What Actually Causes Cancer — Dr. Moritz Przybilla on What the Internet Got Wrong
Few health claims travel as fast online as the idea that cancer is really a metabolic disease — something you can starve, hack or reverse with the right diet. Dr. Moritz Przybilla spends his days in cancer genomics at the Wellcome Sanger Institute, and his argument is that decades of sequencing data point somewhere else entirely. This is an episode about what the evidence actually supports, and about why the most popular version of a story is so often the wrong one.
Cancer is, fundamentally, a disease of DNA
Przybilla's starting point is that cancer begins when mutations accumulate in the genome of a cell until its growth controls fail. Metabolic changes are real and they matter — but in the genomic picture they are largely downstream consequences of the mutations, not the initiating cause. Getting the direction of causation right is not a semantic point: it determines which interventions could plausibly work.
Driver mutations versus passenger mutations
Not every mutation in a tumour is doing something. A small number are drivers that actively push the cell toward malignancy; the great majority are passengers that came along for the ride. Przybilla uses the distinction to explain why tumour sequencing is harder to interpret than headlines suggest, and why counting mutations tells you much less than identifying which ones are steering.
The metabolic theory — and where it breaks
Thomas Seyfried's metabolic theory of cancer gets a full hearing rather than a dismissal. Przybilla grants what is true in it — tumours do rewire their metabolism, and the Warburg effect is real — while explaining why that does not make metabolism the root cause. The evidence he leans on is the genomic record itself: the mutational patterns found across thousands of tumours are hard to reconcile with a purely metabolic origin.
Ketogenic diets and the sugar question
The practical version of the metabolic theory is that sugar feeds cancer and a ketogenic diet can starve it. Przybilla's read of the trial evidence is that keto has not been shown to treat cancer, and that the sugar claim badly misrepresents how glucose metabolism works in the body. He is careful about tone here: the interest in diet is understandable, the mechanism is not what people think, and overselling it costs patients real time.
Why aging is the largest risk factor
Healthy tissue quietly accumulates mutations across a lifetime — this is normal biology, not disease — and the longer you live, the more chances there are for the wrong combination to appear in the wrong cell. That is why age dominates cancer risk. It also reframes prevention: some fraction of cancer traces to modifiable exposures, but a substantial share reflects the arithmetic of time and cell division.
The uncomfortable number: around 60% currently unavoidable
Przybilla defends the estimate that a majority of cancers cannot presently be prevented, and explains why that is not a counsel of despair. It is an argument for shifting effort toward detection and risk prediction rather than promising a lifestyle protocol that eliminates risk. He also weighs where something like metformin and the broader slow-aging thesis might fit — if you could genuinely slow aging, you would expect to move cancer incidence with it.
Detection, prediction and what to actually screen for
His strongest practical claim is that early detection will save more lives than any near-term miracle cure, and he names which screening tests he thinks are worth doing versus which are marketing. The research direction he is most excited about is predicting risk decades before a tumour appears, by reading the mutational history already written into healthy tissue — which would move oncology closer to the preventive posture this show keeps arguing for.
Key takeaways
Cancer is fundamentally a disease of DNA — metabolic changes are mostly downstream, not the root cause. Driver mutations, not mutation counts, determine malignancy. The metabolic theory contains real observations but does not survive the genomic evidence, and ketogenic diets have not been shown to treat cancer. Aging is the dominant risk factor because healthy tissue accumulates mutations over a lifetime. Around 60% of cancers are currently unpreventable, which is an argument for early detection and risk prediction rather than for despair or for lifestyle guarantees.