Ivermectin and Cancer: The Laboratory Evidence, Human Data, Risks and Current Trials

A familiar drug, an intriguing laboratory finding and the hope of a less burdensome cancer treatment: it is easy to understand why ivermectin attracts attention. The question deserves a careful answer. Dismissing the research would miss real scientific work. Presenting it as an established cancer treatment would go well beyond what that work shows.

Evidence at a glance

Evidence level	What is available	What it can tell us
Cells, organoids and animals	Multiple cancer models, proposed signaling and immune effects, and combination experiments.	A biological rationale for testing; it cannot establish a safe, effective human treatment.
Prospective human treatment data	A 2025 phase I/II conference abstract: nine patients with metastatic triple-negative breast cancer received ivermectin plus an anti-PD-1 drug.	An early safety and response signal from a very small, uncontrolled combination study.
Case reports	Small reports involving multiple drugs, including a 2026 melanoma case with an initial response followed by worsening.	Observations worth investigating; they cannot identify which intervention caused a change.
Publication concerns	A widely circulated 197-person cohort was retracted; a prominent breast cancer mouse paper carries an editorial expression of concern.	These reports require explicit qualification and cannot support confident treatment claims.
Routine cancer care	No established oncology dose or demonstrated survival benefit.	Use for cancer should be confined to appropriately designed clinical research.

What ivermectin is—and why it is being studied

Ivermectin is an antiparasitic medicine with established human uses, including treatment of strongyloidiasis and onchocerciasis. It is not a nutraceutical. Drug repurposing asks whether an existing medicine can do something useful beyond its original indication. Its known manufacturing and pharmacology can help researchers, but safety for intermittent parasite treatment does not automatically establish safety for repeated oncology dosing or combinations.

Laboratory studies describe several possible actions: altering chloride movement and oxidative stress, changing growth signaling, interfering with nuclear transport, affecting mitochondria, and modifying tumor–immune interactions. These are different hypotheses in different models, rather than proof of a single universal cancer mechanism.

A map of the preclinical evidence

Model and primary research	Main finding	Important boundary
Acute myeloid leukemia, 2010	Leukemia cell lines and patient-derived samples showed cell death at low micromolar concentrations. Mouse models showed delayed tumor growth; experiments implicated chloride flux and reactive oxygen species.	Patient-derived samples were studied outside the body. This was not treatment of people with leukemia.
WNT-dependent cancers, 2014	Ivermectin inhibited WNT–TCF signaling and growth in selected models, including xenografts.	TCF-independent xenografts did not show the same response: the effect was context dependent.
Epithelial ovarian cancer, 2017	Experiments implicated KPNB1, a nuclear-transport protein; ivermectin suppressed growth in laboratory models.	Target discovery and animal activity do not establish effectiveness in patients.
Esophageal squamous cancer, 2020	Cell and mouse experiments reported reduced growth and metastatic behavior.	The models do not reproduce the full complexity of human disease, exposure or treatment toxicity.
Breast cancer and immunotherapy, 2021	Mouse experiments reported immune changes and synergy with checkpoint blockade.	This paper carries a 2026 editorial expression of concern, discussed below.
Pancreatic cancer, 2022	Combining ivermectin with gemcitabine increased anticancer effects in preclinical models, with mitochondrial dysfunction and apoptosis implicated.	This is not a pancreatic cancer clinical trial and does not justify adding ivermectin to gemcitabine outside research.
Gastric cancer, June 2026	One cell line and two patient-derived organoids responded; combination effects with cisplatin were reported.	Another cell line showed no significant viability effect. Some signaling effects weakened or rebounded over time. Organoids are not treated patients.

The broader literature also includes breast, lung, prostate, brain and other tumor models. A 2020 review maps proposed pathways including PAK1/AKT/mTOR, WNT signaling, mitochondrial stress and drug resistance. A pathway appearing on a long list does not tell us whether it can be altered safely enough, for long enough, to benefit a person.

A June 2026 colon-cancer study adds a metabolic combination: ivermectin plus recombinant methioninase, an enzyme that depletes methionine. It produced a stronger cell-viability effect than four of five other drug–enzyme combinations tested. However, the experiment used one selected cell line, HCT116, for 72 hours; ivermectin’s half-maximal inhibitory concentration was 7.1 micromolar. Its “chemosensitivity index” was a laboratory ratio, not a patient response rate. The authors identify the single-model design as a limitation. My interpretation is that this supports further research; it does not show superiority to clinical colorectal-cancer regimens or establish that dietary methionine restriction reproduces the enzyme experiment.

The exposure problem: a dish is not a bloodstream

Many experiments expose cells directly to micromolar drug concentrations. Human exposure depends on absorption, protein binding, metabolism, tissue distribution and treatment duration. Nominal concentrations in a culture dish cannot simply be converted into a tablet dose. Mouse dosing, injected formulations, nanoparticles or drug derivatives are also not interchangeable with standard oral tablets.

For scale, the oral-product label reports mean peak blood concentrations of 30.6 and 46.6 ng/mL after a single 12-mg dose in fasting healthy volunteers. Converting the major component’s molecular weight gives approximately 0.035–0.053 micromolar. That is well below many micromolar cell experiments. This is an illustrative comparison, not a dose calculation: total blood levels, unbound drug, tumor exposure and concentrations bathing cultured cells are different measurements. Trials need to establish whether a biologically active tumor exposure can be reached safely.

Some investigators argue that relevant exposures may be achievable. That remains a pharmacology question to test, not a reason to escalate doses independently. A 2002 dose-escalation study in 68 healthy volunteers explored substantially higher exposures over short periods. It cannot establish the safety or effectiveness of months of treatment in patients with cancer, impaired organ function or interacting medicines.

What has actually been reported in people?

The breast-cancer study is the clearest prospective clinical experience located in this review. A 2025 ASCO meeting abstract described ivermectin combined with the PD-1 antibody balstilimab in metastatic triple-negative breast cancer. Nine people enrolled and eight were evaluable: one had a partial response, one had stable disease and six had progressive disease. Median progression-free survival was 2.5 months; the four-month clinical-benefit rate was 37.5% with a very wide 95% confidence interval (15.3%–91.7%).

The protocol explored oral ivermectin dose levels of 30, 45 and 60 mg on days 1–3, 8–10 and 15–17 of a 21-day cycle. These are research details, not a recommended cancer regimen. With only eight evaluable participants, no ivermectin-only group and checkpoint immunotherapy given to everyone, the abstract cannot show how much—if any—benefit came from ivermectin. Full-text appraisal was not possible because only the conference abstract was located.

Reported treatment-related events in the abstract included rash, gastrointestinal symptoms, altered taste, weakness, hypothyroidism and grade 3 anemia. The sample was far too small to estimate uncommon serious harms. The corresponding official NCI record for NCT05318469 now says withdrawn.

A 2022 three-patient case series described advanced cancers treated with combinations including ivermectin and dichloroacetate, with other drugs also involved. Reported symptom or marker changes cannot identify the contribution of ivermectin, establish a survival benefit or rule out selection and reporting bias.

A 2026 melanoma case report described initial imaging and circulating-tumor-DNA improvement during self-directed ivermectin, fenbendazole and lifestyle changes. Disease subsequently worsened. The authors considered spontaneous immune-mediated regression at least as plausible as a treatment effect. This is a useful reminder that an initial temporal association is not proof of causation.

Human experience therefore exists, but its evidentiary weight is limited. I did not identify a completed, adequately controlled trial demonstrating improved survival or a reliable added anticancer benefit. That is different from proving ivermectin can never work; it describes the gap that clinical trials still need to close.

Two publication updates that change the conversation

The 197-patient ivermectin/mebendazole observational cohort that circulated widely in 2026 is now marked retracted; see the retraction notice. Its response and survival claims should not be used in decision-making. Retraction is a material reversal in the usable human evidence, not a minor caveat.

Separately, the 2021 breast cancer immunotherapy paper received a correction and then an editorial expression of concern on June 5, 2026. The notice describes additional highly similar images after an earlier correction and states that original data were no longer available. An expression of concern is not a retraction, but it materially lowers confidence in the affected findings. Independent replication is particularly important.

Current human-trial status

ClinicalTrials.gov and the NCI trial record were checked again on October 23, 2026. Registration is not proof of efficacy, and status can lag. The previously listed metastatic triple-negative breast-cancer study, NCT05318469, is now officially withdrawn. No actively recruiting anticancer ivermectin trial was verified in this search.

Study or identifier	Verified status	What it means
NCT05318469	Withdrawn	The 2025 abstract reported eight evaluable participants treated with ivermectin plus balstilimab. The study did not establish ivermectin benefit and is not recruiting.
NCT07487805 / “ICONIC”	Not verified	No corresponding ClinicalTrials.gov record could be confirmed. It is excluded rather than treated as an active study.
NCT02366884	Older multiagent study	A repositioned-drug combination cannot isolate ivermectin’s effect and does not establish efficacy.

The preliminary breast-cancer result remains human evidence worth tracking, but the study’s withdrawal, tiny sample and combination design sharply limit inference. Anyone offered a “trial” should verify the identifier directly in an authoritative registry, confirm an accountable study site and investigator, and review consent and oversight documents.

A circulated listing for an “ICONIC” dose-ranging study used the identifier NCT07487805. Direct registry checks did not confirm that record. It has therefore been removed from the active-trial discussion. An unverified webpage or copied identifier is not evidence that a regulated human trial exists.

The older NCT02366884 study tested multiple repurposed agents in glioblastoma, so even posted outcomes would not isolate ivermectin. Other registry records involving ivermectin for infections or COVID-19 are not cancer-treatment evidence.

Potential harms need equal space

Ivermectin’s established role in parasitic disease does not make it harmless. The official oral-product label includes reports of serious neurotoxicity, including altered consciousness, confusion, stupor, coma and death, after recommended dosing as well as overdose. Dizziness, gastrointestinal effects, hypotension and liver abnormalities also matter in patients already managing treatment side effects.

It is primarily metabolized through CYP3A4, and transport proteins also influence exposure. A pharmacist should review the entire medication list, including anticancer drugs and supplements. Rare increased-INR reports with warfarin are described in the label. A high-fat meal can increase exposure; changing food intake to “boost” absorption is not a safe substitute for a studied protocol.

The relevant cancer protocol used organ-function and safety screening, underscoring why a research regimen cannot be translated into self-treatment. There is no validated cancer dose, and dose escalation can disproportionately increase neurologic and hepatic risk. A pharmacist should review CYP3A4 and transporter-mediated interactions across the entire medication and supplement list.

There are oncology-specific warning signals. A 2026 metastatic breast cancer case report describes life-threatening neurotoxicity after off-label ivermectin use. A 2025 pediatric oncology report also raises concern about severe toxicity in the setting of concurrent cancer therapy. Individual cases cannot estimate frequency, but they show that serious harm is possible.

A 2026 liver-injury report followed combined veterinary fenbendazole and ivermectin use for prostate cancer. Because both were taken, it cannot assign the injury to ivermectin alone. It does illustrate the hazards of unmonitored combinations. Veterinary preparations are not substitutes for human medicines, and products bought outside regulated channels introduce additional uncertainty.

The indirect harm can be just as consequential: postponing an effective treatment, missing a trial window, spending money on unsupported regimens or stopping prescribed therapy when an early scan appears encouraging. New confusion, seizures, fainting or severe vomiting after exposure warrants urgent medical assessment.

What would make the evidence more convincing?

I would look for reproducible preclinical work, measured drug exposure in people, a tolerable regimen and a controlled trial in a defined cancer population. The trial should separate the contribution of ivermectin from the accompanying therapy and measure outcomes patients feel or live: durable tumor response, symptoms, function, progression and survival. Publication of negative results and meaningful follow-up are part of that standard.

ASCO’s May 2026 clinical notice recommends against using ivermectin or fenbendazole as cancer treatments outside a clinical trial. That position reflects the absence of reliable efficacy evidence, the risk of serious harm and the danger of delaying treatments known to improve outcomes.

Questions to bring to the oncology team

• Is the evidence being discussed from cells, animals, a case report, an abstract or a controlled human trial?
• Does the study match my cancer type, stage, prior treatment and current medicines?
• Could an apparent response be explained by another treatment given at the same time?
• Is there a legitimate trial for which I might qualify, and what are its actual endpoints?
• Could this interact with my medicines or increase neurological, liver or bleeding risks?
• If I have already taken it, what details about the product, amount and timing should I share?

There should be room to bring this question into an oncology visit without embarrassment. Curiosity is appropriate. So is asking for evidence strong enough to support the decision in front of us.

How this review was assembled

This is a broad narrative review, not an exhaustive systematic review. It draws on primary laboratory papers, the available human reports and 2025 meeting abstract, official registry records, drug labeling, professional guidance and published toxicity cases. It explicitly excludes the retracted 197-person cohort from usable efficacy evidence and does not treat an unverified registry identifier as a clinical trial.

Evidence checked and trial status verified October 23, 2026. Educational information, not individual medical advice.

Browse the article library · Missing Script podcast — coming soon