A research group at Complutense University of Madrid in Spain studied cannabinoids in the laboratory. In those studies, cannabinoids killed glioma cells, a type of brain tumour cell, and left normal brain cells alone. Other writers have reported this finding accurately. The problem is a sentence that some writers attach to it: that cannabis cures brain tumours. Frightened patients and their families are the people most likely to read that sentence.
This article covers four things. It sets out what the laboratory showed and what has been tested in people. It explains what current treatment for the most aggressive brain tumour achieves. It also explains what to do if someone sends you a “cannabis oil cures brain cancer” article.
Translator’s note: growers in Thailand are often the first person a family asks for “oil” after a cancer diagnosis. This article gives you the facts you need to answer that request. It also gives you the steps for pointing the family back to the treating doctors.
If you or someone close to you has been diagnosed with a malignant brain tumour, read section 5 first. Nothing here is medical advice, and nothing here is a treatment. Nothing here should delay or replace care from a neuro-oncologist, a cancer doctor who specialises in brain and nerve tumours.
In Thailand, cannabis flower is a controlled herb under the Ministry of Public Health Notification on Controlled Herbs (Cannabis) B.E. 2568, in force since 26 June 2025. It may be sold only against a PT 33 (ภ.ท.33) prescription from an authorised practitioner, covering at most 30 days. Advertising and online sale of flower are prohibited. For an overview of what “medical cannabis” covers in Thailand, see this topic.
1. What the Madrid Laboratory Actually Found
The work comes from the Department of Biochemistry and Molecular Biology at Complutense University of Madrid, a public university in Spain. Cannabis books often quote one passage. It is a fair summary of the group’s own review (Velasco et al., Molecular Neurobiology, 2007). Guzmán, a researcher in the same Madrid group, published a related review in Nature Reviews Cancer in 2003.
In plain words, the group’s summary says this:
- What cannabinoids are. Cannabinoids are the active compounds of the cannabis plant (Cannabis sativa L.). They work by copying endocannabinoids, which are similar signalling molecules the body makes itself. Endocannabinoids switch on specific receptors on the surface of cells.
- Effect in patients. In cancer patients, cannabinoids have several palliative effects. “Palliative” means they ease symptoms. It does not mean they treat the disease.
- Effect in animals. In laboratory animals, cannabinoids slowed the growth of several types of tumour cells, including glioma cells. Glioma is a tumour that starts in glial cells, the support cells of the brain.
- How it happens. Cannabinoids change key signals inside the cell. The main one is endoplasmic reticulum stress, which means the part of the cell that folds proteins becomes overloaded.
- The result. This stress pushes the tumour cell into apoptosis, the cell’s built-in programme for dying. It also blocks angiogenesis, which is the growth of new blood vessels that a tumour needs for its food supply.
- Selectivity. Cannabinoids appear to target tumour cells only. They killed glioma cells but did not kill normal astroglial cells, the healthy brain support cells of the matching type.
The passage is careful in its wording. It talks about tumour cells and laboratory animals. For patients, it claims only symptom relief. It makes no claim about treating brain tumours in people.
The detailed mechanism is covered in the topic on cannabis and cancer. In short, THC (tetrahydrocannabinol, the main intoxicating cannabinoid) acts through cannabinoid receptors. This makes the tumour cell produce ceramide, a fat molecule that works as a signal. Ceramide then triggers the endoplasmic reticulum stress described above.
THC also kills glioma cells through autophagy, a process in which the cell digests its own parts. It does this by blocking the Akt/mTORC1 pathway, a chain of signals that tells cells to keep growing. Glioma is the tumour type with the most preclinical research on cannabinoids. “Preclinical” means work done in cell cultures and animals, before any testing in people.
The selectivity is the important part. Many substances kill cancer cells in a laboratory dish. A substance becomes a possible drug only when it kills cancer cells and spares the matching healthy cells. That is why this work drew attention, and the attention is deserved.
What you can do: when you read or share a cannabis study, find the words that say what was tested: cells in a dish, animals, or people. A result in cells or animals shows where research should go next. It does not tell a patient what to take.
2. The Sentence the Researchers Never Wrote
Some popular cannabis books quote the Madrid passage and add a sentence right beside it. The added sentence says that THC “has been shown to kill glioma cancer cells and is an effective cure”. Some books also say that many patients owe their lives to cannabis. These sentences come from the books, not from the Complutense group. The researchers never wrote them.
The passage next to these claims does not support them, and no other evidence supports them either. The topics on why personal testimony is not medical evidence and on cannabis and cancer (linked in section 1) describe this pattern. A writer quotes a careful scientific statement correctly. Then the writer places an unsupported conclusion in the very next line.
The cure claim comes straight after a real quote, so a worried reader takes it as part of the research. Few readers stop to check where the researchers’ words end and the book’s words begin.
What you can do: when a cure claim sits next to a quote from a study, open the original paper or its abstract and look for the cure sentence. If the researchers did not write it, do not pass it on as their finding.
3. What Has Been Tested in People
Very little has been tested in people. Every human study used pharmaceutical cannabinoids made to medical standards, not home-made preparations.
Three terms help you read the two studies below:
- Phase 1 trial: the first stage of testing in people. It checks safety and dose. It does not test whether the drug works.
- Randomised, placebo-controlled: chance decides which patients get the real drug and which get a placebo, a dummy treatment that looks the same. The placebo group is the control group. Comparing the two groups shows whether the drug made a difference.
- Median survival: the time by which half the patients had died. Half lived for a shorter time and half lived longer.
Guzmán and colleagues, 2006
The Madrid group ran a pilot study (British Journal of Cancer, 2006; 95:197–203). It injected delta-9-THC directly into the tumour of nine patients with recurrent glioblastoma. “Recurrent” means the tumour had come back after treatment. Glioblastoma is the most aggressive type of glioma (see section 5).
The study had no randomisation and no control group. The researchers reported that the treatment was safe. A phase 1 study is designed to show safety, and safety does not prove that a drug works.
Median survival was 24 weeks from the first THC dose (9 patients, not randomised). With no comparison group, this figure cannot show whether THC had any effect.
Twelves and colleagues, 2021
A research team led by Chris Twelves of the University of Leeds, UK, ran a phase 1b trial. It was published in the British Journal of Cancer (2021; 124:1379–87). The team added nabiximols to temozolomide for patients whose glioblastoma had come back for the first time.
- Nabiximols is a prescription mouth spray made from cannabis extract that contains THC and CBD (cannabidiol). See the topic on approved cannabinoid medicines.
- Temozolomide is a chemotherapy drug and the standard drug for glioblastoma.
In the randomised, placebo-controlled part of the trial, 12 patients received nabiximols and 9 received placebo (21 in total). After one year, 83% of the nabiximols group were alive, compared with 44% of the placebo group (p = 0.042).
A p value estimates how likely a difference at least this large would be if the drug had no real effect. Researchers usually call a value below 0.05 “significant”. However, this trial was designed to test safety and tolerability, not survival. The survival comparison is therefore exploratory, and its p value of about 0.04 counts only on paper.
The trial report also notes that two placebo patients died within the first 40 days after joining. The authors themselves called for a randomised trial with enough patients to give a reliable answer.
People quote this result often. If you quote it, always include the numbers: 12 patients against 9. In groups this small, with early deaths in one group, chance alone could easily explain the difference. So could groups that were not alike at the start. It is a signal worth testing further, as the authors said, but it is no basis for treatment.
Researchers in this field describe the anticancer evidence as essentially preclinical in their own reviews. One example is Hinz & Ramer, “Cannabinoids as anticancer drugs: current status of preclinical research” (British Journal of Cancer, 2022). Nobody has yet run randomised, placebo-controlled trials in cancer patients with enough patients to give a reliable answer.
Translator’s note: nabiximols is a pharmaceutical product with a fixed dose in every spray. Cannabis oils sold or passed around in Thailand are a different thing. Their THC and CBD content is unknown unless a laboratory has tested that batch. If someone uses the Twelves result to sell you an oil, the result does not apply to that oil. If an oil comes with a lab certificate, see how to read a certificate of analysis.
What you can do: before you quote any trial, write down three things. How many patients were in it? Was there a control group? What was the trial designed to test? If you cannot find all three, do not use the result to advise anyone.
4. Two Laboratory Findings That Are Rarely Quoted
Laboratory studies usually use cannabinoid concentrations far higher than any human dose reaches in the blood. Inside a brain tumour, the gap is even wider. The brain is protected by the blood–brain barrier, a tight lining of blood vessels that keeps many substances out of brain tissue. The topic on how cannabinoids move through the body explains why blood levels stay low.
Some findings also point the other way. At low concentrations, cannabinoids have made some cancer cells grow faster.
A 2022 study in mice found that cannabinoids weakened the immune system’s attack on tumours (Xiong et al., Signal Transduction and Targeted Therapy, 2022). The affected cells were T cells, the white blood cells that attack tumours. The effect worked through the CNR2 gene, which makes the CB2 cannabinoid receptor, and through a chain of signals called JAK/STAT. The cannabis and cancer topic linked in section 1 covers both the concentration argument and the immune argument in detail.
What you can do: if a family member is already being treated for cancer, do not add cannabis products on your own. Because of the immune finding, talk to the oncologist before using any product.
5. What Glioblastoma Involves, and What Current Treatment Achieves
“Malignant glioma” is not one disease. The World Health Organization (WHO) publishes the official classification of brain tumours, and the current edition is called CNS5 (2021). Under CNS5, these groups matter here:
- Glioblastoma, IDH-wildtype, grade 4: the most aggressive form. “IDH-wildtype” means the tumour cells carry no change in the IDH gene.
- IDH-mutant astrocytomas and oligodendrogliomas, grades 2–3: tumours whose cells carry a change in the IDH gene. Patients with these tumours have a much better outlook, and the treatment is different.
- Grade 2 IDH-mutant tumours: treatment for these now includes vorasidenib, a drug that blocks the faulty IDH enzyme. The US Food and Drug Administration (FDA) approved it in 2024.
All the figures below apply to glioblastoma only.
Older cannabis literature gives a median survival of 40 to 50 weeks from diagnosis. That figure dates from an earlier era of treatment.
Today, standard treatment for newly diagnosed glioblastoma has three steps:
- Surgery to remove as much of the tumour as can be removed safely.
- Radiotherapy with temozolomide at the same time. Temozolomide is given at 75 mg/m² per day for six weeks, alongside 60 Gy of radiation in 30 sessions. “mg/m²” means the dose is set by body surface area, which is calculated from height and weight. “Gy” (gray) is the unit of radiation dose.
- Six maintenance cycles of temozolomide after radiotherapy ends.
This standard comes from a large trial run by two cancer research networks. They are the European Organisation for Research and Treatment of Cancer and the National Cancer Institute of Canada (EORTC/NCIC). Stupp and colleagues, the team that ran the trial, reported it in the New England Journal of Medicine (2005; 352:987–96). The five-year results came in a later report (Lancet Oncology, 2009; 10:459–66).
| EORTC/NCIC trial | Median survival | Alive at 2 years | Alive at 5 years |
|---|---|---|---|
| Radiotherapy alone | 12.1 months | 10.4% | 1.9% |
| Radiotherapy + temozolomide | 14.6 months | 26.5% | 9.8% |
Trials select their patients, for example by general fitness. Survival across all glioblastoma patients is therefore generally lower than these figures.
Tumour-treating fields come from a device worn on the shaved scalp. It sends weak, alternating electric fields into the brain, and these fields disrupt the division of tumour cells. In the EF-14 trial (Stupp et al., JAMA, 2017), patients added the device to maintenance temozolomide. Their median survival rose to 20.9 months, compared with 16.0 months for patients on temozolomide alone.
When the tumour comes back, some selected patients can have a second operation or a second course of radiotherapy. Another option is bevacizumab, a drug given by drip that blocks the growth of tumour blood vessels. These options can control the disease or its symptoms for a time. Even so, the tumour grows again in almost all patients.
These numbers are hard to read, and they are stated plainly on purpose. A “cure” claim does the most harm in exactly this situation. The patient is not gullible. The standard options offer so little that any chance looks worth taking.
That is why survival data on alternative medicine matter here. The topic on patients who decline cancer treatment covers a study of curable cancers. Some patients in that study used alternative medicine instead of standard treatment. Their hazard ratio for death was 2.50, which means their risk of dying during the study period was two and a half times as high.
Glioblastoma is not curable, so the balance is different and closer. But surgery, temozolomide with radiotherapy, and tumour-treating fields add months and sometimes years. These are the treatments that randomised trials support.
Translator’s note: in Thailand, ask the treating hospital which of these steps it can provide and which costs your health scheme covers. These treatments need a hospital with neurosurgery and radiotherapy departments. If the family lives in a rural province, ask the local hospital for a referral as soon as possible. Do not wait for symptoms to get worse.
What you can do: ask the treating team to write down the exact diagnosis: tumour type, grade and IDH status. The survival figures and the treatment choices depend on these three details.
6. What Is Reasonable
Symptom relief is a real role
The Madrid group says this in its own words: cannabinoids have several palliative effects in cancer patients. Nausea, appetite, sleep and pain are genuine targets. The strongest evidence of all is for nausea and vomiting caused by chemotherapy.
Tell the treating team about everything
Temozolomide becomes active by breaking down chemically on its own inside the body. CYP enzymes are liver enzymes that process many drugs, and they play only a small part in activating temozolomide. So temozolomide is not the main worry for interactions. The Twelves trial found no clinically important effect of nabiximols on how the body handles temozolomide.
The real risks come from other drugs that brain tumour patients often take:
- Dexamethasone is a steroid widely used to reduce brain swelling. The liver enzyme CYP3A4 processes it, and cannabinoids use the same enzyme.
- Enzyme-inducing anticonvulsants, such as carbamazepine and phenytoin, can change cannabinoid levels in the blood. These are seizure drugs that speed up liver enzymes.
- CBD raises the level of clobazam’s active breakdown product. Clobazam is a seizure drug, so the patient becomes more drowsy.
- Levetiracetam is the seizure drug most often used in brain tumour patients. It has little known interaction potential.
None of this can be judged at home. Give the treating team a full list of everything the patient is taking. The topic on cannabis drug interactions has a list you can take to the doctor.
Translator’s note: in Thailand, families often add herbal remedies, traditional medicines and cannabis oil at the same time. Put every one of them on the list, including products with no label. Many hospitals under the Ministry of Public Health run a medical cannabis clinic. Ask whether the treating hospital has one. If it does, ask for a referral there instead of buying oil privately.
Translator’s note for growers: a relative or friend may ask you to make cannabis oil for a patient. Say no, and send them to the treating doctor. Section 46 of the Thai Traditional Medicine Protection Act B.E. 2542 covers processing a controlled herb without a licence. Under Section 78, the penalty is up to 1 year in prison, a fine of up to 20,000 baht, or both.
Since 26 April 2026, licences for cannabis extracts above 0.2% THC have gone only to three kinds of applicant. These are Thai juristic persons that are not “foreign” under the Foreign Business Act B.E. 2542, state agencies, and the Thai Red Cross Society. The law is not the only issue. An oil of unknown strength can also interact with the patient’s seizure drugs and steroids, as described above.
Clinical trials exist
A patient interested in cannabinoids for glioma should ask their neuro-oncologist about clinical trials. Only a proper trial can turn a result like 12 against 9 into a real answer. Without one, the result stays a rumour.
Translator’s note: you can search for trials on ClinicalTrials.gov, a public registry run by the US National Library of Medicine. You can also search the Thai Clinical Trials Registry (TCTR) at thaiclinicaltrials.org. Search for “glioblastoma” together with “cannabinoid” or “nabiximols”. Take any trial you find to the neuro-oncologist, who can judge whether the patient qualifies.
What to Do With a “Cannabis Cures Brain Cancer” Article
The evidence on cannabinoids and glioma looks like the evidence behind every “cannabis cures cancer” claim. It comes from laboratory and animal work, and no trial in people has had enough patients to give a reliable answer. The one difference is that the Madrid group never made the cure claim. Someone else attached it to their work.
If you have a brain tumour diagnosis and someone has sent you material about cannabis oil:
- Do not start, stop or change any treatment because of it.
- Take it to your neuro-oncology team and ask them to go through it with you.
- If you want the sources checked, post the material in a reply below this topic. Members can compare it with the original papers.
This is general educational information, not medical advice. Cannabis is not a treatment for glioma. No regulator has approved any cannabinoid as a cancer treatment.
References
- Velasco et al. Mol Neurobiol 2007;36:60–67.
- Guzmán. Nat Rev Cancer 2003;3:745–755.
- Guzmán et al. Br J Cancer 2006;95:197–203.
- Twelves et al. Br J Cancer 2021;124:1379–87.
- Hinz & Ramer. “Cannabinoids as anticancer drugs: current status of preclinical research.” Br J Cancer 2022;127:1–13.
- Xiong et al. Signal Transduct Target Ther 2022;7:99.
- World Health Organization Classification of Tumours of the Central Nervous System, 5th edition (CNS5), 2021.
- Stupp et al. N Engl J Med 2005;352:987–96.
- Stupp et al. Lancet Oncol 2009;10:459–66.
- Stupp et al. (EF-14 trial). JAMA 2017.
- US FDA approval of vorasidenib, 2024.