Is cancer fundamentally a fungus that antifungal treatment can cure?

The Claim

The claim says that cancer is not fundamentally uncontrolled growth of altered human cells but a fungal process—often described as Candida—and that antifungal treatment can therefore cure cancer.

Some variants interpret white tumor tissue as fungal colonies. Others point to research detecting fungi in tumors or to antifungal drugs that affect cancer cells in laboratory studies.

Where It Comes From

The modern claim is associated with alternative-cancer narratives that reinterpret tumors as fungal infection and promote antifungal or alkalizing interventions. More recently, legitimate research on the cancer mycobiome has given the claim new scientific vocabulary.

That newer research is important, but it asks how fungi living in or around the body may influence tumor biology. It does not identify malignant human cells as fungal organisms.

Why It Sounds Convincing

Fungal infections can form masses, cause inflammation and be difficult to treat. Some cancers occur in tissues where microbial communities are present. Cancer patients can also be unusually vulnerable to fungal infections because of immune suppression or treatment.

If fungal DNA is detected in a tumor, it is easy to slide from “fungi are present and may matter” to “the tumor is fungus.” Those are different biological propositions.

The Short Answer

Cancer is not fundamentally a fungal disease. Malignant tumor cells arise from human cell lineages with human genomes and acquired alterations. Some studies have detected fungi associated with tumors and mouse experiments suggest that certain fungi can affect pancreatic cancer progression.

The size and reproducibility of those fungal signals are actively disputed, including in direct analyses of the same sequencing datasets. That is important research, but the claim that malignant human cells are themselves fungi—and that antifungal medicines generally cure cancer—does not follow from it.

How It Would Have to Work

If tumors were fundamentally fungal, malignant tissue should be primarily composed of fungal cells or derive from fungal replication. Its genome, cell structures, lineage markers and behavior should match a fungal organism rather than the patient’s altered tissue.

Antifungal treatment should also produce reproducible cancer remission by eradicating the causative fungus across different tumor types.

What Actually Happens

Cancer begins when human cells acquire or maintain changes that allow abnormal growth and survival. Tumor cells can be traced to their tissue lineage and studied through human chromosomes, mutations, gene expression and cell-surface markers.

Microorganisms can coexist with tumors. Bacteria, fungi and viruses may alter inflammation, metabolism, immunity or drug response. A microbe can therefore contribute to carcinogenesis or tumor behavior without being identical to the cancer cell itself.

What the Evidence Shows

Cancer cells have human genomic and tissue-lineage identities. A tumor may contain human immune cells, bacteria and fungi without its malignant cells becoming any of those organisms.

The strongest reason to take the fungal contribution seriously is experimental rather than rhetorical. Aykut and colleagues reported in Nature (2019) that human and mouse pancreatic tumors had about 3,000 times more fungal material than normal pancreas in their studied samples. In mouse models, reducing fungi impeded pancreatic tumor growth and reintroducing a Malassezia species promoted it. The experiment concerns a possible contribution to tumor development in mice, not a clinical trial curing human cancers with antifungal drugs.

A second original study in Cell (2022) examined 17,401 patient tissue, blood and plasma samples across 35 cancer types. Its researchers reported fungal DNA and cells at low levels in some tumors, including supporting microscopic staining, while acknowledging the need to distinguish biological signal from contamination.

These positive reports are not uncontested. Fletcher and colleagues reanalyzed the human pancreatic sequence data used by Aykut, including five normal-pancreas and thirteen pancreatic-cancer tissue samples. With a different processing pipeline they found too few reliable reads to reproduce the claimed tumor-specific pattern. The original research group published a reply disputing the analysis choices. These are competing interpretations of the same original human sequence data, not a positive study and a wholly independent newly recruited negative patient cohort.

A larger computational reassessment by Ge and colleagues, published in Science Translational Medicine in 2025, examined 5,734 cancer-genome sequencing datasets from The Cancer Genome Atlas (TCGA) and identified substantially fewer confidently assigned microbial reads than earlier reports. Of those samples, 4,271 overlapped the TCGA subset used in the 2022 Cell work. That makes it a meaningful methodological challenge, but not an independent patient cohort that simply erases all microscopy findings.

The remaining question in these studies is how much fungi influence particular tumors, not whether the malignant cells are fungi. Antifungal medicines have established uses in preventing or treating fungal infections in patients with cancer; such benefits do not show that the medicines cure the cancer itself.

The Strongest Evidence for the Claim

A serious supportive case must include the Nature 2019 pancreatic study: it did more than detect fungal sequences. Researchers altered fungal abundance in mouse tumor models and observed changes in progression; one Malassezia species was implicated in the experimental effect. The 2022 Cell study also reported fungi in material collected from several cancers, including microscopic validation in selected tissues.

That evidence can support the narrower hypothesis that a fungus sometimes affects a human tumor’s environment or growth. It does not establish that the malignant cells are fungi, that all cancers share one fungal cause or that antifungal treatment generally cures malignancies in patients. A mouse tumor-growth experiment and a human clinical remission trial measure different endpoints.

The Strongest Evidence Against It

Human cancer cells can be traced to their human tissue lineage, chromosomal material and acquired mutations. Fungal organisms have different genomes and cellular identities. Detection of fungal DNA beside a human malignant cell does not reclassify that cell as fungus.

Moreover, even the proposed fungal cofactor is methodologically contested. Fletcher’s 2023 reanalysis did not reproduce the human pancreatic fungal signal from Aykut’s dataset, and the original authors disagreed about analytic assumptions. A 2025 TCGA reanalysis independently challenged many microbial sequence assignments, with partly overlapping specimens. The critical studies do not demonstrate that fungi never occur in tumors; they show why contamination, sequencing pipelines and independent tissue validation matter.

Neither side of this microbiome dispute supplies controlled evidence that broadly administering antifungals eradicates human malignant cell populations. The established clinical indication for antifungals in oncology—managing fungal infection—is not the same outcome as curing a tumor.

Common Viral Arguments

“Candida has been found inside tumors.”

Presence does not establish identity. A tumor can contain immune cells, bacteria and fungi without being made from those organisms.

“Antifungals can kill cancer cells in a dish.”

A drug can have off-target or additional biological effects. In-vitro activity does not establish that cancer is caused by fungus.

“Cancer looks white like Candida.”

Visual similarity is not a biological classification. Histology, molecular markers and genomes are more probative than color.

“Cancer patients get fungal infections.”

Yes, particularly when immunity is impaired. That is one reason antifungal therapy is important in oncology, but it does not make the malignancy fungal.

The Reasoning Error

The central problem is a category error: microbes associated with a tumor are treated as identical to malignant cells. Cherry-picking occurs when mycobiome papers are quoted without their distinction between host tumor and microbial community.

The claim can also neglect mechanism by skipping the large body of cellular and genomic evidence that identifies the tumor’s lineage.

What Would Change Our Conclusion?

The conclusion would change if independent pathology and genomic studies showed that what is currently classified as malignant tissue across cancers is actually fungal tissue, and if fungal eradication reliably produced cancer cures in controlled clinical studies.

Discovery that a specific fungus causally contributes to a specific cancer would be important but narrower. It would not by itself establish that all cancer is fungus.

Evidence Status

Status: Contradicted
Evidence quality: High

Overall: Contradicted for the central cellular-identity claim. Distinct propositions have distinct evidence:

Component A — cancer cells are fundamentally fungal rather than altered human cells: Contradicted. Human tumor pathology and genomic analyses identify malignant human-cell lineages.

Component B — fungi can influence some cancers as microenvironment organisms or cofactors: Supported in specific preclinical settings, qualified in human observational material. Aykut 2019 reported mouse-model perturbation effects, while multicancer and pancreatic fungal sequence claims are contested by reanalyses partly using the same original samples. This narrower finding does not rehabilitate Component A.

Component C — ordinary antifungal treatment generally cures human malignancies: Unsupported. Clinical antifungal treatment of fungal infection is a different outcome from eradication of malignant human cells; no broadly generalizable controlled oncology cure result is established by the cited mycobiome studies.

The inference that cancer cells are fungi, and therefore antifungals broadly cure them, fails at its central premise. Whether fungal organisms can contribute to the progression of a specific human cancer, or whether an antifungal improves the outcome of a specific oncology treatment, remains a separate empirical question.

Sources

The original 2019 Nature mouse and pancreatic-tissue study, 2022 Cell multicancer sampling study, 2023 original sequence reanalysis and original-author reply, and the 2025 Science Translational Medicine TCGA reanalysis are each linked to their own structured Evidence. The reuse of human sequencing datasets is made explicit. The U.S. National Cancer Institute (NCI) explains malignant human-cell identity; the Cochrane source addresses fungal infection care, not antitumor efficacy.

Corrections & Updates

October 8, 2026 — Evidence update: added original positive pancreatic and pan-cancer mycobiome studies alongside 2023 and 2025 critical reanalyses, including shared-dataset limitations. Evidence Status unchanged.

Last Reviewed

October 8, 2026.