Microgreens and Cancer: The Compound Everyone Names Is the One Broccoli Microgreens Have Least Of

This page is not medical guidance. Cancer is not something to treat with food. If you have a diagnosis or take anticancer medicine, talk to your oncology team before changing your diet or taking any supplement.

With that said, there is a chemistry problem at the center of this claim that nobody in the category mentions.

The compound mismatch

Every cancer claim about Brassica microgreens ends at sulforaphane. Sulforaphane is not in the plant. It forms from glucoraphanin when tissue is damaged and the chemistry and microbiome permit conversion.

Broccoli microgreens are not primarily a glucoraphanin crop.

A study in the Journal of Food Composition and Analysis, read at source on 2026-08-12, reports that glucoerucin is the predominant glucosinolate in broccoli microgreens, at a concentration 20-fold higher than glucoraphanin in every microgreen sample tested. Its earlier measurement was glucoerucin 28.1 against glucoraphanin 0.9 micromoles per gram of dry weight.

Glucoraphanin is the major one in the mature tissue, making up over half the glucosinolates in florets.

So the crop everyone recommends for sulforaphane is richest in the other compound. A page selling broccoli microgreens as a sulforaphane source has the crop right and the molecule wrong.

Now the honest completion, because the overcorrection is also wrong. Glucoerucin yields erucin, and erucin and sulforaphane interconvert in people, a point documented in that same paper's own reference list. And a human study has fed fresh broccoli microgreens and measured sulforaphane bioavailability, so exposure does happen.

The accurate statement is therefore narrower and more interesting than either extreme. Broccoli microgreens deliver sulforaphane-related exposure by an indirect route through a different precursor, and none of that is cancer evidence. It is chemistry.

It also breaks a popular number. The famous comparison sets 28.1 micromoles per gram of dry weight in microgreens against 1.08 micromoles per gram of fresh weight in florets. This site has long flagged that as a unit error. It is also a compound error: the first figure is glucoerucin and the second is glucoraphanin. Two molecules, two units, reported as one ratio.

What actually exists, by rung and by material

The evidence ladder has four rungs and the claims are made from the bottom two.

1. Cells in a dish. A compound is put directly on isolated cells. Generates a hypothesis.
2. Animals. A whole organism, so absorption and a tumour environment are partly represented.
3. Human biomarker trial. People take something and an intermediate measure is taken. Shows exposure or target engagement.
4. Human outcome trial. Incidence, recurrence, progression or survival. This is the rung a prevention claim needs.

There are zero completed microgreen studies with cancer incidence, recurrence, progression or survival as an outcome. One study is active in breast cancer survivors. It runs two weeks and measures feasibility, metabolism, microbiome and blood metabolomics. It is not a cancer outcome study.

The two microgreen studies are both dishes. One reported a leukemia-cell IC50 of 446.34 micrograms per milliliter with cell-cycle arrest in 69 percent of analyzed cells. The other treated colon cancer and normal colon cells for 24 hours with diluted digestive fractions of four Brassicaceae microgreens.

What a dish result means, precisely: cells exposed directly to a stated concentration became less viable or stopped dividing. What it bypasses: chewing, conversion from a precursor, digestion, the microbiome, absorption, liver metabolism, delivery to a tumour, the immune system, and toxicity to normal tissue. It also holds a concentration steady in a way food never does.

The dose gap, which is the quiet problem

Dish experiments typically use 5 to 20 micromolar sulforaphane. One review records an ovarian cell IC50 around 8 micromolar; a lung cancer experiment reported full suppression of foci formation at 8 micromolar.

Human plasma does not get there. A pharmacokinetic study gave a single 200 micromole broccoli sprout isothiocyanate preparation and measured peak plasma isothiocyanates of 0.943 to 2.27 micromolar at one hour, with a 1.77 hour half-life.

The peak from a concentrated 200 micromole preparation is roughly three to eight times below the dish concentration, and it is gone within hours.

That is not a clean disproof and it should not be presented as one. Plasma totals include metabolites, tissue concentrations can differ from blood, and different assays measure different chemical species. What it does rule out is the casual assumption that a dish exposure follows from eating a garnish.

Animal doses do not convert either. Tumour experiments run 5 to 100 mg/kg. Multiply crudely by 70 kg and you get 350 to 7,000 mg, which is why crude multiplication is wrong. Using the body-surface-area method, 50 mg/kg in a mouse is about 244 mg a day for a 60 kg adult and 100 mg/kg is about 488 mg. Both are pharmacology screens, not food servings, and no validated conversion exists from an animal anticancer dose to an amount of any named microgreen.

What the human trials actually measured

The best human work measured carcinogen handling, not cancer.

• In Qidong, China, a randomized placebo-controlled trial used a 3-day broccoli sprout infusion and found an inverse association between urinary dithiocarbamates and aflatoxin-DNA adduct excretion, P=0.002, r=0.31. It did not measure liver cancer.
• A 2019 Qidong dose-finding trial assigned 170 adults for 10 days across placebo and three sprout beverage doses, with a benzene detoxification biomarker as its endpoint.
• A breast biopsy trial gave 54 women with abnormal mammograms a glucoraphanin supplement for 2 to 8 weeks. Ki-67 and HDAC3 fell within the supplement group, at P=0.003 and P=0.044, and the between-group changes were not significant after adjusting for multiple comparisons.

That last detail is the whole discipline of this page. A within-group change is not a result. It is the same trap the heart article found in a blood pressure trial.

And the one trial with a clinical endpoint did not meet it. A phase II study gave 20 men with recurrent prostate cancer 200 micromoles a day of sulforaphane-rich broccoli sprout extract for up to 20 weeks. One man had the prespecified PSA decline. The primary endpoint was not achieved. A 2023 systematic review found eight heterogeneous cancer trials and could not combine them.

The observational evidence is about mature vegetables and self-reported diet. A 2024 meta-analysis found broccoli intake associated with relative risk 0.89, 95 percent CI 0.82 to 0.96 in cohorts, which is an 11 percent lower relative risk, alongside a stronger case-control figure more vulnerable to recall bias.

The National Cancer Institute's own summary is the sentence to keep: cell and animal studies identified possible actions, and "studies in humans, however, have shown mixed results." Its cancer-specific review reports little or no prostate association in prospective cohorts, generally no colorectal association, varying lung findings, and no association in a pooled breast cohort analysis.

Confounding is not a technicality here. People who eat more cruciferous vegetables also differ in smoking, alcohol, weight, activity, income, screening access and total vegetable intake, and food recall is imperfect. A relative risk is an association, not a personal prediction.

Safety, which is the part that matters most

Cancer Research UK states it plainly: "There is no reliable evidence that dietary or nutritional supplements can prevent, cure or control the growth of cancer."

There is a documented clinical concern with concentrated products. Memorial Sloan Kettering reports that high-dose broccoli sprout supplements during chemotherapy in people with advanced pancreatic cancer can increase nausea and vomiting, with digestive symptoms in the cited feasibility study.

And there is a plausible interaction nobody has settled. A meta-analysis of human dietary interventions found cruciferous vegetable diets changed CYP1A2 activity by 20 to 40 percent and GST-alpha by 15 to 35 percent. Those enzymes clear many drugs. MSK says explicitly that it is not known whether broccoli sprouts do the same, and no interaction table exists for microgreens at all.

Do not read that gap as reassurance. An unstudied interaction is unknown, not absent. Give your oncology team the actual product, label, dose and frequency, especially for any extract, powder or capsule.

Two more specifics. Vitamin K and warfarin is the clearest food and drug issue here, and the instruction is consistency rather than avoidance. For thyroid disease, the internet advice to avoid all crucifers is stronger than the evidence, and there is no established universal restriction, but high-dose cruciferous extracts are worth asking about, which the thyroid article covers.

Ordinary food portions of broccoli sprouts are generally described as tolerated. The reported problems are at supplement doses. Animal toxicology showing drowsiness at 150 to 300 mg/kg and deaths at 200 to 300 mg/kg is high-dose extract in mice and is not a claim that a salad is dangerous.

What the pages get right, and where they cross

Four consumer and institutional pages were audited on 2026-08-12.

• One of four cited a microgreen-specific cancer study, and it was the colon-cell dish paper.
• Zero of four cited a human microgreen cancer study, because none exists.
• Two of four carried laboratory findings toward a people-facing claim, with language about preventing, blocking or possibly reversing cancer growth, and about fighting cancer and slowing tumour growth.
• None sold a microgreen or supplement on the page as opened.

The NCI pages are the counterexample and deserve saying so. They explicitly separate cell and animal findings from mixed human results, and the active trial listing makes no prevention claim.

The ranking pages are right about real things. Microgreens and sprouts are different stages. Brassica microgreens do contain glucosinolate precursors. There are genuine cell studies of microgreen fractions, a large sulforaphane literature, and an active microgreen study in cancer survivors. Dismissing all of that as marketing would be wrong.

The accurate position is narrower than both extremes: real early-stage science, a material and compound mismatch, and no human microgreen cancer outcome.

What to actually do

• Do not use microgreens instead of cancer screening, prevention advice or treatment. Nothing here supports it and the stakes are not symmetrical.
• Tell your oncology team about any supplement, with the label, dose and frequency. Extracts are the concern, not a salad.
• Eat vegetables, including these, because a varied vegetable-rich diet is good practice. That is the claim the evidence supports.
• Read the material label on every claim: microgreen, sprout, sprout powder, extract, purified compound, mature vegetable, and whether it was cells, animals or people.
• Read the endpoint too. A biomarker, a PSA value or a dish result is not incidence, recurrence or survival.
• Treat a within-group change as no result, as in the breast biopsy trial whose between-group comparison did not survive adjustment.
• Do not buy broccoli microgreens specifically for glucoraphanin. The mature florets are the glucoraphanin tissue; the microgreen is mostly glucoerucin.
• Ignore any "X times more cancer-fighting compound" figure unless it names the compound, the tissue, the cultivar and the weight basis. The famous one gets two of those wrong.

What nobody has measured

• Any human microgreen study with cancer incidence, recurrence, progression or survival. The gap this whole page is about.
• A validated conversion from an animal anticancer dose to an amount of a named microgreen cultivar.
• A stable compositional ratio for microgreens against mature tissue, across cultivar, harvest day, growing system, storage and serving basis.
• Whether the erucin route delivers meaningfully, given that broccoli microgreens supply mostly glucoerucin and the interconversion is documented but unquantified for this food.
• Any specific interaction between an ordinary microgreen portion and a named cancer therapy.
• A dependable upper safe limit for any microgreen or extract product.
• A universal thyroid restriction or safe threshold, in defined thyroid and iodine status groups.

Sources

Opened 2026-08-12. Where a page returned a bot check, its indexed abstract was the accessible record and that is noted. The glucosinolate figures in the opening section were read at source on 2026-08-12 after the publisher's challenge was cleared, which settled a labeling conflict between two of this site's own research files.

• Lu et al., broccoli microgreen glucosinolates, Journal of Food Composition and Analysis, 2018 - that glucoerucin is the predominant glucosinolate in broccoli microgreens at 20-fold the glucoraphanin level in every sample, the earlier measurement of 28.1 against 0.9 micromol/g dry weight, that glucoraphanin is the major one in mature leaves and florets at over half of floret glucosinolates, and, in its own reference list, Clarke et al. 2011 on the bioavailability and interconversion of sulforaphane and erucin in humans.
• National Cancer Institute, cruciferous vegetables and cancer prevention and its diet risk factors page - the proposed mechanisms observed in cells and animals, the statement that human studies have shown mixed results, the cancer-by-cancer cohort summary, and the confounding discussion. The fact sheet was reviewed in June 2012 and its date is stated here for that reason.
• NCI, the active broccoli microgreen trial - the two-week feasibility, metabolism, microbiome and metabolomics study in breast cancer survivors, which has no cancer outcome endpoint.
• NCI on enhancing drug discovery and its patient-derived models program - why conventional cell lines are screening and hypothesis tools rather than clinical evidence. The NIBSC account of TGN1412 is included as the documented example of preclinical work failing to predict a first human trial.
• Applied Biological Research, broccoli microgreen THP-1 experiment, 2024 and Carazo-Diaz et al., Caco-2 microgreen experiment, 2020 - the only two microgreen cancer studies found, both cell culture: the IC50 of 446.34 micrograms per milliliter with 69 percent cell-cycle arrest, and the 24-hour treatment of colon cancer and normal colon cells with diluted bioaccessible fractions of four Brassicaceae microgreens. The second returned a bot check.
• Dashwood and Ho, 2009 - the human pharmacokinetics: a single 200 micromole preparation producing peak plasma isothiocyanates of 0.943 to 2.27 micromolar at one hour with a 1.77 hour half-life. The dose gap on this page rests on it. Returned a bot check.
• A 2017 sulforaphane chemoprevention review and Liu et al., lung cancer cell study, 2015 - the 5 to 20 micromolar dish concentrations and the roughly 8 micromolar IC50 figures compared against that plasma result.
• A 2021 sulforaphane review and Bauman et al., mouse oral cancer model, 2016 - the 5 to 100 mg/kg animal dose range, the review's own warning about crude body-weight multiplication, and the reduced incidence and size of induced tongue tumours.
• Kensler et al., Qidong aflatoxin trial, 2005 and Chen et al., Qidong benzene trial, 2019 - the sprout infusion and the inverse association with aflatoxin-DNA adduct excretion at P=0.002 and r=0.31, and the 170-adult, 10-day dose-finding trial with a benzene detoxification biomarker endpoint. Both returned bot checks.
• Atwell et al., breast biopsy trial, 2015 with its trial record - 54 women, a glucoraphanin supplement, the within-group Ki-67 and HDAC3 falls at P=0.003 and P=0.044, and the between-group changes not significant after multiple-comparison adjustment.
• Cipolla et al., recurrent prostate cancer phase II, 2015 and a 2023 systematic review of sulforaphane cancer trials - 20 men at 200 micromoles a day for up to 20 weeks with one prespecified PSA response and the primary endpoint not achieved, and the eight heterogeneous trials that could not be combined. A prostate biopsy trial report, a sulforaphane absorption study and a multi-targeted prevention review supply the surrounding exposure and pharmacokinetic context. All returned bot checks.
• A 2024 broccoli observational meta-analysis - the cohort relative risk of 0.89 with its confidence interval and the stronger case-control odds ratio. Returned a bot check.
• Fahey et al., 1997, a 2022 microgreen maturity comparison and a human broccoli microgreen bioavailability study - the classic 20-fold sprout against mature comparison which is not a microgreen measurement, the six-species comparison which returned an internal error and is not used for a numerical ratio, and the study that fed fresh broccoli microgreens and measured sulforaphane bioavailability.
• Memorial Sloan Kettering on broccoli sprouts, its herb and botanical FAQs and its warfarin information - the increased nausea and vomiting with high-dose supplements during chemotherapy in advanced pancreatic cancer, the CYP1A2 change of 20 to 40 percent and GST-alpha of 15 to 35 percent with the explicit statement that it is not known whether broccoli sprouts do the same, and the vitamin K consistency advice.
• Cancer Research UK on vitamins, diet and supplements - that there is no reliable evidence that dietary or nutritional supplements can prevent, cure or control the growth of cancer, and the instruction to get advice before taking any.
• Shapiro et al., phase I sprout extract study, 2006 and a 2025 sulforaphane toxicity review - the safety, tolerance and metabolism assessment in healthy volunteers, and the high-dose mouse findings of drowsiness and hypothermia at 150 to 300 mg/kg and deaths at 200 to 300 mg/kg.
• The consumer and institutional pages audited for wording, practice only: Cleveland Clinic and NC State's Plants for Human Health Institute, alongside the two NCI pages above. Counted on 2026-08-12; a page count is not a claim about search rankings.