Clival chordoma research
Home › Diet, supplements and claims

Plant compounds

Research agent report, 4 September 2026, from PubMed (E-utilities searches run today), LiverTox, NCI PDQ, MSKCC herb monographs, UK regulatory sources and published trials. Gaps marked. No prognosis or survival figures. Where a statement is inference or clinical judgement rather than citation it is labelled as such. No em dashes.

Patient context: 34-year-old woman, 3.8 cm invasive clival mass with partial basilar artery encasement and right third and fourth nerve involvement, working diagnosis chordoma (chondrosarcoma the main lookalike), biopsy pending, expected pathway endoscopic surgery then proton beam therapy. Continuous right trigeminal-distribution pain, weight loss (51 kg), on diclofenac, omeprazole and paracetamol, possibly diazepam; likely to be on an opioid and a nerve-pain drug (gabapentinoid, duloxetine or carbamazepine) and possibly dexamethasone. Partner very wary of sedation and cognitive side effects. He has heard that a plant contains "some of the strongest cancer fighting compounds known" and cannot remember which.

0. The short version

  1. None of the nine plants, and none of the compounds in them, has any evidence of treating chordoma in a person. Across all nine there is exactly one prospective chordoma trial (a camptothecin derivative, 15 patients, one response, 2005), a handful of chordoma cell-line experiments (mostly showing resistance), a few paediatric case reports, and nothing at all for five of the nine. For chondrosarcoma the picture is the same: one small taxane trial arm (2 responses in 25), one set of curcumin dish and mouse experiments, and nothing for the rest.

  2. Four of the nine (periwinkle, yew, mayapple, happy tree) are the botanical origin of real, licensed chemotherapy drugs. That is a fact about the history of pharmacology, not about eating the plant. All four plants are poisonous; the drugs are extracted or semi-synthesised under pharmaceutical control, given intravenously or as precisely dosed tablets, and are among the most toxic medicines in use. They are not supplements, they cannot be bought, and the family could not and should not use them. Whether any of them ever enters the patient's treatment is a question for her oncologist and depends on histology (section 12).

  3. The phrase "strongest cancer fighting compounds known" is almost certainly the graviola (soursop) marketing claim ("10,000 times stronger than chemotherapy"), which traces to 1990s dish experiments on acetogenins at Purdue University and has been repeated by supplement sellers ever since. Graviola has no human evidence, is neurotoxic (it causes an atypical Parkinson's syndrome where it is eaten heavily), and is the one plant on this list I would ask the family to keep away from the patient entirely. Details in section 4.

  4. Of the things the family could legally and practically use alongside treatment, only turmeric and black seed are reasonable, and only at food-level doses, declared to the team, and paused around surgery and radiotherapy. Neither treats chordoma. Mistletoe has the most randomised evidence of any plant medicine and its best-designed trial (2024) was negative for survival. Artesunate is a genuine repurposing candidate with one small positive pilot in bowel cancer, but the supplements sold as "artemisinin" are not the drug that was tested, and the drug itself induces liver enzymes in a way that would interfere with her likely medicines.

  5. The honest ranking (section 11) is: none of these treats chordoma; turmeric and black seed are acceptable low-dose adjuncts with conditions; mistletoe and artesunate are "only under a prescriber and only after asking the oncologist"; the four drug-origin plants are irrelevant as plants and toxic; graviola is dangerous.

1. Madagascar periwinkle (Catharanthus roseus): vincristine, vinblastine

a. Compounds and mechanism. The vinca alkaloids (vincristine, vinblastine, and the semi-synthetic vinorelbine and vindesine) bind tubulin and prevent microtubule assembly, so dividing cells cannot form a mitotic spindle and arrest in metaphase, then die. They are classic cell-cycle-specific cytotoxics. Vincristine also damages the microtubules of peripheral nerve axons, which is why peripheral neuropathy is its dose-limiting toxicity.

b. Licensed drug versus plant. Vincristine and vinblastine are licensed intravenous chemotherapy drugs in the UK and worldwide, given only by specialist teams (intrathecal vincristine is fatal and its administration is surrounded by national safety rules). The plant contains these alkaloids at vanishingly small concentrations (vincristine of the order of 0.0005 percent of dry leaf weight, which is why the drug originally needed about half a tonne of leaves per gram) alongside dozens of other alkaloids. Periwinkle leaf teas are sold in some traditional systems for diabetes; there is a 2024 case report of hepatotoxicity with cholestatic jaundice from drinking the juice, mimicking cholangitis, which resolved on stopping (Lin and colleagues, BMC Complementary Medicine and Therapies 2024, PMC10993546). The vinca alkaloids are also poorly absorbed by mouth. There is no vinca "supplement" with any meaningful content of the drug; anything labelled as periwinkle extract is an unstandardised herbal product with a known liver signal.

c. Evidence tiers. As a drug: overwhelming. Vincristine is a component of curative combination regimens in childhood leukaemia, Hodgkin and non-Hodgkin lymphoma, Wilms tumour, rhabdomyosarcoma and others; vinblastine in Hodgkin lymphoma and testicular cancer. Thousands of randomised patients over sixty years. As a plant or supplement: no human evidence of any anticancer effect.

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (vincristine OR vinblastine OR vinca): 11 records, none showing benefit in adult conventional chordoma. - Bayrak and colleagues, "The effects of chemotherapeutic agents on differentiated chordoma cells", Journal of Neurosurgery: Spine 2011;15:620-624, PMID 21905773. U-CH1 cells exposed to vincristine, doxorubicin, cisplatin, etoposide, fludarabine, methotrexate, nilotinib and imatinib, with and without all-trans retinoic acid. Chemotherapy alone had limited effect; the authors' point was that forcing differentiation with retinoic acid made the cells more sensitive. In vitro only. - Azzarelli and colleagues, Journal of Surgical Oncology 1988;37:185-191, PMID 3352273. 33 chordomas, Milan, 1933 to 1983: "No chemotherapeutic regimen achieved any result; one case had a short complete remission after cisplatin plus vinblastine plus bleomycin." - Scimeca and colleagues, Journal of Pediatric Hematology/Oncology 1996;18:237-240, PMID 8846149. A 19-month-old with metastatic cervical chordoma: ifosfamide and doxorubicin produced regression; "no apparent benefit from the use of actinomycin-D, cyclophosphamide and vincristine". - Al-Rahawan and colleagues, Pediatric Blood and Cancer 2012;59:323-325, PMID 21922644. A 7-month-old with clival chordoma had a durable complete response to vincristine/cyclophosphamide/doxorubicin alternating with etoposide/ifosfamide. Infant chordomas are biologically different (usually the poorly differentiated, SMARCB1-deficient type, see section 12) and the authors themselves state that chemotherapy in chordoma "has been generally proven ineffective". - Razis 1974 (Journal of Medicine, PMID 4528471) is a single 1974 case of cervical chordoma treated with vincristine; the abstract is not retrievable and the outcome is unknown. Chondrosarcoma: 32 records, almost all case reports, mesenchymal chondrosarcoma (a distinct, Ewing-like subtype treated with sarcoma regimens that include vincristine), or secondary tumours after Hodgkin treatment. Nothing supporting vinca alkaloids for conventional chondrosarcoma, which is treated as chemotherapy-resistant (Terek 1998, section 2d).

e. Safety for the patient. Not applicable as a plant, because she should not consume it: unstandardised, hepatotoxic in a case report, and the active alkaloids are neurotoxic. As a drug vincristine causes peripheral and cranial neuropathy (including ptosis and ophthalmoplegia), constipation and jaw pain, none of which she needs added to a picture that already includes third, fourth and fifth nerve signs.

f. UK availability. The plant is an ornamental sold in garden centres. Vinca-containing herbal products are not on any approved list and I found no UK retailer of a standardised extract. Vincristine and vinblastine are hospital-only prescription medicines.

g. Verdict. A real drug came from this plant; the plant itself is useless and mildly dangerous to consume, and the drug has no role in conventional chordoma.

2. Pacific yew (Taxus brevifolia) and European yew (Taxus baccata): paclitaxel, docetaxel

a. Compounds and mechanism. Paclitaxel (originally from Pacific yew bark; now made by plant cell culture or semi-synthesis) and docetaxel (semi-synthesised from 10-deacetylbaccatin III extracted from European yew needles) do the opposite of the vinca alkaloids: they bind tubulin and stabilise microtubules so they cannot disassemble, which also freezes the mitotic spindle and triggers cell death. They are among the most widely used chemotherapy drugs in the world.

b. Licensed drug versus plant. Both are licensed intravenous chemotherapy in the UK. The plant is one of the most poisonous in Britain. All parts except the red aril contain taxine alkaloids (taxine B the most potent), which block cardiac sodium and calcium channels; the lethal dose of leaves is estimated at 0.6 to 1.3 g per kg body weight, deaths occur within hours from heart block and ventricular arrhythmia, and there is no antidote (case series: Grobosch 2013 Forensic Science International, eight fatal and non-fatal T. baccata poisonings; fatal heart block case, European Heart Journal Case Reports 2020; cardiogenic shock rescued by ECMO, several 2018 to 2023 reports). Yew contains essentially no paclitaxel in a usable form and no "yew supplement" exists. There is nothing to buy and nothing to take.

c. Evidence tiers. As drugs: enormous randomised evidence in breast, ovarian, lung, prostate, head and neck, gastric and other cancers. As a plant: none, and the plant kills.

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (paclitaxel OR docetaxel OR taxane): 7 records, no trial, no case with benefit. - Jian and colleagues, American Journal of Translational Research 2018;10:936-947, PMID 29636883. In primary chordoma cell lines, the chordoma master transcription factor brachyury blocks paclitaxel-induced apoptosis by switching on carbonic anhydrase IX; in other words, chordoma's defining biology confers taxane resistance. - Kim and colleagues, Journal of Neurosurgery: Spine 2016;25:626-635, PMID 27314550. A new cell line from a dedifferentiated chordoma was more resistant to paclitaxel, cisplatin, etoposide and radiation than U-CH1. - Xiao and colleagues, Colloids and Surfaces B 2022;212:112366, PMID 35144131. An RNA nanoparticle carrying an EGFR aptamer delivered paclitaxel to U-CH2 cells more effectively than free drug. In vitro; a delivery-technology paper. Chondrosarcoma: 25 records. The relevant ones: - Fox and colleagues, The Oncologist 2012;17:321, PMID 22363068 (SARC003). Gemcitabine plus docetaxel in 25 patients with unresectable or recurrent chondrosarcoma: no complete responses, 2 partial responses, arm closed; probability of reaching even a 20 percent response rate was 14 percent. - Li and colleagues, International Journal of Molecular Sciences 2024;25:13003, PMID 39684713. High-throughput screen in IDH1-mutant chondrosarcoma lines: docetaxel, methotrexate, panobinostat, idarubicin, camptothecin and pevonedistat were among the top 20 hits regardless of IDH status. In vitro. - Gorelik 2008 (Cancer Research, PMID 18974149): docetaxel plus bevacizumab in a mesenchymal chondrosarcoma xenograft; mesenchymal is a different disease from skull base conventional chondrosarcoma. - Terek and colleagues, Journal of Orthopaedic Research 1998;16:585-590, PMID 9820282: 37 of 41 cartilage tumours expressed P-glycoprotein, the drug-efflux pump, offered as the reason chondrosarcoma resists chemotherapy.

e. Safety for the patient. The plant: potentially fatal; keep it out of the house if there are any herbal experiments going on. The drugs: not relevant unless an oncologist proposes them; peripheral neuropathy again.

f. UK availability. Yew is in every churchyard. Taxanes are hospital-only. There is no legal or safe consumer product.

g. Verdict. Origin of two of the most important chemotherapy drugs; the plant is lethal; the drugs have shown resistance in chordoma cells and near-zero activity in chondrosarcoma patients.

3. Graviola or soursop (Annona muricata): annonaceous acetogenins

a. Compounds and mechanism. Acetogenins (annonacin, bullatacin, squamocin and about 100 others) are long-chain fatty acid derivatives that are among the most potent known inhibitors of mitochondrial complex I (NADH:ubiquinone oxidoreductase). Blocking complex I starves cells of ATP; in dishes this kills cancer cells at nanomolar concentrations, and the famous "10,000 times more potent than adriamycin" claim came from exactly such dish comparisons at Purdue University in the 1990s (McLaughlin's group; Oberlies 1997 and related papers). The same mechanism is the problem: complex I inhibition is also how MPTP and rotenone cause Parkinson's disease, and the acetogenins cross into the brain.

b. Licensed drug versus plant. No acetogenin has ever become a drug, in any country. Graviola is sold only as a food supplement (leaf capsules, powders, teas, fruit pulp). What is in a capsule varies enormously: a 2023 quality study of leaf products found active-compound concentrations differing markedly between products, and young leaves contain about twice the annonacin of mature leaves (Journal of Pharmacy and Pharmacology 2023;75:1357, evaluation of the quality of Annona muricata leaf products). Dried leaves have been measured at about 306 micrograms of annonacin per gram (Champy 2005, Movement Disorders, quantification of acetogenins in Annona muricata linked to atypical parkinsonism in Guadeloupe).

c. Evidence tiers. In vitro: extensive, in breast, prostate, colon, pancreatic, lung and other lines (review: PMC9497149, 2022). Animal: some xenograft and chemically induced tumour models. Human: no clinical trial of graviola or any acetogenin in any cancer. What exists is survey data: in Trinidad and Jamaica 60 to 81 percent of cancer patients in some clinics report using it (cited in the 2022 review above), which tells us about popularity, not effect. No case report of a documented tumour response was found.

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (annona OR graviola OR acetogenin): zero. (chondrosarcoma) AND the same: zero. Nothing in any chordoma cell line, animal model or patient.

e. Safety for the patient. This is where graviola stands apart from the rest of the list. - Neurotoxicity: on Guadeloupe an unusual form of atypical parkinsonism (levodopa-unresponsive, with early gait and eye-movement problems) was linked epidemiologically to heavy consumption of Annona fruit and herbal teas (Caparros-Lefebvre and Elbaz, Lancet 1999;354:281-286; Lannuzel and colleagues, Brain 2007;130:816-827). Annonacin given to rats enters the brain, lowers ATP and destroys basal ganglia neurons (Champy 2004, Journal of Neurochemistry); it is roughly 100 times more potent than MPP+ (the active metabolite of MPTP) at this. Höllerhage and colleagues, International Journal of Toxicology 2015;34:543-553, tested commercial Annonaceae dietary supplements and found them neurotoxic to cultured dopaminergic neurons at concentrations achievable from labelled doses. The FDA and Health Canada have both noted the concern; MSKCC's monograph lists neurotoxicity as the principal risk (https://www.mskcc.org/cancer-care/integrative-medicine/herbs/graviola). - Why that matters for the patient specifically (clinical judgement): her tumour sits against the brainstem and has already taken the third and fourth nerves; the surgical and radiotherapy teams will be tracking her eye movements, gait, balance and cognition closely for months. A compound that damages the basal ganglia and brainstem nuclei and produces eye-movement and gait abnormalities is the last thing to add to that assessment. A complex I inhibitor is also the wrong thing for someone who is losing weight. - Other effects (MSKCC): hypotension and hypoglycaemia in animals; CYP3A4 activation and CYP2C8 inhibition reported in vitro, so unpredictable effects on her opioid, dexamethasone and carbamazepine levels. - Radiotherapy: no data.

f. UK availability. Legal, cheap and everywhere: capsules of 500 mg leaf for about 10 to 20 pounds per 100 on Amazon UK, eBay and health shops. No quality standard, no upper limit set by the FSA, no requirement to declare acetogenin content.

g. Verdict. The likely source of the "strongest compounds known" claim; no human evidence; a documented neurotoxin; the one item on this list to keep away from the patient.

4. Sweet wormwood (Artemisia annua): artemisinin, artesunate, dihydroartemisinin

a. Compounds and mechanism. Artemisinin is a sesquiterpene lactone with an endoperoxide bridge. Ferrous iron cleaves the bridge and generates carbon-centred free radicals and reactive oxygen species. In malaria this kills the parasite inside iron-rich red cells; in cancer cells, which take up more iron via the transferrin receptor, the same reaction damages proteins and lipids and triggers ferroptosis (iron-dependent cell death), with additional reported effects on angiogenesis and cell cycle (reviews: Frontiers in Pharmacology 2024, "Anti-tumor mechanism of artesunate"; Hu 2022, Oxidative Medicine and Cellular Longevity). Artesunate (water-soluble prodrug) and dihydroartemisinin (the active metabolite of all of them) are the forms used clinically. This is a pro-oxidant mechanism, not an antioxidant one, which matters for radiotherapy (e).

b. Licensed drug versus plant. Intravenous artesunate (Artesunate Amivas) received UK marketing authorisation from the MHRA in April 2022 for severe malaria; artemether-lumefantrine tablets (Riamet) are licensed for uncomplicated malaria; oral artesunate tablets are not licensed in the UK or EU and are obtained for trials as unlicensed imports. All are prescription-only. What is sold in UK health shops as "artemisinin" (typically 100 mg capsules, about 36 pounds per 90, Doctor's Best and others) is a food supplement of the parent compound, not artesunate, with poor and variable oral absorption and no medicinal licence. Artemisia annua tea is even less reliable: measured artemisinin content ranges from 0.3 mg/L (plants grown in France) to 15.7 mg/L (Benin), with a third to a half lost after a year's storage (Journal of Pharmaceutical and Biomedical Analysis 2012; Malaria Journal analyses). Private "artemisia infusion" clinics in the UK offer intravenous artesunate for about 500 pounds a session, outside any trial and without published outcomes.

c. Evidence tiers. - In vitro and animal: extensive across tumour types, including radiosensitisation in glioblastoma cells (Reichert and colleagues, Radiotherapy and Oncology 2012, PMID 22560712) and cervical cancer cells and xenografts (Luo 2014, Radiation Oncology 9:84). - Human, the best evidence: Krishna and colleagues, "A randomised, double blind, placebo-controlled pilot study of oral artesunate therapy for colorectal cancer", EBioMedicine 2015;2:82-90, PMID 26137537. 23 patients randomised to 200 mg oral artesunate daily or placebo for 14 days before bowel cancer surgery; 20 completed per protocol; over a median 42 months, 1 of 9 on artesunate versus 6 of 11 on placebo recurred. A pilot, not powered for outcomes, and the authors say so. The follow-on phase II trial NeoART-V (NCT03093129, 140 patients, primary endpoint recurrence-free survival at two years) has, as far as I can find today, not published results. - von Hagens and colleagues, "Prospective open uncontrolled phase I study to define a well-tolerated dose of oral artesunate as add-on therapy in patients with metastatic breast cancer (ARTIC M33/2)", Breast Cancer Research and Treatment 2017;164:359-369, PMID 28439738. 23 patients, 100 to 200 mg daily for 4 weeks; 200 mg tolerated; no efficacy claim. Long-term compassionate use up to 37 months in the same patients without major safety concerns (Phytomedicine 2019, PMID 30668363); a dedicated audiology substudy found no clinical ototoxicity (Cancer Chemotherapy and Pharmacology 2016). - Deeken and colleagues, Cancer Chemotherapy and Pharmacology 2018: phase I of intravenous artesunate in solid tumours, dose-limiting toxicity at higher doses, no responses. - Case reports in glioblastoma (Zeitschrift für Phytotherapie 2024) are uncontrolled and in a different tumour. - Cervical precancer: a small trial of oral artesunate for CIN2/3 (Trimble 2020, Gynecologic Oncology) showed regressions; a larger trial (NCT07095478) is planned. Precancer, not cancer. No systematic review finds any cancer in which artesunate has proven benefit; the field's own editorial calls it a promising repurposed drug awaiting phase II data (Krishna 2015, ecancer editorial "The wisdom of crowds and the repurposing of artesunate").

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (artemisinin OR artesunate OR dihydroartemisinin OR artemisia): zero. (chondrosarcoma) AND the same: zero. Nothing in any cell line, animal or patient.

e. Safety for the patient. - Liver: LiverTox (NBK548419) grades artemisinins as a rare cause of hepatotoxicity; transaminase rises in 1 to 4 percent, usually mild and self-limiting; clinically apparent injury is rare but reported, including a glioblastoma patient taking artesunate with temozolomide and Chinese herbs. Artesunate combined with dichloroacetate has caused severe hepatotoxicity and marrow suppression. - Enzyme induction, the interaction that matters most for her: artemisinin induces its own metabolism. In healthy volunteers two days of oral artemisinin increased CYP2B6 activity 2.1-fold and CYP3A4 activity 3.2-fold (Malaria Journal 2014;13:214, PMID 24889062); in hepatocytes it induces CYP3A4 and CYP2B6 through PXR and CAR and also MDR1/P-glycoprotein (Xing 2012, Drug Metabolism and Disposition 40:1757, PMID 22679214; Burk 2005, Molecular Pharmacology). Oxycodone, fentanyl, methadone and tramadol, dexamethasone, and carbamazepine are all CYP3A4 substrates. Induction would lower their levels, so pain control could slip and dexamethasone could under-deliver, unpredictably, over the first week of use and again on stopping. Carbamazepine in turn induces CYP3A4 and would lower artemisinin levels. Gabapentinoids are renally cleared and unaffected. This is mechanism-based inference from human pharmacokinetic data, not a documented case in a cancer patient. - Neurotoxicity: high-dose artemether and arteether in dogs and rats damage brainstem nuclei (the vestibular and auditory nuclei in particular; Brewer 1994, American Journal of Tropical Medicine and Hygiene). This has not been seen in humans at malaria doses or in the breast cancer long-term series, but it is worth stating for a patient whose tumour abuts the brainstem: it is an animal signal at high doses, not a human finding. - Blood: delayed haemolysis after intravenous artesunate is well documented (CDC MMWR 2013); not reported with oral supplement doses. - Radiotherapy: artesunate is a pro-oxidant and radiosensitiser in dishes and in a cervical cancer mouse model; there is no human radiotherapy trial. The antioxidant-interference worry does not apply, but the "unknown addition during proton therapy" worry does. - Sedation and cognition: none reported at studied doses. Weight: none. - Surgery: no antiplatelet effect; but enzyme induction affects anaesthetic drug metabolism (midazolam, fentanyl are 3A4 substrates); the anaesthetist would want to know.

f. UK availability. Supplement capsules legal and cheap (about 12 pounds a month at 100 mg daily); the drug that was actually trialled (artesunate 200 mg daily) is unlicensed orally and would need a doctor prepared to prescribe it off-label from a specials importer, which NHS oncologists will not do outside a trial. Quality of supplements is unregulated; content of teas is unpredictable.

g. Verdict. The most scientifically respectable repurposing candidate on the list, with one small positive pilot in bowel cancer and a phase II still unreported; zero chordoma data; a liver-enzyme inducer that would interfere with her opioid, steroid and possibly carbamazepine levels; not something to take as a shop supplement, and only ever under the oncologist.

5. Mayapple (Podophyllum peltatum): podophyllotoxin, etoposide

a. Compounds and mechanism. Podophyllotoxin is a lignan that, like the vinca alkaloids, binds tubulin and blocks microtubule assembly. Its semi-synthetic derivatives etoposide and teniposide work differently: they inhibit topoisomerase II, the enzyme that cuts and reseals both strands of DNA during replication, trapping the enzyme on DNA and causing double-strand breaks that kill dividing cells.

b. Licensed drug versus plant. Etoposide is licensed in the UK as intravenous and oral (capsule) chemotherapy, hospital-initiated. Podophyllotoxin 0.15 to 0.5 percent is a prescription-only topical treatment for genital warts (Warticon, Condyline) with explicit warnings never to swallow it. The plant (root, rhizome, unripe fruit and leaves; only the ripe fruit pulp is edible) is toxic: ingestion causes vomiting and diarrhoea followed in serious cases by confusion, hallucinations, seizures, peripheral neuropathy and bone marrow suppression; podophyllin resin poisonings in the Chinese herbal literature ("bajiaolian") have caused permanent sensory neuropathy and death (review: Journal of Ethnopharmacology / Pharmacological Research reviews 2023 on the toxicological profile of Podophyllum; a historical two-child case is in PMC9987978). There is no oral mayapple supplement with any legitimate use.

c. Evidence tiers. As a drug: etoposide is part of curative regimens in testicular cancer, small cell lung cancer, lymphomas and several paediatric sarcomas; huge randomised base. As a plant: none.

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (podophyllotoxin OR etoposide): 10 records. This is the compound with the most chordoma patient data on the list, all paediatric and all in the ifosfamide-etoposide combination: - Dhall and colleagues, "The role of chemotherapy in pediatric clival chordomas", Journal of Neuro-Oncology 2011;103:657-662, PMID 21052774. Six children with clival chordoma at Los Angeles Children's, 1995 to 2005; four survivors at a median 9 years had all received ifosfamide and etoposide, with or without radiotherapy. Retrospective, six patients, paediatric, and the authors say only that it "may have a role". - Al-Rahawan 2012 (infant, section 1d): etoposide/ifosfamide alternating with VCD, durable complete response. - Wong and colleagues, Pediatric Blood and Cancer 2023;70:e30386, PMID 37218455: tolerability of an ifosfamide-containing regimen in INI-1 (SMARCB1)-deficient tumours, which include poorly differentiated chordoma. - Cavalheiro and colleagues, Journal of Cancer Research and Clinical Oncology 2026;152:147, PMID 42191910: an ex vivo chemoresistance test on paediatric clival chordoma tissue, presented as a way to pick agents for tumours "generally considered chemotherapy-refractory". - Cell lines: Bayrak 2011 (U-CH1, limited effect alone); Kim 2016 (dedifferentiated line more resistant to etoposide than U-CH1); Tecimel and colleagues, Naunyn-Schmiedeberg's Archives of Pharmacology 2026;399:2481-2493, PMID 40888882 (knocking out CD90 in U-CH1 increased sensitivity to etoposide and metformin, i.e. the parent line is relatively resistant). Chondrosarcoma: 20 records; etoposide appears in the Issels hyperthermia-plus-ifosfamide/etoposide sarcoma series (1990 to 1993, mixed histologies) and in case reports of mesenchymal and dedifferentiated chondrosarcoma. Nothing for conventional chondrosarcoma.

e. Safety for the patient. The plant: do not consume. The drug: relevant only if histology shows a poorly differentiated or dedifferentiated tumour and an oncologist proposes it; marrow suppression, hair loss, secondary leukaemia risk (Dhall's series includes a child who developed leukaemia after treatment; Larkin 2020, PMID 31764514, asks the same question).

f. UK availability. Ornamental plant, not sold as a supplement. Etoposide hospital-only.

g. Verdict. Origin of a real drug; plant toxic; etoposide is the one agent on this list with a plausible, if thin, role in paediatric or poorly differentiated chordoma, decided by the oncologist after histology, not by the family.

6. Happy tree (Camptotheca acuminata): camptothecin, irinotecan, topotecan

a. Compounds and mechanism. Camptothecin, a quinoline alkaloid from the bark and fruit, inhibits topoisomerase I by trapping the enzyme-DNA complex, so replication forks collide with it and DNA breaks. Camptothecin itself was too toxic and insoluble (haemorrhagic cystitis, unpredictable diarrhoea) in 1970s trials; the semi-synthetic derivatives irinotecan and topotecan solved this and became standard drugs.

b. Licensed drug versus plant. Irinotecan and topotecan are licensed hospital chemotherapy (colorectal, pancreatic, small cell lung, ovarian, cervical, paediatric sarcomas). The tree is a Chinese ornamental; it is not a food, not sold as a supplement in the UK, and crude extracts carry camptothecin's own toxicity. There is nothing for a family to buy.

c. Evidence tiers. As drugs: very large randomised base. As a plant: none.

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (camptothecin OR irinotecan OR topotecan): 8 records, including the only prospective chordoma trial of any compound on this list: - Chugh and colleagues, "Phase II study of 9-nitro-camptothecin in patients with advanced chordoma or soft tissue sarcoma", Journal of Clinical Oncology 2005;23:3597-3604, PMID 15908670. 15 patients with advanced or metastatic chordoma took oral 9-nitrocamptothecin (rubitecan) 1.25 mg/m2 five days a week. One objective response (7 percent); median time to progression 9.9 weeks; 3- and 6-month progression-free survival 47 and 33 percent; anaemia, leukopenia, fatigue, nausea and diarrhoea common, 10 percent stopped for toxicity. The authors' conclusion: "modest activity in delaying progression" with "moderate toxicity". Rubitecan was never licensed. - Kato and colleagues, Radiation Oncology 2011;6:116, PMID 21914223. A fast-growing U-CH1 subline tested against X-rays, carbon and iron ions, camptothecin, mitomycin C, cisplatin and bleomycin; bleomycin was the most cytotoxic chemical, and the main finding was the high relative biological effectiveness of heavy ions (relevant to why proton and carbon therapy are used). - Wang and colleagues, Cell Death and Disease 2019;10:887, PMID 31767864. Irinotecan (and doxorubicin) switched on cytokeratin 8 in U-CH1 and CM319 chordoma cells, which drove resistance; knocking it down re-sensitised them. In vitro and xenograft. A resistance paper. - Cavalheiro 2026 (section 5d) includes camptothecin-class agents in its ex vivo panel. - Falcón 2025 and 2026, Ponce 2026: lurbinectedin plus irinotecan phase I and pooled safety analyses in advanced solid tumours; chordoma is indexed but no chordoma-specific result is reported. Chondrosarcoma: 3 records: Li 2024 (camptothecin among top in vitro hits, section 2d); Aydın 2021 (amygdalin combinations in SW1353 cells, PMID 33733877); an electroporation methods paper.

e. Safety for the patient. Not consumable as a plant. Irinotecan's severe diarrhoea and neutropenia, and its CYP3A4/UGT1A1 metabolism, would make it a difficult drug alongside her regimen; academic, since no one is proposing it.

f. UK availability. Drugs hospital-only. No plant product.

g. Verdict. Origin of two real drugs; the only compound on the list actually trialled in chordoma, where it produced one response in fifteen at meaningful toxicity; irrelevant as a plant.

7. Turmeric (Curcuma longa): curcumin

a. Compounds and mechanism. Curcumin (with demethoxy- and bisdemethoxycurcumin, together "curcuminoids", about 2 to 5 percent of turmeric root) is a polyphenol reported to inhibit NF-kappaB, STAT3, COX-2, cyclin D1, VEGF and dozens of other targets, to induce apoptosis and to chelate iron. The breadth is itself a warning: medicinal chemists classify curcumin as a PAINS compound (pan-assay interference), chemically unstable, poorly absorbed, rapidly metabolised, and prone to giving positive results in almost any dish assay; the definitive critique is Nelson and colleagues, "The essential medicinal chemistry of curcumin", Journal of Medicinal Chemistry 2017;60:1620-1637, which found no double-blind placebo-controlled trial of curcumin had succeeded for any indication at that date.

b. Licensed drug versus plant. Not a medicine anywhere. Sold as a spice and as food supplements: plain 95 percent curcuminoid extracts (typically 500 to 1500 mg a day), extracts with piperine (black pepper alkaloid) to block glucuronidation and raise absorption up to twenty-fold, and formulated products (Meriva phytosome, Theracurmin, BCM-95, Longvida) with claimed 10 to 30-fold better absorption. Content in supplements is reasonably reliable for reputable brands; the safety problem (e) comes with the high-absorption products, not with the spice. Culinary turmeric delivers perhaps 50 to 200 mg of curcuminoids a day, almost none of it absorbed.

c. Evidence tiers. - In vitro and animal: thousands of papers across every tumour type; NCI PDQ "Curcumin (Curcuma, Turmeric) and Cancer" summarises them and notes the translation problem (https://www.cancer.gov/about-cancer/treatment/cam/hp/curcumin-pdq). - Human, cancer treatment: a 2023 systematic review found seven randomised trials, 2016 to 2022, in prostate, colorectal and breast cancer, multiple myeloma and oral leukoplakia (PMC10144810, "Exploring the contribution of curcumin to cancer therapy: a systematic review of randomized controlled trials"); results mixed, mostly biomarker endpoints, small samples, high dropout. A 2024 critical analysis of bioavailability and outcome trials reaches the same conclusion (PMC11074217). - The best-known trial: Howells and colleagues, "Curcumin combined with FOLFOX chemotherapy is safe and tolerable in patients with metastatic colorectal cancer in a randomized phase IIa trial" (CUFOX), Journal of Nutrition 2019;149:1133-1139, PMC6602900. 28 patients, open-label, 2 g curcumin daily plus FOLFOX versus FOLFOX; primary endpoint safety, met; an exploratory overall survival hazard ratio of 0.34 (95 percent CI 0.14 to 0.82) that the authors present as hypothesis-generating from a tiny unblinded sample. No confirmatory trial has reported. - Golombick 2012, American Journal of Hematology (PMID 22473809): 36 patients with MGUS or smouldering myeloma, 4 g curcumin crossover, small falls in free light chain ratio and a bone resorption marker; biomarker only. - With radiotherapy: a prostate radiotherapy RCT (Hejazi 2016, NCT01917890) found no difference in outcomes with curcumin; a cervical cancer RCT (Indonesia 2021, PMC8184198) reported lower serum survivin in the curcumin arm, a biomarker. Randomised trials of curcumin for radiation mucositis and dermatitis show modest symptomatic benefit in some (PMC10108802, 2023) and none in others. CurChexia (PMC9184232, 2022): a phase IIa in head and neck cancer cachexia reported improved muscle mass; small. - Systematic review of curcumin as a complementary treatment in oncology (PMC11695395, 2024): supportive-care signals, no proven effect on tumour outcomes.

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (curcumin OR curcuma OR turmeric): zero. Chondrosarcoma: 8 records, two relevant, both laboratory: - Lee and colleagues, "Curcumin induces cell apoptosis in human chondrosarcoma through extrinsic death receptor pathway", International Immunopharmacology 2012;13:163-169, PMID 22522053. Curcumin killed JJ012 and SW1353 chondrosarcoma cells but not normal chondrocytes, via Fas/FasL/DR5 and p53; in a mouse xenograft, 21 days of curcumin reduced tumour volume by about 60 percent. Dose and route in the mouse are not comparable to oral supplement use in a person; the cell lines are from conventional chondrosarcoma, not skull base. - Kalinski and colleagues, "Curcumin blocks interleukin-1 signaling in chondrosarcoma cells", PLoS One 2014;9:e99296, PMID 24901233. Curcumin blocked IL-1beta-induced NF-kappaB activation, VEGF-A expression and in vitro angiogenesis in C3842 and SW1353 cells. In vitro only. This is the only "supplement" plant on the list with any tumour-specific laboratory data, and it is for the lookalike diagnosis, not for chordoma, and stops at the mouse.

e. Safety for the patient. - Liver: this is now a well-documented hazard. LiverTox (NBK548561) grades turmeric/curcumin category A, a well-documented cause of clinically apparent liver injury: several dozen published cases; the US DILIN series (Halegoua-DeMarzio and colleagues, American Journal of Medicine 2023) found 10 cases 2011 to 2022, 80 percent women, hepatocellular pattern, half jaundiced, one death; HLA-B*35:01 present in over 70 percent of cases versus 10 to 15 percent of controls; an Italian cluster in 2019 traced to high-bioavailability and piperine-containing products (Lombardi and colleagues, British Journal of Clinical Pharmacology 2021;87:741-753, Tuscany phytovigilance analysis); more piperine-product cases in 2023 to 2025 (Annals of Internal Medicine Clinical Cases 2023; Case Reports in Gastroenterology 2025; Cureus 2025, PMID 41281037). Onset typically 1 to 4 months; most recover within 1 to 3 months of stopping. She is a woman, will be on paracetamol regularly and possibly carbamazepine (both hepatically handled), and will need clean liver function tests through surgery and proton therapy; the case for avoiding high-absorption curcumin products is strong. - Enzyme and transporter interactions: curcumin and especially piperine inhibit CYP3A4, CYP2C9 and P-glycoprotein and glucuronidation (Welsh Medicines Advice Service, "Turmeric: potential interactions", https://www.wmic.wales.nhs.uk/turmeric-potential-interactions/; Bahramsoltani 2017, Journal of Ethnopharmacology review of curcuminoid pharmacokinetic interactions, PMID 28734960). A raised INR with bleeding risk has been reported with warfarin. In principle this would raise levels of oxycodone, fentanyl, methadone, tramadol (3A4 or P-gp), dexamethasone (3A4) and carbamazepine (3A4), and could increase sedation from opioids. Human data are thin: the review above found only one clinical trial demonstrating a significant interaction. Gabapentin and pregabalin are not affected (renal clearance). Duloxetine is CYP1A2/2D6, no established interaction. - Bleeding and surgery: antiplatelet effect in vitro and in animals; the standard perioperative advice (MSKCC, ASA herbal guidance) is to stop turmeric supplements at least two weeks before surgery, and skull base surgery with basilar artery encasement is not the operation to take a bleeding-risk chance on. Clinical judgement, but conservative and cheap to follow. - Radiotherapy: the evidence runs both ways. In dishes curcumin radiosensitises via reactive oxygen species and thioredoxin reductase inhibition (Javvadi 2008, Molecular Pharmacology; Javvadi 2010, Cancer Research 70:1941) but it is also an antioxidant and radioprotector of normal tissue (review PMC8869399, "Radiosensitization and radioprotection by curcumin in glioblastoma and other cancers"). No human trial has measured tumour control with curcumin during radiotherapy for any tumour, and the general antioxidant caution during radiotherapy (Bairati 2005, Journal of Clinical Oncology, worse local control with beta-carotene and vitamin E in head and neck radiotherapy) is the standard reason oncologists ask patients to stop high-dose antioxidant supplements during treatment. Proton therapy relies on the same radical chemistry as photons. - Cognition and sedation: no impairment; meta-analyses of curcumin and cognition find small improvements or no effect (Frontiers in Nutrition 2025; PMC8708668). This is the one item on the list with no cognitive downside. - Iron and weight: curcumin is an iron chelator in animals (Jiao 2009, Blood 113:462); a randomised human study of a bioavailable curcumin with ferrous sulphate found no impairment of iron absorption (PMC8308323, 2021). Curcumin reduces weight in metabolic syndrome trials (PMC6582779), which is the wrong direction for a patient at 51 kg, though the effect is small.

f. UK availability. Spice: pennies. Supplements: legal, everywhere; plain 95 percent extract 500 mg about 10 to 15 pounds a month; Meriva or Theracurmin products 20 to 50 pounds a month. FSA has no specific upper limit; EFSA's acceptable daily intake for curcumin as a food additive is 3 mg per kg (about 150 mg a day for the patient), which is far below supplement doses. Quality of reputable brands is reasonable; contamination (lead chromate in adulterated turmeric powder from South Asia) is a documented problem with loose spice from unverified sources.

g. Verdict. Safe and pleasant as a spice; as a high-dose supplement it has small, unconfirmed human trial signals in other cancers, laboratory data only for chondrosarcoma, none for chordoma, a real liver-injury risk (particularly with piperine), and interaction and bleeding issues that make it something to stop before surgery and probably during proton therapy.

8. European mistletoe (Viscum album): Iscador, Helixor, mistletoe lectins

a. Compounds and mechanism. Mistletoe lectins I, II and III (ribosome-inactivating proteins that block protein synthesis and induce apoptosis), viscotoxins (small cytotoxic proteins), and polysaccharides. In humans the proposed action of the injected extracts is immunomodulation: rises in NK cells, cytokines and DNA-repair markers, and a fever response. The anthroposophic preparations (Iscador, Helixor, Abnoba, Iscucin) are fermented or unfermented aqueous extracts standardised by host tree (apple "M", pine "P", oak "Qu", fir "A") rather than by lectin content, and are given by subcutaneous injection two or three times a week, escalating dose. Oral mistletoe is inert and the berries are toxic.

b. Licensed drug versus plant. In the UK Iscador holds a limited anthroposophic product licence and is a prescription-only medicine; since April 2018 it is not NHS-funded and is prescribed privately, mainly through integrative medicine centres (NCIM Bristol patient information, February 2025; Yes to Life; https://ncim.org.uk). Helixor is available on private prescription as an unlicensed import. In Germany and Switzerland mistletoe is the most commonly prescribed complementary cancer medicine and is reimbursed. There is no "mistletoe supplement" worth the name; tablets and tinctures sold online have no evidence and no standardisation.

c. Evidence tiers. Mistletoe has more randomised human evidence than any other plant on this list, and that evidence is instructive. - Tröger and colleagues, "Viscum album [L.] extract therapy in patients with locally advanced or metastatic pancreatic cancer", European Journal of Cancer 2013;49:3788-3797. 220 patients on best supportive care only (no chemotherapy), open-label, randomised to Iscador Qu or nothing; the interim analysis showed longer overall survival on mistletoe and the trial was stopped early. Widely quoted; unblinded, no placebo, stopped early, and the control group had no active treatment at all. - MISTRAL: Wode and colleagues, "Mistletoe extract in patients with advanced pancreatic cancer: a double-blind, randomized, placebo-controlled trial", Deutsches Ärzteblatt International 2024;121:347-354, PMC11539882. 290 patients at Swedish oncology centres (143 mistletoe, 147 placebo), stratified by chemotherapy eligibility, subcutaneous injections three times a week for nine months. No effect on the primary endpoint, overall survival; the authors conclude there is "no clinical reason to recommend mistletoe extract in patients with advanced pancreatic cancer". The health-related quality of life analysis (Palliative Medicine 2026) is published and did not show a convincing benefit either. This is the best-designed mistletoe trial ever run and it is negative. - Hofinger and colleagues, "Systematic assessment of the influence of quality of studies on mistletoe in cancer care on the results of a meta-analysis on overall survival", 2024, PMC11056339. Randomised trials at low risk of bias: hazard ratio 0.78, not statistically significant; trials at high risk of bias: 0.66, significant. In other words the survival benefit lives in the weaker studies. - Ostermann 2022 (PMID 36324298), meta-analysis of survival with non-fermented extracts: positive but built largely on the same lower-quality studies. - Cochrane: Horneber and colleagues, "Mistletoe therapy in oncology", Cochrane Database of Systematic Reviews 2008, CD003297, PMID 18425885: 21 trials, too heterogeneous to pool; some evidence of quality of life benefit in breast cancer during chemotherapy; survival evidence weak. An updated Cochrane review (Wider 2022, CD014782) is at protocol stage. - NCI PDQ "Mistletoe Extracts": reports of improved survival or quality of life are common but "many studies had major weaknesses"; use "cannot be recommended outside the context of well-designed clinical trials". - Intravenous mistletoe: Paller, Nelson and colleagues (Johns Hopkins), "Phase I trial of intravenous mistletoe extract in advanced cancer", Cancer Research Communications 2023;3:338-346, PMID 36860652. 21 patients, Helixor M intravenously, maximum tolerated dose 600 mg, disease control (stable disease mostly) in 24 percent for a median 15 weeks, quality of life scores rose. Phase I, no control group. - Brain tumours: Lenartz and colleagues, "Survival of glioma patients after complementary treatment with galactoside-specific lectin from mistletoe", Anticancer Research 2000;20:2073-2076, PMID 10928154: randomised, ML-1 added to standard care; overall no meaningful prolongation of relapse-free or overall survival. Mouse glioma work (PMC6129785, PMC6636536, 2018 to 2019) reports synergy with radiochemotherapy; preclinical. - Quality of life and fatigue: the most consistent signal across reviews (Cochrane 2008; a 2018 systematic review in Journal of Cancer Research and Clinical Oncology found benefits mainly in fatigue and chemotherapy tolerance in lower-quality trials). The patient is not having chemotherapy. - A UK pilot RCT in breast cancer (MAB, University of Bristol) has a published protocol but I found no results.

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (mistletoe OR viscum OR iscador): zero. Chondrosarcoma: zero. Nothing in any cell line, animal or patient; Helixor's own 38-cell-line panel (which showed pine extract the most cytotoxic in vitro) did not include a chordoma line.

e. Safety for the patient. - Common: injection-site inflammation, fever, flu-like symptoms, fatigue, headache; these are dose-related and expected; anaphylaxis is rare but reported. - Brain tumours: the standard cautions (CancerChoices safety page; product literature) advise close supervision in primary brain tumours or brain metastases because fever and inflammatory responses can raise intracranial pressure. Her tumour is extradural and skull base rather than intracerebral, but it sits against the brainstem with cranial nerve compression; a febrile, inflammatory treatment in the peri-operative period is not attractive. Clinical judgement. - Steroids: dexamethasone suppresses exactly the immune responses mistletoe is supposed to provoke; the combination is common in practice but makes the rationale weaker. - No CYP interactions known; no sedation or cognitive effects; no bleeding effect. This is why it scores reasonably on safety despite the lack of efficacy. - Radiotherapy: no human data on tumour control; mouse glioma data suggest no antagonism. - Surgery: not a bleeding risk; but fevers in the days before or after skull base surgery would confuse the picture (infection versus reaction) so most prescribers pause it around operations.

f. UK availability. Private prescription only. NCIM Bristol (2025 leaflet): 274 pounds per five weeks of clinical supervision for the first 20 weeks, plus ampoules at 80 to 138 pounds per 2.5 weeks depending on dose, plus postage; roughly 2,500 to 3,500 pounds in the first year, self-injected subcutaneously two or three times a week. Other providers (Integrated Medicine London) similar.

g. Verdict. The plant medicine with the most human trials; the only large double-blind one is negative for survival; possible quality of life benefit in weaker studies; injectable, expensive, no chordoma data; not dangerous, but not a treatment.

9. Black seed (Nigella sativa): thymoquinone

a. Compounds and mechanism. Thymoquinone (0.2 to 2.5 percent of the seed oil, highly variable by origin and pressing) is a quinone reported to act as both antioxidant and, at higher concentrations, pro-oxidant; in cancer cells it is reported to inhibit NF-kappaB, STAT3, PI3K/Akt and to induce apoptosis and cell-cycle arrest, and to radiosensitise breast cancer cells in vitro (Velho-Pereira 2011, PMID 21557727). The seed also contains nigellone, alpha-hederin and fixed oils. As with curcumin, the target list is long and the human data short.

b. Licensed drug versus plant. Not a medicine anywhere. Sold as a culinary seed (used in Indian and Middle Eastern cooking), as cold-pressed oil (about 10 pounds per 100 ml), and as capsules. Thymoquinone content in commercial oils varies several-fold and is rarely stated; no UK standard applies.

c. Evidence tiers. - In vitro and animal: extensive, including glioblastoma models (ScienceDirect review 2024 of thymoquinone in glioblastoma) and small cell lung cancer (PMID 39206985, 2024). - Human, cancer: very little. The best is a randomised trial of a Nigella sativa muco-adhesive tablet in 48 patients with oral leukoplakia (a precancer), Egypt, 2024, which reported smaller lesions and lower Ki-67 than placebo (Exploration of Medicine 2024); a supportive-care RCT in 80 children with brain tumours on chemotherapy reported fewer episodes of febrile neutropenia with Nigella sativa (Mousa 2017, Childs Nervous System). A 2025 systematic review of thymoquinone clinical trials (ScienceDirect, "Thymoquinone and therapeutic potentials: updated evidences from clinical trials") found ten completed studies across diabetes, cancer, COVID-19, epilepsy and periodontitis and seven ongoing; none shows a tumour outcome benefit. Larger RCT evidence exists for glycaemic control, blood pressure, lipids and asthma, which is not what the family is asking about.

d. Chordoma and chondrosarcoma. PubMed (chordoma) AND (thymoquinone OR nigella): zero. Chondrosarcoma: zero.

e. Safety for the patient. - Liver: LiverTox (NBK591552) grades black cumin seed E, unlikely to cause liver injury: no published reports of hepatotoxicity despite wide use; trials show if anything small falls in ALT. - Interactions: Nigella and thymoquinone inhibit CYP2D6, CYP3A4 and CYP2C9 in vitro and in animals (MSKCC monograph, https://www.mskcc.org/cancer-care/integrative-medicine/herbs/nigella-sativa; thymoquinone-warfarin and thymoquinone-phenytoin interaction studies, Chemico-Biological Interactions 2022, PMID 35921950). The warfarin paper estimates that more than about 18 mg thymoquinone a day, or more than about 1 g of seed or oil a day, could alter warfarin handling; clinical cases have not been reported. The same enzymes handle codeine and tramadol (2D6), oxycodone and fentanyl (3A4), dexamethasone and carbamazepine (3A4), and diclofenac (2C9). At culinary amounts this is not a concern; at supplement doses it is a theoretical one and no clinical study exists. - Sedation and cognition: this is the item the partner should know about. Aqueous and methanolic Nigella seed extracts have a "potent sedative and depressive effect on the central nervous system" in animals, attributed partly to GABA content (Al-Naggar 2003, Journal of Ethnopharmacology, PMID 12902052; review PMC4884225). Human trials at 1 to 3 g a day do not report sedation, but she will be on opioids, possibly diazepam and possibly a gabapentinoid, and a fourth CNS depressant, however weak, is not what her partner wants. Animal work also shows thymoquinone attenuating morphine tolerance and withdrawal (PMC4884218), which is interesting but purely preclinical. - Glucose and blood pressure: lowers both in meta-analyses (PMC6950756; PMC11195321), modestly; relevant if dexamethasone pushes her glucose up and then black seed pulls it down unpredictably, and relevant to peri-operative blood pressure management. Minor at food doses. - Bleeding: reduced platelet counts in mice; no human bleeding reports; the sensible perioperative rule (stop two weeks before) applies as for turmeric. - Radiotherapy: radiosensitiser in breast cancer cells, radioprotector in rat lung and brain (PMID 38235870, 2024); the same antioxidant ambiguity as curcumin; no human data. - Weight: no adverse effect; some trials show small weight loss in obese subjects.

f. UK availability. Legal, cheap, everywhere; oil 10 to 27 pounds per 100 ml, seeds pennies. No quality standard; thymoquinone content unstated on most products.

g. Verdict. Safe as a food and probably as a low-dose oil; essentially no human anticancer evidence; no chordoma data; theoretical CYP and sedation issues at supplement doses; stop before surgery.

10. Chordoma, chondrosarcoma and chemotherapy: the context for all nine

Conventional chordoma is regarded as resistant to cytotoxic chemotherapy. The 2015 global consensus (Stacchiotti, Sommer and the Chordoma Global Consensus Group, "Building a global consensus approach to chordoma", Lancet Oncology 2015;16:e71-e83, PMID 25638683) and the ESMO-EURACAN-GENTURIS-ERN PaedCan bone sarcoma guideline (Strauss and colleagues, Annals of Oncology 2021;32:1520-1536) give conventional chemotherapy no established role; systemic treatment for advanced chordoma is imatinib and other targeted or trial agents, after surgery and high-dose radiotherapy (protons or carbon ions) have been used. The cell-line papers above (Bayrak 2011, Jian 2018, Wang 2019, Kim 2016, Tecimel 2026) are largely studies of why chordoma resists drugs, including the taxane, camptothecin and etoposide classes on this list. Skull base conventional chondrosarcoma is likewise treated with surgery and radiotherapy; chemotherapy is reserved for dedifferentiated and mesenchymal subtypes (Terek 1998; Fox 2012; ESMO 2021).

Two exceptions matter for section 12. Poorly differentiated chordoma (SMARCB1/INI-1 loss, mostly children and young adults, mostly skull base and cervical spine, more aggressive) and dedifferentiated chordoma are treated more like high-grade sarcomas, and the paediatric literature above (Dhall 2011, Al-Rahawan 2012, Wong 2023, Cavalheiro 2026) is where etoposide, ifosfamide, vincristine and doxorubicin appear. That is a decision for the oncologist after histology and immunohistochemistry (brachyury, INI-1), not something the family can act on.

What this means for the nine plants: even where a plant is the origin of a real chemotherapy drug, that drug has either been shown to be resisted by chordoma cells or has produced at most one response in a small trial; and the supplement plants have no chordoma data at all.

11. Ranking

Scores 0 to 5 on each criterion. Criterion 1 is scored twice where the plant is the origin of a licensed drug, because the two questions are different; the overall ranking uses the "as the family could use it" score, since that is the question asked.

Plant 1. Human evidence of anticancer benefit (drug / as usable by family) 2. Relevance to chordoma or chondrosarcoma 3. Safety in the patient (as usable by family) 4. UK practicality alongside standard care Total (usable form)
7 Turmeric n/a / 2 1 (chondrosarcoma dish and mouse; chordoma zero) 3 at supplement dose, 5 as spice 5 11 (13 as spice)
9 Black seed n/a / 1 0 3 at supplement dose, 4 as food 5 9 (10 as food)
8 Mistletoe n/a / 2 0 3 2 7
4 Sweet wormwood (artesunate) 2 / 0 (supplement is not the trialled drug) 0 2 2 4 (6 if artesunate under a prescriber)
6 Happy tree (camptothecins) 5 / 0 2 (the only chordoma trial: 1 of 15) 0 (plant not consumable; drug toxic) 0 2
5 Mayapple (etoposide) 5 / 0 2 (paediatric clival series; cell lines resistant) 0 0 2
2 Yew (taxanes) 5 / 0 1 (resistance papers; 2 of 25 in chondrosarcoma) 0 (lethal plant) 0 1
1 Periwinkle (vinca) 5 / 0 1 (no benefit in U-CH1 or in reported cases) 0 (toxic plant) 0 1
3 Graviola n/a / 0 0 0 to 1 (neurotoxin) 3 (legal and cheap, which is the problem) 3

Overall ranking with justification

  1. Turmeric. The only "supplement" plant with any tumour-specific laboratory data (chondrosarcoma, the lookalike, not chordoma), a few small and unconfirmed human trial signals in other cancers, no cognitive downside, and a well-characterised risk profile that can be managed by keeping to culinary or modest doses, avoiding piperine and high-absorption products, and stopping two weeks before surgery and during proton therapy. Acceptable adjunct; not a treatment.
  2. Black seed. Safe as food, unlikely to harm the liver, but almost no human anticancer evidence and none for chordoma; theoretical enzyme inhibition and animal sedation signals argue against supplement doses on top of opioids and diazepam. Acceptable at culinary levels; nothing more.
  3. Mistletoe. The most-studied plant medicine and therefore the most honest lesson: when finally tested double-blind at scale it did not extend survival. Possible quality of life and fatigue benefit from weaker trials. Injectable, private, roughly 3,000 pounds a year, no chordoma data, and fever reactions that are unwelcome around skull base surgery. "Only if the family wants it for wellbeing, with the oncologist informed, and not before recovery from surgery."
  4. Sweet wormwood / artesunate. Scientifically the most interesting repurposing candidate, but the evidence is a 23-patient pilot in a different cancer with the confirmatory trial unreported; the shop supplement is not the drug; and the drug's enzyme induction would destabilise her opioid, steroid and possibly carbamazepine levels. Only ever via the oncologist or a trial, and there is no chordoma trial. 5 to 8. Happy tree, mayapple, yew, periwinkle. These are the origins of irinotecan/topotecan, etoposide, paclitaxel/docetaxel and vincristine/vinblastine. As plants they are poisonous (yew fatally so) and there is nothing for the family to buy or take. As drugs they are handled by oncologists, and for conventional chordoma the evidence is resistance in the dish and one response in fifteen in the only trial. Etoposide (5) and the camptothecins (6) edge above the taxanes and vinca because they are the only ones with any chordoma patient data; that data is paediatric and thin, and only becomes relevant if histology shows a poorly differentiated or dedifferentiated tumour. This is a ranking of pharmacology history, not of options.
  5. Graviola. Last, and separately: no human evidence, a mitochondrial toxin with an epidemiological link to atypical parkinsonism, sold without standards, and specifically unsuited to a patient whose brainstem and cranial nerve function will be under close examination for the next year. The likely source of the "strongest compounds" claim, and the one to say no to.

The honest one-line ranking: none of these treats chordoma; turmeric and black seed at food doses are safe adjuncts with conditions; mistletoe and artesunate are prescriber-only questions with no chordoma evidence; the four drug-origin plants are irrelevant as plants and toxic; graviola is dangerous.

12. What would change the picture

13. Questions the family can put to the oncology team

  1. Which of the patient's medicines are metabolised by CYP3A4, CYP2D6 or CYP2C9 (opioids, dexamethasone, carbamazepine if used, diclofenac), and do you want her to avoid herbal supplements that inhibit or induce those enzymes?
  2. What is your policy on supplements, particularly antioxidants, in the two weeks before surgery and during proton therapy? Is culinary turmeric or black seed acceptable, and at what point should any supplement stop?
  3. Will you record any supplement she takes in her notes and tell the anaesthetist and the proton centre?
  4. Once histology is back: is this conventional, poorly differentiated or dedifferentiated chordoma (or chondrosarcoma), and does that change whether any chemotherapy has a role?
  5. Is there any systemic therapy or radiosensitiser trial for skull base chordoma or chondrosarcoma she could be considered for, in the UK or Europe?
  6. Given her weight loss, would you rather she saw a dietitian than added supplements, and is there anything she should specifically avoid because of the trigeminal pain and appetite?
  7. If the family want to use mistletoe or another prescribed complementary product later, would you object, and would you want to be told the product, dose and prescriber?

14. Sources

PubMed searches run 4 September 2026 via NCBI E-utilities: (chordoma) AND each compound set; (chondrosarcoma) AND each compound set. Zero results recorded for chordoma with graviola/acetogenins, artemisinins, curcumin, mistletoe and thymoquinone; zero for chondrosarcoma with graviola, artemisinins, mistletoe and thymoquinone.

Chordoma and chondrosarcoma - Stacchiotti S, Sommer J, Chordoma Global Consensus Group. Lancet Oncol 2015;16:e71-e83. PMID 25638683. - Strauss SJ, et al. ESMO-EURACAN-GENTURIS-ERN PaedCan bone sarcoma guideline. Ann Oncol 2021;32:1520-1536. - Bayrak OF, et al. J Neurosurg Spine 2011;15:620-624. PMID 21905773. - Azzarelli A, et al. J Surg Oncol 1988;37:185-191. PMID 3352273. - Scimeca PG, et al. J Pediatr Hematol Oncol 1996;18:237-240. PMID 8846149. - Al-Rahawan MM, et al. Pediatr Blood Cancer 2012;59:323-325. PMID 21922644. - Dhall G, et al. J Neurooncol 2011;103:657-662. PMID 21052774. - Wong CI, et al. Pediatr Blood Cancer 2023. PMID 37218455. - Larkin T, et al. J Pediatr Hematol Oncol 2020. PMID 31764514. - Cavalheiro S, et al. J Cancer Res Clin Oncol 2026;152:147. PMID 42191910. - Jian J, et al. Am J Transl Res 2018;10:936-947. PMID 29636883. - Kim JY, et al. J Neurosurg Spine 2016;25:626-635. PMID 27314550. - Xiao D, et al. Colloids Surf B 2022;212:112366. PMID 35144131. - Tecimel D, et al. Naunyn Schmiedebergs Arch Pharmacol 2026;399:2481-2493. PMID 40888882. - Chugh R, et al. J Clin Oncol 2005;23:3597-3604. PMID 15908670. https://ascopubs.org/doi/10.1200/JCO.2005.02.170 - Kato TA, et al. Radiat Oncol 2011;6:116. PMID 21914223. - Wang D, et al. Cell Death Dis 2019;10:887. PMID 31767864. - Fox E, et al. Oncologist 2012;17:321. PMID 22363068. - Li L, et al. Int J Mol Sci 2024;25:13003. PMID 39684713. - Terek RM, et al. J Orthop Res 1998;16:585-590. PMID 9820282. - Gorelik B, et al. Cancer Res 2008. PMID 18974149. - Lee HP, et al. Int Immunopharmacol 2012;13:163-169. PMID 22522053. - Kalinski T, et al. PLoS One 2014;9:e99296. PMID 24901233. - Aydın D, et al. J Med Food 2021. PMID 33733877. - SMARCB1 and poorly differentiated chordoma: Am J Pathol 2023, PMC10123523; NCI PDQ Childhood Chordoma Treatment, https://www.cancer.gov/types/bone/hp/child-chordoma-treatment-pdq

Periwinkle and yew - Catharanthus roseus intoxication mimicking cholangitis. BMC Complement Med Ther 2024. PMC10993546. - Grobosch T, et al. Eight cases of fatal and non-fatal poisoning with Taxus baccata. Forensic Sci Int 2013. - Fatal heart block from Taxus baccata ingestion. Eur Heart J Case Rep 2020;4:1. https://academic.oup.com/ehjcr/article/4/1/1/5686032 - Cardiotoxicity of yew. Cardiovasc Toxicol 2014 (ScienceDirect S0010865014000952).

Graviola - Caparros-Lefebvre D, Elbaz A. Lancet 1999;354:281-286. - Lannuzel A, et al. Brain 2007;130:816-827. - Champy P, et al. Quantification of acetogenins in Annona muricata linked to atypical parkinsonism in Guadeloupe. Mov Disord 2005. - Höllerhage M, et al. Neurotoxicity of dietary supplements from Annonaceae species. Int J Toxicol 2015;34:543-553. https://journals.sagepub.com/doi/10.1177/1091581815602252 - Evaluation of the quality of Annona muricata leaf products. J Pharm Pharmacol 2023;75:1357. - Review of A. muricata anticancer activity. PMC9497149 (2022). - MSKCC Graviola monograph: https://www.mskcc.org/cancer-care/integrative-medicine/herbs/graviola - ASCO Post, Graviola, 2019: https://ascopost.com/issues/october-10-2019/graviola/

Artemisia annua and artesunate - Krishna S, et al. EBioMedicine 2015;2:82-90. PMID 26137537. https://www.thelancet.com/article/S2352-3964(14)00034-6/fulltext - NeoART-V, NCT03093129: https://clinicaltrials.gov/study/NCT03093129 (no results found 4 Sep 2026). - von Hagens C, et al. Breast Cancer Res Treat 2017;164:359-369. PMID 28439738. Long-term follow-up: Phytomedicine 2019, PMID 30668363. Ototoxicity: Cancer Chemother Pharmacol 2016. - Deeken JF, et al. Cancer Chemother Pharmacol 2018 (phase I IV artesunate). - Reichert S, et al. Radiother Oncol 2012. PMID 22560712. Luo J, et al. Radiat Oncol 2014;9:84. - Xing J, et al. Drug Metab Dispos 2012;40:1757. PMID 22679214. Autoinduction in volunteers: Malar J 2014;13:214, PMID 24889062. Burk O, et al. Mol Pharmacol 2005 (PXR/CAR induction). - LiverTox, Artemisinin: https://www.ncbi.nlm.nih.gov/books/NBK548419/ - Brewer TG, et al. Am J Trop Med Hyg 1994 (brainstem neurotoxicity in animals). - Artesunate Amivas MHRA authorisation April 2022: https://amivas.com/amivas-ireland-ltd-artesunate/; EMA EPAR: https://www.ema.europa.eu/en/medicines/human/EPAR/artesunate-amivas - Artemisia tea artemisinin content: J Pharm Biomed Anal 2012 (S0731708512000192); Molecules 2026;31:1854 (supplement quality). - Mechanism reviews: Front Pharmacol 2024 (anti-tumour mechanism of artesunate); Hu 2022 Oxid Med Cell Longev.

Mayapple and etoposide - Pharmacognostic and toxicological profile of Podophyllum peltatum. 2023 review (ScienceDirect S2667142523001392). - Two cases of poisoning by the fruit of Podophyllum peltatum (historical). PMC9987978.

Turmeric and curcumin - Nelson KM, et al. The essential medicinal chemistry of curcumin. J Med Chem 2017;60:1620-1637. - NCI PDQ Curcumin and Cancer (HP): https://www.cancer.gov/about-cancer/treatment/cam/hp/curcumin-pdq - Systematic review of curcumin RCTs in cancer, 2023: PMC10144810. Critical analysis of bioavailability and outcome trials, 2024: PMC11074217. Complementary treatment systematic review, 2024: PMC11695395. - Howells LM, et al. J Nutr 2019;149:1133-1139. PMC6602900. - Golombick T, et al. Am J Hematol 2012. PMID 22473809. - Prostate radiotherapy RCT NCT01917890 (Hejazi 2016); cervical radiotherapy RCT PMC8184198; mucositis RCT PMC10108802; CurChexia PMC9184232. - Radiosensitisation and radioprotection review: PMC8869399. Javvadi P, et al. Mol Pharmacol 2008, PMID 18252805; Cancer Res 2010;70:1941. - Bairati I, et al. J Clin Oncol 2005;23:5805-5813. - LiverTox, Turmeric: https://www.ncbi.nlm.nih.gov/books/NBK548561/ - Halegoua-DeMarzio D, et al. Liver injury associated with turmeric, a growing problem: ten cases from the DILIN. Am J Med 2023. - Lombardi N, et al. Br J Clin Pharmacol 2021;87:741-753. https://bpspubs.onlinelibrary.wiley.com/doi/10.1111/bcp.14460 - Turmeric-piperine liver injury cases: Ann Intern Med Clin Cases 2023; Case Rep Gastroenterol 2025;19:96; Cureus 2025, PMID 41281037. - Welsh Medicines Advice Service, Turmeric potential interactions: https://www.wmic.wales.nhs.uk/turmeric-potential-interactions/ - Bahramsoltani R, et al. Pharmacokinetic interactions of curcuminoids with conventional drugs. J Ethnopharmacol 2017. PMID 28734960. - Jiao Y, et al. Curcumin is a biologically active iron chelator. Blood 2009;113:462. Iron absorption RCT: PMC8308323. - Cognition meta-analyses: Front Nutr 2025 (fnut.2025.1549509); PMC8708668. - Weight meta-analysis: PMC6582779.

Mistletoe - Tröger W, et al. Eur J Cancer 2013;49:3788-3797. - Wode K, et al. MISTRAL. Dtsch Arztebl Int 2024;121:347-354. PMC11539882. HRQoL: Palliat Med 2026, https://doi.org/10.1177/02692163261437609. Protocol: Trials 2020, PMC7488501. - Hofinger et al. 2024 meta-analysis by risk of bias: PMC11056339. - Ostermann T, et al. 2022. PMID 36324298. - Horneber M, et al. Cochrane 2008, CD003297. PMID 18425885. Wider B, et al. Cochrane protocol 2022, CD014782. - NCI PDQ Mistletoe Extracts (HP). - Paller CJ, et al. Cancer Res Commun 2023;3:338-346. PMID 36860652. PMC9973409. - Lenartz D, et al. Anticancer Res 2000;20:2073-2076. PMID 10928154. - Mouse glioma: PMC6129785; PMC6636536. - Safety and brain tumour caution: https://cancerchoices.org/therapy/mistletoe-european/safety/ - UK access and cost: NCIM Mistletoe Therapy patient information Feb 2025, https://ncim.org.uk/wp-content/uploads/2025/02/NCIM-Mistletoe-Therapy-Patient-Information-Feb25.pdf; https://yestolife.org.uk/therapy/mistletoe-therapy/; https://integratedmedicine.co/therapies/mistletoe-therapy/cost/

Black seed - LiverTox, Black Cumin Seed: https://www.ncbi.nlm.nih.gov/books/NBK591552/ - MSKCC Nigella sativa monograph: https://www.mskcc.org/cancer-care/integrative-medicine/herbs/nigella-sativa - Thymoquinone-warfarin: Chem Biol Interact 2022, PMID 35921950. Thymoquinone-phenytoin: Chem Biol Interact 2022 (S0009279722000060). - Oral leukoplakia RCT, 2024: https://www.explorationpub.com/Journals/em/Article/1001290 - Mousa HFM, et al. Childs Nerv Syst 2017 (febrile neutropenia in children with brain tumours). - Thymoquinone clinical trials review 2025 (ScienceDirect S2950199725001417). Glioblastoma review 2024 (S2950199724000958). - Radiosensitisation in breast cancer cells: PMID 21557727. Radioprotection in rat lung: PMID 38235870. - CNS depressant activity of extracts: Al-Naggar TB, et al. J Ethnopharmacol 2003, PMID 12902052; review PMC4884225. Morphine tolerance: PMC4884218. - Glycaemic and endothelial meta-analyses: PMC6950756; PMC11195321.

Addendum, 10 September 2026: COX-2 inhibitors (a friend's question)

Addendum, 10 September 2026: holy basil (tulsi), from a friend's papers

Research thread 11 · source file em-research/11-plant-compounds.md · last edited 10 September 2026