Clival chordoma research
Home › Diet, supplements and claims

Ketone utilisation, metabolic adaptability and immunogenicity

Agent report for Kya, 14 September 2026, written after her partner asked in em-research: "something that worries me is the ability of brain cancers to utilise ketones... Are these sarcoma like tumors more adaptable in terms of their metabolic options? Also how fast do they tend to switch tactics and is there anything on making these tumors more identifiable to Ems own immune system... this is why I'm concerned about opiates lowering her innate immune system." Working frame: probable clival chordoma, chondrosarcoma the main imaging lookalike, histology pending, 51 kg, and currently eating a deliberately ketogenic diet. Every PMID verified through NCBI E-utilities on 14 September 2026. No survival figures, by rule.

Evidence levels. L1 randomised trial, meta-analysis or guideline. L2 prospective non-randomised trial or large cohort. L3 retrospective series. L4 case reports or very small series. L5 laboratory, animal or cell-line work, and expert reasoning.

This thread changes a standing position in the repo. Until today the file line on ketogenic diets was "no evidence of benefit, and ESPEN advises against restriction in cancer". That was too soft. There is now a mechanism by which it could actively help the tumour, demonstrated in chondrosarcoma specifically, and one animal study in which an unrestricted ketogenic diet shortened survival. Her partner's instinct was right and the file was behind.

Tier A. The short answers to his four questions

  1. Can these tumours burn ketones? Yes, and the most directly relevant paper is in chondrosarcoma. Not a lookalike disease, not an analogy: the main entry on the patient's own differential.
  2. Are sarcoma-like tumours more metabolically adaptable? On the available evidence, yes, and the adaptation is switched on by exactly the conditions a ketogenic or calorie-restricted diet creates.
  3. How fast do they switch? In cell culture, within the span of a serum-starvation experiment, hours to days. In a person, nobody has measured it.
  4. Can the tumour be made visible to her immune system? Yes in principle, and chordoma has a genuinely excellent target in brachyury. It has been tried, in a randomised placebo-controlled trial, and the trial was negative.

And one correction to the thread he drew between them, in Tier E.

Tier B. Ketone utilisation, and the chondrosarcoma paper

The classic premise is that tumours cannot use ketones. The Warburg-derived argument, as popularised by Seyfried, is that cancer cells have damaged mitochondria, depend on aerobic glycolysis, and therefore starve when glucose is withdrawn and ketones replace it. Normal brain adapts to ketones; the tumour supposedly cannot. That is the entire theoretical basis for ketogenic diets in brain tumours.

The premise is now known to be false in several tumour types, and the closest one to the patient is chondrosarcoma.

Vargas-López 2024 (Heliyon, PMID 38694129, L5) is the paper her partner needs to see. In conventional chondrosarcoma cells:

The authors note the context: conventional chondrosarcoma grows inside a dense cartilaginous matrix, an environment already poor in oxygen and nutrients, which selects for exactly this kind of metabolic flexibility. A tumour that evolved to survive starvation is a poor target for a starvation strategy (clinical reasoning, L5).

The glioblastoma evidence points the same way and adds a harm signal. Sperry 2020 (iScience, PMID 32861192, L5, UCLA): fatty acid utilisation was found throughout the GBM metabolome; FAO inhibitors inhibited growth in nearly every patient-derived line tested; knocking down CPT1A, the rate-limiting enzyme of fatty acid oxidation, slowed tumour growth and increased survival in vivo. And then the finding that matters here: "the unrestricted ketogenic diet did not reduce tumor growth and for some models significantly reduced survival." The paper's title states the conclusion plainly: glioblastoma utilises fatty acids and ketone bodies for growth, allowing progression during ketogenic diet therapy.

The clinical trial evidence is small and negative. ERGO2 (Voss 2020, Int J Radiat Oncol Biol Phys, PMID 32619561, L1) is the first randomised trial of a calorie-restricted ketogenic diet with intermittent fasting in brain tumours: 50 patients with recurrent malignant glioma randomised 1:1 alongside reirradiation. The diet worked as a diet, with 17 of 20 completers in ketosis by day 6 and glucose significantly down, and it was well tolerated with modest weight loss. Progression-free survival at 6 months was 20 per cent with the diet and 16 per cent without. No difference in PFS, local PFS or overall survival. An exploratory finding that patients achieving glucose below the median did better is hypothesis-generating and unrandomised within the arm.

Where that leaves the ketogenic rationale. The mechanism it depends on is contradicted in the tumour type closest to the patient's; it failed in the only randomised trial in brain tumours; and in one animal model an unrestricted version shortened survival. That is a worse position than "unproven".

Tier C. Metabolic adaptability and the speed of the switch

Adaptability. The Sperry work is the clearest demonstration that a brain tumour does not have one fuel but several, running fatty acid oxidation alongside glycolysis. The Vargas-López work shows chondrosarcoma adding ketolysis on demand. The general literature on metabolic plasticity describes tumours moving between glycolysis and oxidative phosphorylation according to substrate availability and microenvironment rather than being locked into either (L5).

Chordoma specifically is covered in research thread 21. It is a slow-growing tumour with few mutations, notochordal in origin, and there is no published work at all on ketone utilisation in chordoma. That absence is a gap, not a reassurance. Given that chondrosarcoma, its nearest histological neighbour and its main differential, does this, the prior should be that chordoma can too until somebody looks (clinical reasoning, L5).

Speed. The honest answer has two halves.

The practical implication of the speed finding is uncomfortable. If the adaptation is regulatory rather than mutational, then a ketogenic diet does not need weeks to be outmanoeuvred, and the window in which it might restrict the tumour may be very short or absent. And the substrate itself upregulates the importer and the enzyme, so the diet supplies both the fuel and the signal to build the machinery that burns it (clinical reasoning from L5).

Tier D. Making the tumour visible to her immune system

This is the genuinely hopeful part of her partner's question and chordoma has an unusually good target.

Brachyury is a transcription factor which is overexpressed in essentially all chordomas, is the diagnostic marker for the disease, is largely absent from normal adult tissue, and is associated with chemotherapy resistance and epithelial-to-mesenchymal transition. That combination, near-universal in the tumour and near-absent in the host, is what a therapeutic cancer vaccine wants.

It has been tried properly. Three trials.

  1. GI-6301 phase 1 (Heery 2015, Cancer Immunol Res, PMID 26130065, L2), a heat-killed recombinant yeast vaccine targeting brachyury. Safe, with a signal of clinical activity in chordoma.
  2. GI-6301 randomised phase 2 (DeMaria 2021, Oncologist, PMID 33594772, L1). The first randomised study in chordoma. 24 adults with locally advanced unresectable chordoma, double-blind, placebo-controlled, three vaccine or placebo doses then radiotherapy then continued vaccine or placebo to progression. One partial response in each arm. No complete responses. Median PFS 20.6 months with vaccine and 25.9 months with placebo. Hazard ratio 1.02 (95% CI 0.38 to 2.71). Negative.
  3. BN-Brachyury phase 2 (Cote 2024, Cancer, PMID 38985843, L2), MVA-BN-Brachyury priming plus fowlpox-Brachyury boosting with radiotherapy, 29 patients with advanced chordoma. Two durable partial responses, 80.8 per cent stable disease, median PFS not reached. Did not meet its predefined goal of four responses. The authors note that in an ultra-rare disease with low response rates, a randomised study or novel trial design will be needed to confirm any activity.

Checkpoint inhibitors. Wang 2022 (Cancers, PMID 36612259, L3 systematic review of 22 studies, all 2015 to 2022, mostly case reports and small series). PD-1/PD-L1 inhibitors, most often pembrolizumab, are the commonest agents used. The review recommends PD-1/PD-L1 monotherapy over combinations in advanced chordoma, notes combinations are more toxic without demonstrated added benefit, and observes that PD-L1-negative patients may still benefit. This is salvage-setting evidence in refractory disease, not first-line.

Honest position. The idea is sound, the target is unusually good, real trials have been run, and the one that was randomised did not show benefit. Immunotherapy in chordoma is a live research field and not yet a treatment. It is for later and it is not an alternative to surgery and proton therapy, which remain the only interventions with established effect on this disease.

Tier E. The thread her partner drew between opioids and immunogenicity, and the correction

Her partner: "this is why I'm concerned about opiates lowering her innate immune system."

The link is reasonable but the arms do not match, and the distinction is worth having.

So morphine suppressing NK function is not the reason a brachyury vaccine would or would not work. They are different arms of the immune system doing different jobs (clinical reasoning, L5).

Where the two do meet, and it is worth stating because it is the version of his point that survives: NK cells are the main defence against circulating tumour cells, which is exactly what the MADB106 model in thread 13 measures, and NK function matters around surgery. So the opioid concern is genuine and its natural home is the perioperative window, which is already question 21 on the appointment sheet. It is not an argument about whether her immune system can be taught to see the tumour.

Tier F. What this means for what the patient eats, which is the part that matters tomorrow

She is eating ketogenically now. The food diary records 14 September as a second consecutive clearly ketogenic day, at about 18 g of net carbohydrate. This thread was written because that is a live situation, not a theoretical one.

Separate the two things the food plan is doing.

  1. Getting calories and protein into a 51 kg woman before major skull base surgery. This is evidenced, it is urgent, and it is working: the butter, the cream, the coconut oil and the cheese have been doing the heavy lifting. Nothing in this thread argues against any of that. ESPEN guidance is explicitly against dietary restriction in cancer patients, and being underweight going into surgery is a measurable risk she can actually do something about.
  2. Aiming at ketosis in the belief that it starves the tumour. This is the part that has just got worse, not better. The mechanism it relies on is contradicted in chondrosarcoma, it failed in the only randomised brain tumour trial, and in one animal model an unrestricted ketogenic diet shortened survival.

The recommendation, therefore: keep the fat, drop the goal. A high-fat, high-calorie, low-sugar diet is an excellent way to get 2,200 kcal into someone with a poor appetite, and that is reason enough to continue exactly as she is eating. Stop counting ketosis as a win. Stop treating carbohydrate restriction as treatment. If a piece of fruit or a potato helps her eat more overall, it is now the better choice, not the worse one.

This should be raised with the team rather than decided in the house, and it belongs on Wednesday's sheet alongside the antioxidant question, which is the same shape of problem: a well-meant intervention with a mechanism that might cut the wrong way.

Sources

Research thread 24 · source file em-research/24-ketone-utilisation-and-immunogenicity.md · last edited 14 September 2026