Coach — Mitochondrial-Metabolic Hypothesis of Cancer

Cancer is usually told as a story that starts with a mutation and ends with a drug aimed at that mutation. The working hypothesis here starts one layer upstream: a cell that has lost clean oxidative power, switches on ancient fermentation as a rescue, and then, inside a mixed population of tumor cells, natural selection favors the clones best at that rescue.

Safety before ideas

This essay is not cancer care. The extended report says so in plain terms: "Cancer / colorectal: This is not oncology or gastroenterology care. Do not delay, replace, or abandon recommended cancer evaluation, colonoscopy, polypectomy follow-up, or treatment because of a hypothesis on this site." It adds that "Surgery, chemo, radiation, targeted agents, and immunotherapy keep their place," and that "Hypothesis language is not a promise that adenomas, CRC, or metastatic disease will reverse."

Several practical cautions follow from that. A genome panel or a single driver mutation is not proof that metabolism is irrelevant, and a metabolic idea is not proof that the panel is irrelevant. Unsupervised extreme ketosis, megadose antioxidants, "cancer detox" programs, and supplement stacks are not oncology. Chemotherapy, radiation, immunotherapy, targeted drugs, hormone blockade, and supportive medicines are not stopped because an essay reframed tumor metabolism; the treating team owns those decisions and their timing. Biopsy, staging, surgery, indicated systemic therapy, colonoscopy, and surveillance are not delayed while someone tries to think the case through from a web page. Nothing here promises a cure, the erasure of metastatic disease, or a reset of the colon after a polyp, and nothing here suggests that food, light, sleep, or training replaces oncology. Peptide, hormone, or GLP-1 "cancer" protocols are not offered on this page, and there is no dosing for methylene blue, CoQ10, PQQ, or butyrate. If the question is whether to pause treatment for a metabolic plan, that decision belongs with the oncologist who knows the case.

The standard model and its gaps

The dominant view treats cancer as a genetic disease. Mutations in genes such as KRAS, BRAF, or PIK3CA, the loss of guardians such as TP53 or PTEN, and amplified growth signals give cells an advantage, and the clones that carry them expand. This view is right that cancers are clonal, and its tools have saved lives: genomic profiling, targeted kinase inhibitors, PARP inhibitors, and checkpoint immunotherapy in selected settings. The report respects that record and does not reject it.

Its charge is that the model is incomplete. Older adults carry many somatic mutations in normal-looking tissue that usually never becomes cancer, though such mutations in blood do raise the risk of blood cancer. Some tumors carry remarkably few mutations. Not every carrier of a high-risk BRCA variant develops breast or ovarian cancer, and metabolic and lifestyle factors may change that risk. Obesity, type 2 diabetes, and inactivity raise incidence across many sites, while weight loss and exercise are linked to lower risk. And resistance after an early dramatic response to a targeted drug remains common, often with metabolic rewiring along the way. The genetic lens describes these facts but, in the report's view, does not explain why mutations become fixed in some cells and settings and not in others.

The report also places this in history. The War on Cancer, launched with the 1971 National Cancer Act, and the large sequencing projects that followed catalogued thousands of mutations and enabled precision treatment for some patients. Yet most common solid tumors, once metastatic, are still not cured by genetic targeting alone, and prevention has lagged. The report's worry is that powerful AI tools applied to the same genetic-first assumptions may meet the same ceiling, unless energy thresholds and selection enter the model.

One layer upstream

The proposal begins with mitochondria. In precancerous or early transformed cells, oxidative phosphorylation and ATP production falter. The cell falls back on glycolysis, an ancestral fermentation program suited to low oxygen and unreliable fuel. Inside a tumor, cells differ genetically, epigenetically, and metabolically, and natural selection acts on that variety. Subclones that are best at grabbing carbon, running glycolysis fast, and building biomass win, and the Warburg phenotype gets locked in. Mutations, in this account, arise largely from oxidative damage and replication stress, and they persist when they help the glycolytic state. Mutations matter. In this hypothesis they often arrive after, and stick because of, the energy crisis.

Body-wide metabolism shapes that field. High insulin and IGF-1 add growth and survival signals and more available glucose. Chronic inflammation adds oxidative stress. Obesity creates inflamed, low-oxygen fat tissue that stresses neighboring cells. Aging, the dominant risk factor, brings slower mitochondrial renewal and cleanup, more damaged mitochondrial DNA, and less metabolic flexibility. The report treats this as one upstream defect that accounts for many dashboards: variable penetrance, low-mutation tumors, metabolic risk factors, and resistance read as ongoing adaptation under selection.

The report's practical surrogates for systemic mitochondrial capacity are ordinary ones: grip strength, cardiorespiratory fitness, the trend in lean mass, recovery after effort, sleep regularity, and insulin tone. They sit beside staging, histology, molecular profiling, tumor markers, and imaging, which belong to the treating team. A better scan on a background of collapsing fitness and sleep is, in this view, an incomplete picture of the person.

The colon as a clear example

Colorectal cancer gives the hypothesis its clearest gut-level example. A 2026 Harvard study in Cell Host & Microbe (Nogal and colleagues) looked at stool from 354 women about 12 years after a precancerous polyp was removed. Their gut bacteria and stool chemistry still differed from controls and partly matched profiles seen in colorectal cancer, and the differences were larger in women with poorer diets and less activity. One of the depleted species, Faecalibacterium prausnitzii, is a keystone producer of butyrate, the fatty acid that colon cells burn as their main fuel. A sulfide-producing species, Bilophila wadsworthia, can inhibit mitochondrial respiration.

The report reads the polyp as a warning light on the dashboard: a visible, focal exaggeration of a wider field in which colon cells, their mitochondria, and their microbes are under energy stress. Removing the polyp clears the light. It does not, by itself, repair the field. The report is careful about what follows from this. Screening and polypectomy "demonstrably reduce CRC incidence and mortality by interrupting the sequence at visible stages; they remain necessary." And it states the boundary directly: "This does not authorize skipping colonoscopy, surveillance intervals, or oncology follow-up." The deeper science is laid out in the report's sections on the adenoma as a dashboard and the colon cell, mitochondria, and microbiome axis. The same partnership is described from the gut side in the GI essay and from the inflammation side in the immune essay.

Where restoration is discussed, the report frames it as orientation within its three pillars, coordinated with gastroenterology and oncology. Fuel means nutrient-dense whole foods, traditionally fermented foods and fermentable carbohydrates that feed butyrate producers, introduced slowly, with less refined carbohydrate, fewer industrial seed oils, and fewer ultra-processed foods. Under medical supervision and in appropriate contexts, the report says, time-restricted or ketogenic approaches "may exert selective pressure against glycolysis-dependent cells", while watching for different effects on primary and metastatic disease. Circadian alignment means morning outdoor light, consistent sleep, dim evenings, and eating windows aligned with daylight. Strength means progressive resistance and aerobic work. The report insists these are not a home protocol. Where stool panels, fasting insulin, hs-CRP, HbA1c, sleep recovery, or grip are available, they are read as field readings beside colonoscopy, never as substitutes for it, and no home test is proof of tumor bioenergetics. The general measures of capacity are gathered on the common-ground page.

The ketogenic diet paradox

The report uses one observation to block two opposite cartoons. On the report's hypothesis, starting a ketogenic diet may shrink a primary tumor, as glucose-dependent cells lose fuel and insulin signaling, while metastatic behavior shifts or energy-adapted subclones persist. Mouse evidence so far is mixed. The first cartoon says ketosis shrinks everything everywhere, so metabolic change is a cure. The second says metabolism is irrelevant because cancer is genetic. Selection predicts neither: changing fuel pressures some clones and spares others. The report labels this plainly as "hypothesized mechanism language, not a retail keto-for-cancer order set".

That paradox also sets the order of questions. First, acute and specialty-critical problems go to oncology and gastroenterology: suspicious findings, staging, surveillance, and treatment. Next comes the everyday ground of food, light, sleep, and training, coordinated with the treating team during active therapy or surveillance. Then capacity is tracked beside whatever dashboards the physicians already run. The report's own ordering is "integrative and pro-root-cause, not anti-oncology", and it repeats the line that matters most: "None of this authorizes delaying biopsy, staging, surgery, chemo, radiation, indicated systemic therapy, or colonoscopy/surveillance from a webpage." Two opposite errors are worth naming. One chases panels and markers while sleep and strength collapse. The other rejects indicated oncology or skips colonoscopy out of ideology. Both treat one dashboard as the whole person.

Related commentaries

Two shorter commentaries sit beside the main report. The first concerns an October 2026 KAIST experiment in which colon-cancer cells were pushed toward a more normal-looking, less proliferative state by keeping three regulators, MYB, HDAC2, and FOXA2, switched off before implantation. The report reads that as a forced phenotype under a clamp, not proof that the tissue's defended set point moved, and notes that durability was not measured. It is useful for keeping "conversion" headlines from replacing questions about energy and selection. It is not a treatment protocol. The full commentary is Forced Differentiation Is Not a Moved Set Point.

The second commentary, from June 2026, is a hypothesis, not a finding. The extended report states it this way: "Hypothesis (June 2026): clinical observers including Dalgleish describe long-remission cancer patients relapsing or progressing aggressively after COVID-19 mRNA boosters. The essay reframes that pattern as an iatrogenic metabolic stressor — proteomic load from spike translation, spike- and adjuvant-related repair cost, and the energetic cost of sustained immune activation — that can exceed mitochondrial reserve in susceptible people, driving Warburg-like glycolytic rescue, T-cell exhaustion with loss of surveillance, and selection for carbon-optimized clones." It is not medical advice and not a treatment protocol, and any question about vaccination or cancer follow-up belongs with the physicians who know the case. The full essay is mRNA Vaccine-Associated Cancer Relapse and Progression. A broader companion explains when fermentation is an acute rescue and when it becomes a chronic lock: Warburg and Crabtree effects in acute and chronic compensation.

Limits and where to read more

This is a working hypothesis. The report is explicit that it is "Not a randomized-trial verdict that mitochondrial-first strategies outperform guideline oncology as monotherapy," and that it "Does not claim ketogenic or metabolic interventions cure cancer, erase metastatic disease, reverse every adenoma field, or replace oncology / colonoscopy clocks." It does not deny mutation biology, genomic profiling, targeted drugs, immunotherapy, or standard treatment when indicated. Established cancer, hereditary syndromes, and active treatment remain high-stakes, and they stay with specialists. The model is also testable: prospective studies of mitochondrial markers and later cancer risk, trials of mitochondria-supportive care added to standard care in high-risk groups, and post-polypectomy cohorts that track stool and insulin measures beside recurrence would all strengthen or weaken it.

Cancer topics elsewhere on the site continue the map. Skin cancer is discussed in the dermatology report. Sulindac has its own essay. Breast cancer has its own essay too: A Metabolic Hypothesis of Breast Cancer. The two commentaries above are linked again here: Forced Differentiation Is Not a Moved Set Point and mRNA Vaccine-Associated Cancer Relapse and Progression. For the full argument, read the extended report on cancer. The short version is in the lessons. Talk with your own physician before changing medications or supplements.

A working hypothesis, open to correction and revised as new evidence arrives. Last revised November 1, 2026.

See the notice at the top of this page. Working hypothesis, not medical advice.