could apply to ovary, prostate, endometrium or even pancreas, but not brain. Ovarian cancer (really fallopian-tube in origin) has known antigens and several CAR-T attempts. Pancreas lacks good unique antigens, and most KRAS sits inside the cell, making it a small-molecule target instead.

• 09

  Pancreatic cancer survival depends heavily on anatomy, and new KRAS drugs may be most valuable after surgery.

  Tail tumours are often found incidentally and are anatomically separate; head tumours sit in 'the traffic intersection of everything,' presenting with pain or jaundice. Mukherjee has patients alive 10 years after tail cancers. The KRAS inhibitor daraxonrasib extends life in metastatic disease despite skin side effects; adjuvant trials after a Whipple are 'in the works in 1,000 ways.'

• 10

  Preventing the big killers may not equal longevity; the next frontier is 'regenerative failure.'

  Supercentenarians tend not to die of cancer but of skin, bone, blood and immune failure. Attia counters that atherosclerosis can be made nearly an 'orphan disease' with known drivers—no smoking, insulin sensitivity, very low apoB, treated blood pressure—but cancer can't be eliminated that way. Both agree on a parallel strategy: remove proximal diseases while systematically studying regeneration.

• 11

  Aging science lacks its 'cholesterol'—a validated biomarker—and finding one is a decades-long Manhattan Project.

  Cardiology moved fast once blood pressure and cholesterol served as proxies. A longevity biomarker study could take 15–30 years prospectively, perhaps 15 with AI; Attia suggests mining metabolomes and proteomes of Interventions Testing Program winners. Echoing David Sabatini, they argue funding must anchor in basic science rather than rushing to drug discovery, and likely sits outside the NIH's illness-focused mandate.