When Keto Works and When It Doesn’t - A Cancer-by-Cancer Guide

In my How to Starve Hormone-Positive (ER+) Breast Cancer on Tamoxifen: I made the case for why I question ketogenic diet as a default for ER+ breast cancer, and why intermittent fasting plus metformin is the better lever, anchored in the Caffa/Longo Nature 2020 trial and the Elgendy/Cazzoli PP2A-GSK3beta-MCL-1 mechanism.

This piece is the practical follow-through on specific questions:

• for which cancers does strict ketogenic diet genuinely earn its place,

• for which is it actively harmful, and

• for which does running a full metabolic cocktail change the answer?

Part 1B of this series covers the seven-lever stack and the four-question decision logic in full.

I want to be specific about this because I do not want to be misread as anti-keto. Ketogenic diet is a precision tool. The problem is not the tool. The problem is that it is being applied without precision.

Glioma and glioblastoma. This is the one indication where the case is unambiguous and I will defend it without qualification. Glioma cells are largely OXCT1-deficient and glycolysis-addicted. Beta-hydroxybutyrate is not just non-fuel for them; at sufficient concentration it is directly cytotoxic to glioma cells while being neuroprotective to normal brain tissue. KD radiosensitises glioma. The ERGO2 trial and multiple case series support this. The 2026 Seyfried/Mukherjee Cell Rep Med paper showed KD combined with mebendazole and devimistat produced the greatest reductions in tumour invasion and prolonged survival in both VM-M3 and CT-2A juvenile glioma models, supporting further investigation of KD as a metabolic vehicle for drug sensitisation specifically in glioma. This is where Seyfried’s lifetime of work is most clearly correct.