Food, Dietary Patterns, and Cancer Prevention: Evidence Review Through September 22, 2026

Review period: 1994-01-01 to 2026-09-22
Scope: Primary prevention evidence for foods, beverages, dietary patterns, harmful exposures, supplements, and a separate review of nutrition during treatment and survivorship.
Source policy: Priority was given to authoritative cancer-prevention reviews, systematic reviews and meta-analyses, randomized trials, prospective cohorts, and original primary sources; laboratory and animal findings are explicitly separated from human outcomes.

Executive summary

The most defensible dietary strategy for cancer prevention is not a single “superfood,” supplement, extract, or detox regimen. It is a long-term pattern rich in whole grains, legumes, vegetables and fruit; low in processed meat and alcohol; moderate in red meat; and supportive of healthy body weight. WCRF/AICR and the American Cancer Society emphasize the overall pattern rather than isolated nutrients. (pressroom.cancer.org)

The strongest food-specific protective evidence concerns whole grains and dietary fibre for colorectal cancer. Coffee probably lowers liver and endometrial cancer risk, although it is not necessary to begin drinking coffee for prevention. Evidence for Mediterranean-style and other high-quality plant-forward patterns is increasingly persuasive—particularly for colorectal and breast cancer—but most incidence data remain observational. (wcrf.org)

The clearest dietary hazards are processed meat, alcohol, and probably high red-meat intake. Processed meat causally increases colorectal cancer risk; approximately 50 g/day was associated with an 18% relative increase. Alcohol causally increases cancers of the oral cavity, pharynx, larynx, oesophagus, liver, colorectum and breast, with risk beginning at low intake. Ultra-processed food is an emerging concern, especially for colorectal cancer, but its evidence remains less causal because all incidence studies are observational and separation of processing from poor diet quality, excess energy intake and adiposity is difficult. (publications.iarc.who.int)

Evidence for broccoli-derived sulforaphane, berries, lycopene, garlic, onions, leafy greens, fish and nuts ranges from limited to moderate. These are reasonable components of a healthy diet, but none has a proven stand-alone anticancer dose. Biomarker changes—such as altered detoxification products, gene expression or cellular proliferation—must not be interpreted as demonstrated reductions in cancer incidence. (pmc.ncbi.nlm.nih.gov)

Randomized trials do not support routine vitamin D, omega-3, selenium, vitamin E, beta-carotene or antioxidant supplements for primary cancer prevention. Beta-carotene increases lung-cancer risk in smokers or asbestos-exposed people, and 400 IU/day vitamin E increased prostate-cancer incidence in SELECT. Supplements remain appropriate for a diagnosed deficiency, pregnancy-related needs or another clinical indication—not as self-prescribed cancer prophylaxis. (cancer.gov)

No food or dietary pattern has been shown to cure cancer or replace surgery, radiotherapy, systemic treatment, screening, vaccination or tobacco cessation.

1. How the evidence was graded

• Strong: WCRF/AICR “convincing” or “probable” causal judgment, consistent high-quality prospective evidence, or replicated randomized evidence.
• Moderate: reasonably consistent prospective associations with dose-response or supporting mechanisms, but meaningful residual confounding or limited trials remain.
• Preliminary/limited: inconsistent findings, case-control evidence stronger than cohort evidence, few events, surrogate biomarkers, or predominantly laboratory/animal research.
• Evidence of no benefit or harm: adequately powered randomized trials show no preventive effect or identify adverse outcomes.

Nutrition research faces recurring limitations: food-frequency questionnaires measure diet imprecisely; healthy diets correlate with screening, exercise, education and lower smoking; dietary components are highly correlated; reference categories and serving sizes vary; and decades-long randomized cancer-incidence trials are difficult. Publication and selective-reporting bias are credible concerns—for example, the WCRF fibre analysis detected funnel-plot asymmetry and a significant Egger test despite an overall protective conclusion. (wcrf.org)

Interpreting relative and absolute risk

A relative risk of 1.18 does not mean that 18% of consumers will develop cancer. Illustratively, if a comparable population had a 10-year risk of 1%, an RR of 1.18 corresponds to about 1.18%, or 0.18 additional percentage point. If baseline risk were 4%, it would correspond to about 4.72%, or 0.72 additional percentage point. These examples are mathematical illustrations, not individualized forecasts.

2. Evidence map

3. Cruciferous vegetables, broccoli and sulforaphane

Human cancer-incidence evidence

A 2024 meta-analysis reported lower overall cancer incidence among higher broccoli consumers in prospective cohorts (RR 0.89, 95% CI 0.82–0.96). The association was substantially larger in case-control studies (OR 0.64), a pattern that raises concern about recall and selection bias. Site-specific benefits were generally confined to case-control rather than cohort analyses. (pmc.ncbi.nlm.nih.gov)

A prospective lung-cancer meta-analysis found an inverse association for cruciferous vegetables (RR 0.82, 95% CI 0.75–0.91), but smoking is a particularly difficult residual confounder: people who eat more vegetables often smoke less and differ in socioeconomic status, occupational exposures and screening. (pubmed.ncbi.nlm.nih.gov)

Judgment: Crucifers are healthful vegetables and contribute fibre, folate and micronutrients, but evidence that broccoli itself materially prevents a particular cancer is limited to moderate observational evidence, not proof from incidence trials.

Sulforaphane trials

Sulforaphane is generated when glucoraphanin in cruciferous vegetables is hydrolysed by plant or microbial myrosinase. Human trials have mostly measured intermediate endpoints rather than cancer.

• In a randomized trial of 170 adults, a high-dose broccoli-sprout beverage increased urinary benzene detoxification products by 63.2% over ten days; lower doses were not significantly effective. This showed altered carcinogen metabolism—not fewer cancers. (pmc.ncbi.nlm.nih.gov)
• A 12-week randomized trial in 291 adults found 61% greater excretion of a benzene metabolite and 23% greater excretion of an acrolein metabolite after a beverage supplying 600 μmol glucoraphanin plus 40 μmol sulforaphane daily. Again, incidence was not measured. (pmc.ncbi.nlm.nih.gov)
• In a phase II trial of 43 high-risk former smokers, 37 completed 12 months. A preparation estimated to deliver 95 μmol sulforaphane daily produced a 20% decline in bronchial Ki-67 versus a 65% increase with placebo (P=0.014), but did not improve bronchial histopathology, apoptosis markers or cancer incidence. Small size, baseline imbalance and pandemic-related adherence disruptions limit inference. (pmc.ncbi.nlm.nih.gov)

Mechanisms

Laboratory and animal evidence suggests activation of NRF2-regulated detoxification enzymes, modulation of glutathione pathways, inhibition of histone deacetylases, altered cell-cycle control and apoptosis. These mechanisms establish biological plausibility, not clinical prevention. (pubmed.ncbi.nlm.nih.gov)

Practical conclusion: Include broccoli, cabbage, cauliflower, Brussels sprouts, kale and related vegetables regularly for dietary variety. Sulforaphane supplements cannot presently be recommended for cancer prevention.

4. Berries, tomatoes, garlic, onions and leafy greens

Berries

Berry anthocyanins and ellagitannins have antiproliferative, anti-inflammatory and epigenetic effects in laboratory models. Human evidence is largely limited to small biomarker studies in people with precancerous lesions or existing disease.

A phase I study gave 20 colorectal-cancer patients 60 g/day of freeze-dried black-raspberry powder for one to nine weeks and reported favourable changes in selected methylation, proliferation and angiogenesis biomarkers. It had no randomized control group and did not test recurrence or survival. (pubmed.ncbi.nlm.nih.gov)

Judgment: Preliminary. There is no demonstrated berry dose that prevents clinical cancer.

Tomatoes and lycopene

A recent prospective-cohort meta-analysis found no association between total tomato intake and breast cancer (RR 1.04) or prostate cancer (RR 1.02). Higher dietary lycopene was associated with a small reduction in overall cancer incidence (RR 0.95), but estimates varied by exposure measurement and cancer site; blood lycopene may also mark a generally healthier diet and metabolism. (pmc.ncbi.nlm.nih.gov)

Randomized evidence has not established that lycopene supplements prevent prostate cancer; a WCRF review of four trials reported RR 0.92 (95% CI 0.66–1.29). (wcrf.org)

Judgment: Tomatoes are nutritious, but claims that tomato sauce or lycopene supplements specifically prevent prostate cancer remain unproven.

Garlic, onions and other alliums

Prospective evidence is inconsistent. A systematic review concluded that neither allium vegetables nor garlic supplements produced a clinically relevant reduction in overall cancer incidence, and prospective meta-analysis found no colorectal-cancer benefit from garlic. (pmc.ncbi.nlm.nih.gov)

Organosulfur compounds can affect carcinogen metabolism, inflammation and cell signalling in laboratory systems, but supplement bioavailability and dose differ greatly from ordinary foods.

Judgment: Use garlic and onions as flavourful plant foods, not as anticancer treatments or pills.

Leafy greens

A 2025 meta-analysis of prospective studies reported lower lung-cancer risk with greater green-leafy vegetable intake (RR 0.85, 95% CI 0.76–0.94). This is plausible through folate, carotenoids, fibre and displacement of less-healthful foods, but site specificity and smoking-related residual confounding reduce certainty. (pubmed.ncbi.nlm.nih.gov)

Judgment: Limited site-specific evidence, stronger support as part of an overall vegetable-rich pattern.

5. Whole grains, dietary fibre and legumes

Whole grains and fibre: the strongest protective food evidence

WCRF judges whole grains and foods containing dietary fibre to decrease colorectal-cancer risk and recommends at least 30 g/day of fibre from foods. (wcrf.org)

An influential dose-response meta-analysis estimated approximately 17% lower colorectal-cancer risk per 90 g/day—roughly three servings—of whole grains. The updated WCRF analysis found RR 0.93 per 10 g/day higher fibre intake, based on 16,562 cases, although the upper confidence limit was 1.00, heterogeneity was high (I²=72%) and Egger testing indicated small-study bias. (bmj.com)

Contradictory randomized evidence

The Polyp Prevention Trial prescribed a low-fat diet with high fibre, fruit and vegetables but did not reduce adenoma recurrence over four years or continued follow-up. A separate wheat-bran trial likewise found no significant recurrence reduction (OR 0.88, 95% CI 0.70–1.11). (pubmed.ncbi.nlm.nih.gov)

These null results do not necessarily refute primary prevention: participants already had adenomas, intervention began late in carcinogenesis, adenoma recurrence is an imperfect surrogate, fibre type matters, and four years may be too short. They do show that adding bran or changing diet after polyp removal is not a proven substitute for colonoscopic surveillance.

Legumes

Beans, lentils, chickpeas and peas supply fibre, resistant starch, folate and plant protein. Evidence for legumes as an isolated anticancer exposure is insufficient; WCRF therefore recommends them as part of the whole-grain/vegetable/fruit/bean pattern rather than assigning a cancer-specific dose. (wcrf.org)

Mechanisms

Humanly plausible mechanisms include greater stool bulk and faster transit, dilution of luminal carcinogens, microbial fermentation to short-chain fatty acids such as butyrate, better insulin sensitivity, and displacement of processed meat and refined carbohydrates. (wcrf.org)

Practical quantity: Build gradually toward approximately 30 g/day fibre using whole grains, legumes, vegetables, fruit, nuts and seeds; increase fluids as fibre rises.

6. Fish, omega-3 fatty acids and nuts

Fish

WCRF classifies fish and colorectal cancer as limited–suggestive rather than strong evidence. A prospective meta-analysis found a 4% lower colorectal-cancer risk per 50 g/day (RR 0.96, 95% CI 0.92–0.99), but one large study supplied 66.5% of the statistical weight, and the apparent benefit plateaued around 50 g/day. (wcrf.org)

Fish may also work indirectly by replacing processed or red meat. Definitions vary substantially—oily versus white fish, fried versus baked, fresh versus salted—and contaminants and preparation can differ.

Omega-3 supplements

The VITAL randomized trial assigned 25,871 adults to 1 g/day marine omega-3 containing 460 mg EPA and 380 mg DHA or placebo. Over a median 5.3 years, invasive cancer occurred in 820 versus 797 participants (HR 1.03, 95% CI 0.93–1.13): no preventive benefit. (nejm.org)

Interpretation: Eating fish and taking fish-oil capsules are not equivalent interventions. The modest observational fish association may involve food substitution, other nutrients or residual confounding.

Nuts

Prospective meta-analyses have sometimes suggested lower cancer incidence or mortality with greater nut intake, but three large US cohorts with repeatedly measured diets found no association with total cancer or common site-specific cancers. (aacrjournals.org)

In PREDIMED, a Mediterranean diet supplemented with 30 g/day mixed nuts produced a nonsignificant breast-cancer HR of 0.53 (95% CI 0.23–1.26), based on only ten cases in that arm. (jamanetwork.com)

Judgment: Nuts are valuable for overall cardiometabolic health and as replacements for processed snacks or meat, but no cancer-preventive nut dose is established.

7. Coffee, green tea and other beverages

Coffee

WCRF judges coffee to probably protect against liver and endometrial cancers. The endometrial association occurs with caffeinated and decaffeinated coffee, arguing against caffeine as the sole explanation. No threshold or universally applicable preparation has been established. (wcrf.org)

Umbrella reviews find the most credible inverse associations for liver and endometrial cancer, while associations at other sites are less robust and sometimes show publication or excess-significance bias. Potential mechanisms include improved insulin sensitivity, altered oestrogen metabolism, lower liver inflammation and actions of chlorogenic acids and other polyphenols. (bmj.com)

Practical conclusion: For adults who tolerate it, unsweetened coffee can fit a preventive diet. Evidence does not justify starting coffee, consuming very high doses, or adding large amounts of sugar and cream solely for cancer prevention.

Green tea

The Cochrane review included 142 studies and more than 1.1 million observational participants, but only 1,795 randomized participants. Results were inconsistent. In trials, prostate-cancer incidence had RR 0.50 but a very wide CI of 0.18–1.36; overall observational incidence was RR 0.83, 95% CI 0.65–1.07. (cochrane.org)

High-dose extracts can cause gastrointestinal adverse effects and elevated liver enzymes. Therefore, brewed green tea is a reasonable low-calorie beverage, but concentrated catechin/EGCG supplements should not be presented as cancer prevention.

Sugar-sweetened beverages and fruit juice

Associations between sugary drinks and individual cancers are generally low-certainty, but these beverages facilitate excess energy intake and weight gain, which is causally related to multiple cancers. A dose-response review found positive associations with several cancer sites but rated the absolute effects small and certainty low or very low. Whole fruit is preferable to juice because of fibre and lower intake rate. (pubmed.ncbi.nlm.nih.gov)

8. Mediterranean and plant-based dietary patterns

Mediterranean pattern

A Mediterranean-style pattern generally emphasizes vegetables, fruit, legumes, whole grains, nuts, olive oil and fish, with limited processed meat, refined foods and—under modern cancer-prevention guidance—little or no alcohol.

The 2025 WCRF review of 170 studies concluded that integrated healthy dietary and lifestyle patterns have strong evidence for lower colorectal- and breast-cancer risk. For breast cancer, however, the WCRF/AICR and ACS scores had stronger evidence than the Mediterranean score alone, underscoring the importance of weight, activity, alcohol and smoking alongside food. (wcrf.org)

PREDIMED randomized evidence

Among 4,152 women without previous breast cancer, PREDIMED identified only 35 invasive cases over a median 4.8 years. Rates were:

• Mediterranean diet plus extra-virgin olive oil: 1.1 per 1,000 person-years;
• Mediterranean diet plus nuts: 1.8 per 1,000 person-years;
• control diet: 2.9 per 1,000 person-years.

The adjusted HR was 0.31 (95% CI 0.13–0.77) for the olive-oil arm and 0.53 (0.23–1.26) for the nut arm. The absolute rate difference for olive oil versus control was about 1.8 cases per 1,000 person-years. Breast cancer was a secondary endpoint, the event count was very small, and participants were older Spanish women at high cardiovascular risk, limiting generalizability. (jamanetwork.com)

Vegetarian, vegan and plant-based patterns

A 2026 meta-analysis of prospective studies found vegetarian diets associated with lower total cancer incidence (RR 0.87, 95% CI 0.84–0.91), colorectal cancer (RR 0.86), colon cancer (RR 0.79) and breast cancer (RR 0.92). For total cancer, four studies included approximately 613,000 participants and 65,523 cases. These results remain observational, and vegetarian status can proxy lower smoking, lower alcohol use, greater screening and other health-conscious behaviours. (pmc.ncbi.nlm.nih.gov)

“Plant-based” quality matters. A diet dominated by refined grains, sweets, fries and sugary drinks may be plant-based yet metabolically poor; recent reviews distinguish healthful from unhealthful plant-based indices. (pubmed.ncbi.nlm.nih.gov)

Judgment: Moderate evidence supports high-quality plant-forward and Mediterranean-style patterns. Complete exclusion of animal foods has not been shown necessary for cancer prevention.

9. Red meat, processed meat, alcohol and ultra-processed foods

Processed meat

IARC classified processed meat as carcinogenic to humans based on sufficient evidence for colorectal cancer after reviewing more than 800 studies. An analysis of ten studies estimated an 18% relative increase in colorectal-cancer risk per 50 g/day. Classification identifies causal confidence, not equivalence in hazard magnitude to smoking. (publications.iarc.who.int)

Likely mechanisms include haem iron, endogenous N-nitroso compounds, nitrites/nitrates, salt, and heterocyclic amines or polycyclic aromatic hydrocarbons generated by high-temperature cooking. The contribution of each pathway in humans is not fully quantified.

Red meat

Evidence for unprocessed red meat is weaker than for processed meat but is judged probably causal for colorectal cancer. WCRF recommends no more than approximately 350–500 g cooked weight weekly and little, if any, processed meat. (wcrf.org)

Red meat is nutritionally dense, so the evidence supports moderation and substitution—not the proposition that one serving inevitably causes cancer.

Alcohol

Alcohol is an established carcinogenic exposure. WCRF finds strong evidence for cancers of the mouth, pharynx, larynx, oesophagus, liver, colorectum and postmenopausal breast, and probable evidence for premenopausal breast and stomach cancer. (wcrf.org)

For oral-cavity cancer, WCRF estimated RR 1.15 per additional 10 g/day ethanol. A global analysis attributed approximately 741,300 incident cancers in 2020—4.1% of all cases—to alcohol; about 103,100 were associated with consumption below 20 g/day. (wcrf.org)

Mechanisms include acetaldehyde-mediated DNA damage, oxidative stress, altered folate metabolism, increased oestrogen and synergistic injury with tobacco. Wine is not exempt, and resveratrol does not neutralize ethanol’s carcinogenicity.

Practical conclusion: From a cancer perspective, not drinking is best. People who drink should reduce frequency and quantity; no type of alcohol is protective against cancer.

Ultra-processed foods

Recent systematic reviews associate high UPF intake with colorectal and other digestive cancers, but certainty is below that for processed meat or alcohol. All cancer-incidence evidence is observational, UPF definitions can vary, and residual confounding by fibre, adiposity, smoking, socioeconomic status and overall dietary quality is difficult to remove. (pubmed.ncbi.nlm.nih.gov)

Plausible pathways include passive overconsumption and weight gain, low fibre, high glycaemic load, displacement of protective foods, emulsifier-related microbiome changes, processing contaminants and packaging chemicals. It is not yet clear how much risk is attributable to processing itself rather than nutrient profile and excess energy intake.

Practical conclusion: Reduce frequently consumed processed meats, packaged sweets, salty snack foods, fast food and sugary drinks. “Ultra-processed” is not synonymous with uniformly harmful: fortified whole-grain bread, unsweetened yogurt and some convenient legume products can contribute useful nutrients.

10. Supplements and “anti-cancer superfood” claims

Randomized evidence

The USPSTF recommends against beta-carotene and vitamin E for cancer prevention and finds evidence insufficient for routine multivitamins or most other single/paired supplements. (uspreventiveservicestaskforce.org)

Popular claims without clinical-incidence proof

Claims involving turmeric/curcumin, medicinal mushrooms, apricot kernels/amygdalin, alkaline diets, ketogenic diets, juicing, megadose vitamin C, concentrated antioxidants, detox teas and proprietary “immune” powders generally rest on laboratory data, uncontrolled studies, pharmacologic speculation or selective testimonials. NCI states that no diet, food, vitamin, mineral, herb or supplement has been proven to cure cancer, slow it reliably or prevent recurrence. (cancer.gov)

Food-derived nutrients and high-dose purified supplements are not biologically interchangeable. During radiotherapy or chemotherapy, antioxidant, herbal and concentrated botanical products may alter drug metabolism or treatment-related oxidative injury and should be disclosed to the oncology team. (cancer.gov)

11. Primary prevention versus treatment and recurrence prevention

Primary prevention

This is where evidence is strongest. The priority hierarchy is:

1. avoid tobacco and limit or avoid alcohol;
2. maintain a healthy body weight and remain physically active;
3. emphasize fibre-rich plant foods and whole grains;
4. minimize processed meat and moderate red meat;
5. use healthy dietary patterns rather than supplements or isolated superfoods. (cancer.org)

During active treatment

Nutrition goals can differ sharply from prevention advice. Maintaining energy intake, protein, hydration, body weight and muscle may be more important than achieving an idealized high-fibre diet, particularly with nausea, diarrhoea, obstruction, mucositis, altered taste or swallowing difficulty. ACS recommends malnutrition screening, individualized medical nutrition therapy and ongoing surveillance. (cancer.org)

Restrictive fasting, ketogenic, raw-food, juice-only or supplement-heavy regimens can worsen malnutrition or interact with treatment. Decisions should involve the oncology team and, where possible, an oncology-credentialed registered dietitian.

Recurrence and survivorship

Evidence is much less secure than for primary prevention. WCRF explicitly notes a lack of randomized trials and advises survivors, where clinically appropriate, to follow general prevention recommendations after acute treatment. (wcrf.org)

The WHEL randomized trial assigned 3,088 early-stage breast-cancer survivors to a diet very high in vegetables, fruit and fibre and low in fat or to standard advice. After 7.3 years:

• recurrent/new breast-cancer event: 16.7% versus 16.9%; HR 0.96;
• death: 10.1% versus 10.3%; HR 0.91.

Neither difference was significant despite substantial dietary change. (jamanetwork.com)

Recent WCRF reviews suggest that healthier patterns may improve outcomes after breast or colorectal cancer, but much of the evidence is observational and subject to confounding by disease severity, treatment, physical activity and prediagnosis health. (wcrf.org)

Clinical conclusion: A healthy diet supports overall health and may reduce second-cancer and chronic-disease risks, but no food has been shown to prevent recurrence reliably. Standard oncology treatment and surveillance remain essential.

12. Practical evidence-based dietary framework

1. Make fibre-rich plant foods the foundation. Include vegetables or fruit at most meals, use whole grains rather than refined grains, and eat legumes regularly. Work toward approximately 30 g/day food-derived fibre. (wcrf.org)
2. Prioritize variety over superfoods. Rotate cruciferous and leafy vegetables, tomatoes, onions, garlic, berries, citrus and other produce. Different colours and food families provide different nutrients; no single item needs to be consumed daily.
3. Choose protein sources strategically. Use beans, lentils, soy foods, fish, poultry and nuts more often; keep red meat within roughly 350–500 g cooked per week and processed meat as close to zero as practical. (wcrf.org)
4. Avoid alcohol for maximum cancer-risk reduction. If drinking continues, less is better; wine has no cancer-preventive exemption. (wcrf.org)
5. Use water as the default beverage. Unsweetened coffee or tea can fit individual tolerance, but green-tea extracts and other concentrated products are unnecessary.
6. Reduce foods that promote excess energy intake. Limit sugary beverages, packaged sweets, frequent fast food and highly processed snack products. This supports weight control even where direct cancer causality remains uncertain.
7. Do not use high-dose supplements as insurance. Correct genuine deficiencies and meet condition-specific needs under medical guidance. Avoid beta-carotene in smokers and routine high-dose vitamin E. (uspreventiveservicestaskforce.org)
8. Pair diet with the larger prevention package. Weight, activity, tobacco, vaccination, infection treatment and age-appropriate screening often have larger and better-established effects than any individual food. (wcrf.org)

A simple meal template is approximately half non-starchy vegetables and fruit, one quarter whole grains or other minimally processed starch, and one quarter legumes, fish or another minimally processed protein, with nuts/seeds or unsaturated plant oils as appropriate.

13. Important unresolved questions

1. Can long-duration randomized dietary-pattern interventions reduce cancer incidence without relying on surrogate markers?
2. Which fibre types, whole grains and microbial metabolites provide the greatest colorectal protection, and how does baseline microbiome composition modify response?
3. Do sulforaphane-related biomarker changes translate into fewer lung, prostate or colorectal cancers, and what food preparation maximizes safe bioavailability?
4. Are coffee’s liver and endometrial associations causal, and how do roast, brewing method, additives, decaffeination and genetics alter them?
5. Does fish itself lower colorectal risk, or is the association explained by replacing processed/red meat and other lifestyle factors?
6. Which components of a Mediterranean pattern drive any breast-cancer benefit, and can PREDIMED’s result be replicated with substantially more cancer events?
7. Is UPF-associated risk caused by processing, additives and packaging, or mainly by excess energy intake, adiposity, low fibre and poor overall diet quality?
8. How do healthy plant-based patterns compare with equally high-quality omnivorous patterns under conditions of similar weight, alcohol, activity and screening?
9. Which dietary interventions during and after treatment improve recurrence or survival without worsening malnutrition or treatment tolerance?
10. How should dietary recommendations be individualized by cancer subtype, genetic risk, metabolic health, treatment and socioeconomic access?

Sources and direct links

Authoritative evidence reviews

• World Cancer Research Fund: evidence behind cancer-prevention recommendations
• WCRF: whole grains, vegetables, fruit and beans
• WCRF 2025 dietary and lifestyle patterns report
• American Cancer Society cancer-prevention guideline
• IARC monograph: red and processed meat
• USPSTF supplement recommendation

Primary randomized trials and original studies

• Sulforaphane in high-risk former smokers
• Broccoli-sprout beverage and pollutant detoxification
• PREDIMED Mediterranean diet and breast-cancer incidence
• VITAL omega-3 trial
• VITAL vitamin D trial
• SELECT vitamin E and prostate-cancer risk
• Physicians’ Health Study II multivitamin trial
• Polyp Prevention Trial
• Wheat Bran Fiber Trial
• WHEL breast-cancer survivorship trial
• Global cancer burden attributable to alcohol

Key systematic reviews and meta-analyses

• Whole grains, fibre and colorectal cancer
• Broccoli consumption and cancer
• Tomato, lycopene and cancer
• Allium vegetables and cancer
• Fish consumption and colorectal cancer
• Green tea and cancer, Cochrane review
• Coffee and cancer umbrella review
• Vegetarian and vegan diets and cancer incidence
• Ultra-processed foods and cancer

Bottom line: A varied, fibre-rich, minimally processed, plant-forward diet with little processed meat and no or minimal alcohol is the most evidence-based framework. Specific foods may contribute, but none cures cancer, guarantees prevention or replaces established medical care.