Professor Dalgleish checked

Written 7 September 2026 for her partner. Every study is cited by PubMed ID (PMID) so it can be checked. The full research thread with the table of papers is em-research/15.

Short answer

Dalgleish is a real oncologist with a real record, and his vitamin D advice is mostly opinion, not his own research. His published vitamin D work is three laboratory papers on cells; his statements about 2,000 patients and "several papers" have no matching publication. The independent trial evidence supports correcting a low vitamin D level to the normal range with a daily dose, and does not support pushing the level to 100 nmol/L. The compound that "lets chemotherapy work at a lower dose" is his CBD plus THC pairing, shown in leukaemia cells in a dish; there is no human trial anywhere. For the patient, what transfers is a blood test and 2,000 to 4,000 IU a day; the rest does not, yet.

Who he is

Professor of Oncology at St George's, London, a melanoma and immunotherapy specialist, first author of the 1984 paper identifying the HIV receptor (PMID 6096719), with over 300 publications. He also has interests that colour his enthusiasms: adviser to the company that makes IMM-101, named on patents for low-dose naltrexone in cancer, funded by a cannabinoid company (his own declarations, PMIDs 35761946, 29500634). During the pandemic he made public claims about vaccines and cancer that have no peer-reviewed basis. None of that makes his laboratory work wrong. It does mean his public statements should be checked against what has actually been published, which is what follows.

Vitamin D and the immune system

What he has published. Three papers, 2000 to 2012, all on cells in a dish, on vitamin D compounds and prostate or melanoma cell lines. One of them found the vitamin D analogue blocked another drug's anticancer effect. Nothing in patients, nothing on the innate immune system.

What he says. That most of his melanoma patients are low in vitamin D, that low patients respond poorly to immunotherapy, that vitamin D is vital for the immune response, and that every cancer patient should have the level measured and corrected first. The biology behind that is real and independent of him: immune cells need vitamin D to make an antimicrobial peptide (Science 2006, PMID 16497887), and in 872 melanoma patients a lower level at diagnosis went with thicker tumours and worse survival (PMID 19770375). That is an association. It does not show that correcting the level changes the outcome.

What the big trials show. - VITAL, 25,871 adults, 2,000 IU a day for five years: no fall in cancer diagnoses; in follow-up analysis, fewer advanced or fatal cancers, confined to people of normal weight (PMIDs 30415629, 33206192). - D-Health, 21,315 older adults, a large monthly dose that pushed levels to 115 nmol/L: no fall in deaths, and cancer deaths trended the wrong way (PMID 35026158). - Meta-analyses of all randomised trials: no effect on getting cancer, a modest reduction in dying of cancer, seen with daily dosing rather than big monthly doses (PMIDs 30796437, 35676320).

The 100 nmol/L target. UK labs call under 25 deficient, 25 to 50 insufficient, over 50 sufficient. Some advocates argue for 75 to 125 (PMID 18844846). The 2024 Endocrine Society guideline now recommends against chasing levels above the standard thresholds (PMID 38828931). No trial shows 100 beats 50 to 75; the one trial that reached 115 showed no benefit. Toxicity, which is high calcium, appears from about 220 to 250 nmol/L upward (PMID 18290725). At the patient's weight, going from a low level to 100 needs about 3,000 IU a day (PMID 24037880).

For the patient. Have the level and calcium measured at the GP. Take 2,000 IU a day now; if the result is low, 3,000 to 4,000 a day until it is above 50 to 75, then back to 2,000. High calcium causes nausea, thirst, constipation, confusion and drowsiness, which are the very things being watched in her, so the dose stays where a blood test can justify it.

The compound that "reduces the chemotherapy dose"

It is the pairing of CBD with THC, from Dalgleish's laboratory with Wai Liu (PMID 28560402, building on PMID 24123005). In leukaemia cells, the pair made two chemotherapy drugs work at much lower doses, provided the cannabinoids were given after the chemotherapy. A 2022 paper adds low-dose naltrexone before CBD, in cells and mice (PMID 35179218). All of it is dishes and mice. There is no human trial of cannabinoids lowering a chemotherapy dose in any cancer.

Whether it matters for the patient depends on the tissue. - Conventional chordoma, the most likely result: chemotherapy is not used because the tumour does not respond to it (PMIDs 28750274, 38691867). A compound that lowers the dose of a drug she will not be given has no bearing. - Correction, 12 Oct 2026 (Kya): "most likely" is out of date. On 6 Oct the surgeon said the tumour is "not a well behaved, aggressive tumour, possibly a form of sarcoma", with INI1 possibly lost. The two branches below, where chemotherapy is used, are now the live ones. 13 Oct 2026: the diagnosis is poorly differentiated chordoma, so chemotherapy is on the table. The evidence that it helps in this type is weak: small case series and case reports, no trials. - Dedifferentiated chordoma: sarcoma-type chemotherapy is offered. The largest pooled analysis found that complete surgery and radiotherapy predicted survival and chemotherapy did not separate outcomes (PMID 36804481). Here the idea is at least about a drug she might receive, but the data are leukaemia cells with different drugs. - Poorly differentiated (INI-1 deficient) chordoma: a rare, aggressive, mainly childhood form, treated with surgery, radiotherapy and chemotherapy together (PMID 29483606); the targeted drug tazemetostat has produced responses (PMID 30642912). No cannabinoid data at all.

[Correction 11 Oct 2026: tazemetostat was withdrawn from all markets in March 2026 (Ipsen, after SYMPHONY-1 showed second blood cancers in 5.7% vs 0); trials and compassionate access stopped. PMID 42599162; research 12 A3. Read anything above or below about it as history, not an option.]

Radiotherapy, which is her likely treatment. The same group showed CBD plus THC made glioma cells more sensitive to radiation and shrank tumours in mice (PMID 25398831); other laboratories reproduced it (PMIDs 29088769, 30783513). In other brain tumour mouse models the dish result did not translate into any survival benefit (PMID 33477420). Human data: two uncontrolled case series of 9 and 15 glioblastoma patients (PMIDs 31570484, 35403130) and one 21-patient randomised safety trial (PMID 33623076). Nothing with protons, nothing in chordoma. This is the one idea worth raising with the proton team when the time comes, as a question, not a plan.

IMM-101

A heat-killed mycobacterium injected under the skin, made by the company he advises. One randomised trial in 110 pancreatic cancer patients missed its main endpoint, with a signal in a subgroup (PMID 27599039). It was safe alongside stereotactic radiotherapy in 38 pancreatic patients with no sign of added benefit (PMID 36813171). No phase 3, no licence, no trial in sarcoma or any bone tumour. The company went into administration in December 2024 and remains there; its compassionate-access programme is closed and a 2024 academic trial was withdrawn before recruiting, so the product has no maker and no supply. Nothing for the patient here.

Vitamin C

Dalgleish has published nothing on it. The evidence stands on its own.

Tablets. Absorption saturates: blood levels plateau on 200 to 400 mg a day and doses above 1 g add little (PMID 15068981). Two randomised trials of 10 g a day in advanced cancer showed no benefit (PMIDs 384241, 3880867). Safe up to 1 g a day; above that, loose stools and more oxalate for the kidneys.

High-dose intravenous. At drip levels vitamin C generates hydrogen peroxide in the tissue fluid and kills cancer cells selectively in the dish and in mice (PMIDs 16157892, 18678913). Human studies: small phase 1 trials with chemotherapy in pancreatic and ovarian cancer showing safety and less chemotherapy toxicity (PMIDs 22272248, 24500406, 29215048); the Iowa glioblastoma work, giving it during radiotherapy, safe, with survival better than historical controls in 11 and then in a single-arm phase 2 (PMIDs 28366679, 37773633). Historical controls are unreliable in that disease and a randomised trial has not been done. Randomised evidence elsewhere: one small positive trial (PMID 39369582) against two larger negative ones, including a 442-patient colorectal trial (PMID 35929990) and a prostate trial stopped for futility (PMID 39076107). No randomised trial shows it prolongs survival; the best-supported effect is on quality of life and chemotherapy tolerance (PMID 24867961).

Safety. A G6PD blood test is compulsory first, because the infusion can destroy red blood cells in people who lack that enzyme (PMID 32561141). Kidney function must be normal because of oxalate. It makes fingerprick glucose meters read falsely high. Sodium ascorbate is the form used and it carries about 110 mg of sodium per gram.

With radiotherapy: genuinely contested. At infusion doses it behaves as a pro-oxidant and sensitised tumours to radiation in animals (PMID 30254147) and was given during photon radiotherapy in the glioblastoma trials without excess harm. At tablet doses it is an antioxidant and most radiotherapy units advise against high-dose antioxidants during a course. Neither side has proton data. If it is ever tried, it is with the radiation oncologist's agreement, not around them.

What transfers to the patient, and what does not

• Transfers: a vitamin D and calcium blood test, 2,000 IU a day now, replete to 50 to 75 nmol/L if low, re-test. Cheap, safe, evidence-based.
• Does not transfer yet: cannabinoids to lower a chemotherapy dose (cells only, and chemotherapy is unlikely to be her treatment); IMM-101 (nothing in her tumour type); high-dose vitamin C (no survival evidence, real safety checks, contested with radiotherapy); 100 nmol/L as a target (no trial supports it over 50 to 75); low-dose naltrexone (blocks her morphine, covered in the earlier brief).
• Worth one question to the proton team later: is there any place for cannabinoid or vitamin C radiosensitisation in her plan, and are they aware of the glioblastoma data. Expect the answer to be no for now; the point is to have asked.
• Safety to keep in view: high calcium from vitamin D mimics the drowsiness and confusion being watched; cannabinoids add to the sedation from morphine and gabapentin and can affect eye movements on top of her existing double vision.