10/1/26  Cancer Is Your Fault: An Exploration of Agency and Risk Reduction   Author: Jeremy Smith & ChatGPT

Is that really true? That's the question I'm examining. I'll admit my bias up front: I WANT it to be true. I think saying “no, it’s not your fault” is sometimes tantamount to saying it's not within your control, and that can dramatically diminish a person’s sense of agency and hope. People are less likely to make hard decisions or radically alter their lifestyles if they believe the outcome is beyond their capacity. Fatalism makes sacrifice seem pointless.

So after admitting my bias, what I WANT is to demonstrate a scientifically defensible way that a very high percentage of cancer risk—say 90%—can be mitigated through lifestyle, prevention, surveillance, and medicine. Even higher would be great. The question is whether the science can get anywhere near that number without cheating.

There is also a personal reason for this exploration. This past summer, my mother was diagnosed with grade 4 glioblastoma at age 70. Besides the emotional devastation of this happening to someone I love deeply, another frustrating part is that it happened to probably the most health-conscious person in our family for decades. She didn't smoke or drink, maintained a healthy weight, ate mostly whole foods, exercised, and lived much of her life away from major air pollution and other environmental toxins.

Glioblastoma is also among the cancers with the fewest established modifiable causes. Looking at her life and diagnosis, it would be cruel—and scientifically difficult—to say anything other than: This was terrible luck. This was not your fault.

But that conclusion bothers me for another reason. I have declared that I intend to live to 100, and believe others can take steps to achieve the same thing. If a healthy person can simply be lightning struck by an essentially unpreventable cancer, how much control do any of us really have?

That is what this article is actually about. How much power can be reclaimed from the clutches of this seemingly random, indiscriminate monster we call cancer?

The answer lies in analyzing two different kinds of percentages: 1) how much can we reduce our probability of developing cancer, and 2) how much can we reduce our probability of eventually dying from it?

My hope going into this investigation is to reach somewhere near 90% or better, enough that we can declare some agency and responsibility in this journey. Let's see how close the evidence can get us without cheating.

Start With 40%

Globally, a 2026 IARC analysis covering 36 cancer types and 185 countries reached almost the same answer: 37.8% of new cancers—7.1 million cases in 2022—were attributable to 30 modifiable risk factors. [3]

So roughly 40% is our first hard number, but it is not necessarily our ceiling.

What Happens When People Actually Live Differently?

A 2016 Harvard study asked a different question. Its low-risk group had four characteristics: little or no smoking, no more than moderate alcohol, BMI between 18.5 and 27.5, and at least 150 minutes of moderate or 75 minutes of vigorous exercise weekly. [4] Compared with the general U.S. white population, the estimated preventable fractions were 41% of cases in women and 63% in men, and 59% of deaths in women and 67% in men. [4] The study excluded several cancers, including brain and hematologic cancers, so that number cannot simply replace the ACS's 40%. But it demonstrates something important:

A fairly ordinary definition of healthy living—not some fanatical longevity protocol—was associated with a cancer-mortality difference approaching two-thirds!

Real populations provide another clue. Ohio's Amish had age-adjusted cancer incidence only 56% of the national rate. [24] Seventh-day Adventists provide another useful comparison: in the 95,863-person Adventist Health Study, vegetarians had 12% lower overall cancer incidence than nonvegetarians, including 21% lower colorectal, 45% lower stomach, and 25% lower lymphoproliferative cancer incidence. [33]

There are also claims of spectacularly low cancer among hunter-gatherer societies, but they are hard to trust because they have shorter lifespans, and cancer more often appears in later ages. It’s also more difficult to identify and diagnose cancer incidents as it is done in modern societies.

The important point is not that these groups have discovered a secret cancer cure. It is that the average American cancer rate is demonstrably not a biological constant. These populations represent bundles of behaviors and exposures, so they cannot tell us which factor deserves which share of the credit. It is still useful, however, to itemize some the best-established levers:

Cancer Isn't Just “Bad Luck”

The famous Tomasetti and Vogelstein papers found that tissues undergoing more stem-cell divisions generally develop more cancers and later estimated that roughly two-thirds of mutations found in cancers arise from unavoidable DNA-replication errors. [6,7] The headlines translated that into a “cancer is mostly bad luck” story.

But mutations and cancers aren't the same thing. A cancer usually requires multiple alterations. If two happen randomly while a third depends upon an avoidable exposure, eliminating the third can still prevent the cancer. Normal tissues also contain cells carrying cancer-associated mutations that never become malignant tumors.

The useful model is therefore not genes VERSUS lifestyle. Rather, it is mutation PLUS environment: which abnormal cells arise, which survive, which expand, which escape immune control, and which ultimately become lethal. That creates many more intervention points than merely preventing mutations.

A year later, Yusuf Hannun's group attacked the problem differently and estimated that intrinsic factors often contributed only 10–30% of cancer risk, with extrinsic factors accounting for 70-90%. [7] Their estimates were controversial, but they demonstrated why “random mutation” does not equate to “unpreventable cancer.”

Change the Terrain

Obesity provides a useful human experiment. In the SPLENDID study, people with severe obesity who underwent bariatric surgery experienced about 2.9% cumulative incidence of obesity-associated cancer over ten years versus 4.9% in matched nonsurgical controls. [8] Their inherited genomes didn't change. Their physiology did!

Body fat, insulin/IGF signaling, sex hormones, inflammation, metabolism and immune function all interact with carcinogenesis. That doesn't prove the strong claim that cancer is primarily a metabolic disease. It supports a more useful proposition: Reduce the mutations we can reduce, and make the biological terrain as hostile as possible to the ones we cannot.

The metabolic view of cancer is here worth an introduction as well. In 1923 Otto Warburg famously observed that cancer cells consume glucose and produce lactate at unusually high rates even when oxygen is available. That phenomenon—the Warburg effect—is real and remains central to cancer biology.

Today, the stronger “metabolic theory of cancer,” associated today with Thomas Seyfried, argues that mitochondrial/metabolic dysfunction is the primary cause and mutations are largely

downstream. That universal claim goes beyond the evidence. But the weaker metabolic proposition is much more interesting: mutations alone do not determine which abnormal clones become lethal cancers.

Normal tissues can contain clones carrying cancer-associated driver mutations without containing cancer. Tumor promotion also occurs within a metabolic, inflammatory, hormonal, immune, and tissue environment. Classic nuclear-cytoplasmic transfer experiments add an intriguing wrinkle: in several cell and animal systems, tumor nuclei placed into normal cytoplasm showed suppressed tumorigenicity despite retaining tumor-associated mutations, while normal nuclei placed into tumor-cell cytoplasm could acquire malignant behavior or fail to restore a normal phenotype. [34] These experiments do not prove that mitochondrial dysfunction is the universal cause of cancer, but they do show that the nucleus does not act in isolation; cellular context can profoundly alter whether a cancer-associated genome expresses a malignant phenotype.

The useful synthesis is therefore not “mutation theory versus metabolic theory.” It is mutation plus terrain. Reduce the mutations we can reduce, and make the terrain as hostile as possible to the ones we cannot.

The Immune System Is Already Fighting Cancer

One enormous part of that terrain is immunity.

Cancer immunosurveillance is established biology. Immune cells identify and eliminate some transformed cells; surviving tumors evolve mechanisms that evade or suppress those defenses. The concept of cancer immunoediting describes this contest as elimination, equilibrium and eventual escape.

We can see the importance of immune surveillance when it fails dramatically. Severe HIV-associated immunosuppression has been associated with approximately 200-fold higher Kaposi sarcoma risk, ~15-fold higher Burkitt lymphoma risk, ~6-fold higher diffuse large B-cell and Hodgkin lymphoma risk, and 3–4-fold higher cervical-cancer risk. Much of this excess occurs because weakened immunity can no longer suppress cancer-causing viruses or virus-infected cells as effectively—one of the clearest demonstrations that immune defense is part of cancer defense. [35]

The harder question is how much a healthy person can deliberately improve ordinary immune surveillance.

Exercise is one prominent lever of control here. Human epidemiology consistently associates physical activity with lower incidence of several cancers, while animal experiments provide a possible mechanism: exercise can mobilize natural killer cells and other immune components into tumors. In one widely cited mouse experiment,

voluntary running reduced tumor incidence and growth by more than 60% across several tumor models, with NK cells playing an important role. That does not translate into a human 60% prevention number, however. Human exercise trials have not consistently shown higher resting NK-cell activity, so the immune mechanism remains harder to quantify directly. [36] What human cancer data do show is that immune state matters enormously: in an international study of 2,681 colon cancers, five-year recurrence was 8% in patients with a high tumor Immunoscore versus 32% with a low score. [37] We cannot yet translate that into “do X and improve immune surveillance by Y%,” but it makes the biological importance of antitumor immunity difficult to dismiss.

Feed the System

Diet also gives us more specific numbers than “eat healthy.”

A meta-analysis of prospective studies found approximately 10% lower colorectal-cancer risk for every additional 10 g/day of dietary fiber. Three additional daily servings—90 g—of whole grains were associated with about 17% lower risk.

Tree nuts are interesting and stronger than a single cohort might suggest. A comprehensive meta-analysis found high total-nut intake associated with 14% lower overall cancer risk; tree nuts specifically were associated with 13% lower incidence, and higher tree-nut intake with 18% lower cancer mortality. [38] Observational nutrition data can still be confounded, so these numbers do not earn an independent 13–18% slot in our final arithmetic. They do, however, make nuts a reasonable component of a cancer-conscious dietary pattern.

The gut microbiome may also connect diet, metabolism, and immunity. Gut microbes convert fiber into short-chain fatty acids such as butyrate, while microbiome composition appears capable of influencing cancer treatment. In 128 melanoma patients receiving immune-checkpoint therapy, higher dietary fiber was significantly associated with longer progression-free survival. [39] The study was observational and does not supply a universal percentage benefit, but it gives this pathway human clinical—not merely theoretical—support.

This demonstrates an important interaction among body systems: some foods may influence both the probability that cancer develops and the immune system's ability to respond after it does. The magnitude remains unclear, but the mechanism is no longer fanciful.

Push the Variables Harder

There is another limitation in conventional prevention research: “healthy” is usually fairly modest. One example is the Harvard low-risk group, mentioned earlier, required only 150 minutes of moderate exercise per week and allowed BMI as high as 27.5. [4] A massive 2023 dose-response meta-analysis found roughly 15% lower cancer mortality around the recommended activity level, with smaller additional associations at higher activity. Exercise-response curves eventually flatten, but public-health minimums plainly aren't designed to identify the maximal achievable benefit [45] .

The same problem applies to body composition, metabolic health, diet and probably other variables. Most epidemiology asks what happens when ordinary people become somewhat healthier. It rarely studies people who deliberately optimize every plausible cancer-related variable for forty years.

That extreme end of the distribution is precisely what interests me.

Attack Aging Too

Age may be the largest cancer risk factor of all. Mutations accumulate. Clones expand. Senescent cells accumulate. Immune function changes. Chronic inflammation rises. Tissue architecture deteriorates.

Yet prevention statistics normally put aging entirely in the uncontrollable column.

Geroscience asks whether some of that biology can itself be modified. In pooled prospective data, every five years of acceleration in several epigenetic-aging measures has been associated with roughly 4–9% higher subsequent cancer risk. [44]

Does that mean reducing an epigenetic clock five years automatically reduces cancer risk by 4–9%? We don't yet have that experimental data, but it creates another category that matters to this article: unknown additional control.

Today's 40% preventable-cancer estimate includes essentially none of the potential benefit of successfully slowing the biology of aging.

Knowing Your Genetic Hand

Genetics is usually placed in the uncontrollable column. That's only half right. The DNA sequence may be inherited. What happens after discovering it is not.

Population genomic screening has found actionable pathogenic variants for conditions including hereditary breast/ovarian cancer and Lynch syndrome in people who would never have been identified by ordinary family-history criteria. Discovering BRCA1/2, Lynch syndrome or another high-risk mutation can radically alter screening age, frequency and modality and sometimes justify preventive medication or surgery.

For a maximal-agency strategy, leaving your actionable inherited risks undiscovered makes little sense. The test is worth doing. In one population-screening study of 26,906 people, 1.33% carried pathogenic variants for hereditary breast/ovarian cancer, Lynch syndrome, or familial hypercholesterolemia. Those are only three pathology categories. Broader modern genomic-screening studies find still much more medically actionable variants. [40] Whole-genome sequencing and/or comprehensive clinical genetic risk assessment can tell us where to concentrate surveillance decades before cancer appears.

Genetics should help aim the weapons, not merely predict the casualty.

Prevent Cancer Before It Exists

Medical prevention adds still another layer. HPV vaccination provides an extraordinary example: Swedish registry data found 88% lower invasive cervical-cancer risk among women vaccinated before age 17 compared with unvaccinated women.

Cervical screening can then identify precancer. Colonoscopy can find and remove adenomas. H. pylori eradication reduces gastric-cancer risk. Hepatitis vaccination and treatment attack pathways leading to liver cancer.

These aren't merely ways to find cancer earlier. They break carcinogenesis before invasive cancer exists.

Find It Before It Finds You

The next defense is surveillance. Whole-body MRI is interesting for an aggressive strategy because it can survey large portions of the body without ionizing radiation. A 2026 meta-analysis covering 9,024 asymptomatic people found previously unknown cancers in 1.57%. We don’t have data on subsequent reduced mortality, but it demonstrates that occult cancers can be found before symptoms bring apparently healthy people to medicine. [42]

Although experiencing a slow start in terms of effectiveness, blood tests for emerging circulating-tumor-DNA and multi-cancer detection may eventually make surveillance considerably more powerful as well.

For someone willing to accept more testing, expense, incidental findings, and false alarms in exchange for a chance at earlier detection, going beyond population-minimum screening can be a rational and deliberate part of a maximal-agency strategy. The evidence should determine which tests survive in your repertoire—but the default doesn’t need to be “stop exactly where population guidelines stop.”

Build the Future Cancer Patient Now

Suppose prevention and surveillance both fail. Agency doesn't end.

The 2025 CHALLENGE randomized trial studied 889 patients with resected colon cancer after chemotherapy. Structured exercise produced 80.3% five-year disease-free survival versus 73.9% in controls. Eight-year overall survival was 90.3% versus 83.2%, an absolute difference of 7.1 percentage points. More importantly, Over the follow-up period, the exercise group had a 37% lower hazard of death than the control group. [31] These people had already developed cancer, undergone surgery, and received chemotherapy. Behavior still changed the outcome.

That trial is not the only punch. A 2025 meta-analysis of 42 studies involving 46,694 people with cancer found that high muscle strength or cardiorespiratory fitness was associated with 31–46% lower all-cause mortality than low fitness; each increment in cardiorespiratory fitness was also associated with an 18% lower risk of cancer-specific death. [41] These are observational associations rather than randomized proof, but they reinforce the same idea: the body you have when you finally get cancer does indeed make a difference.

Exercise therefore appears repeatedly: lower cancer incidence, better metabolic health, greater treatment reserve, and—now with randomized evidence in colon cancer—better disease-free and overall survival. High fitness also tracks with substantially better survival across many cancers. That makes fitness perhaps the closest thing we currently have to a multi-stage cancer intervention.

The Cancers That Terrify Us

Then there are cancers that seem to mock the entire strategy. Glioblastoma, pancreatic cancer, ovarian cancer, some aggressive hematologic malignancies…These are the diseases that make “just live healthy” sound painfully inadequate.

When primary prevention becomes weak, control has to move downstream.

That means identifying inherited risk where possible, recognizing symptoms quickly, finding disease earlier where technology permits, reaching expert treatment rapidly, and entering treatment with maximal physiological reserve.

For brain tumors specifically, MRI offers a non-ionizing surveillance technology. Routine brain-MRI screening has not been shown to reduce glioblastoma mortality, but MRI can detect asymptomatic tumors. It is a radical step I am prepared to consider in light of the positive tradeoff.

Frankly, glioblastoma remains the hardest challenge to my overall thesis. My mother's experience guarantees that I cannot honestly claim cancer is entirely controllable, totally within one’s sphere of responsibility. But weak primary prevention is not the same thing as zero agency. It means more of the available agency lies later in the chain.

The Outer Ring

Cancer epidemiology contains another intriguing category: characteristics repeatedly associated with cancer incidence or survival without enough evidence to assign them independent causal effects.

This list can become enormous, so only associations with meaningful numerical signals are worth mentioning.

These associations are not equivalent. Some have plausible biological pathways and repeated dose-response evidence; others may largely reflect smoking, wealth, healthcare access, relationship stability, or other confounders. Their value here is not that they can be added together, but that they identify potentially favorable states worth investigating.

The practical rule is simple: when the favorable side of a correlation is independently healthy, inexpensive, and low-risk, there may be little reason to wait twenty years for perfect causal decomposition before occupying that side of the distribution.

But as before, none of these numbers gets independently added to our final calculation. That would double-count the same healthy people repeatedly.

Now Pile Up the Numbers

Here is where the argument changes. We cannot add 40% preventable cancer + 44% lower Amish incidence + 67% lower Harvard mortality. Those measurements overlap. But neither should we pretend only one matters. Cancer has to traverse a chain:

initiation → promotion → immune escape → progression → failure of interception → failure of detection → failure of treatment → death.

Break the chain anywhere and cancer loses.

The table below gathers the strongest numerical anchors from the preceding sections before we move from measured evidence to an explicitly illustrative model.

None proves 90% overall control. Collectively, however, they make “40% controllable, 60% fate” look like a very poor description of the problem.

An 86% Model

Now comes the speculative part—and perhaps the most useful part. Instead of adding overlapping epidemiological percentages, model cancer as a series of defenses acting on whatever lethal risk survives the previous defense. These are illustrative assumptions, not clinical-trial estimates of a real protocol.

The equation: 1 − (0.50 × 0.85 × 0.75 × 0.45) = 85.7%

There is the 86%. It is not a measured fact. But it isn't numerology either.

A 50% prevention assumption lies between the ACS's roughly 40–44% attributable fractions and the Harvard study's much larger 59–67% mortality difference. Precancer interception already works spectacularly for particular cancers. Earlier detection can convert some advanced cancers into localized ones. Modern oncology already cures many cancer types. The uncertainty is in the magnitude and independence of those layers.

So let’s change the assumptions a little more, building one more step on this same logic…

Several imperfect defenses can compound into enormous control without requiring any one intervention to perform a miracle. That is the mathematical rebuttal to fatalism.

The Last Intervention Is Time

There is one more layer that cannot honestly be assigned a percentage.

Survival buys future medicine.

Prevent a cancer at 55 and develop it at 75, and you haven't merely gained twenty years. The cancer will be treated with the oncology of 2046 instead of 2026.

The strongest version of this idea is “longevity escape velocity”: medical progress eventually increases remaining healthy lifespan faster than aging consumes it. Nobody knows how soon that will happen, but the weaker version requires much less faith.

Cancer treatment in 2026 is dramatically different from cancer treatment several decades ago. Immunotherapy has already transformed treatment of multiple cancers. Personalized vaccines, engineered immune cells, bispecific antibodies, molecularly targeted drugs and increasingly individualized combinations are under development.

AI could further accelerate imaging, pathology, molecular diagnosis, drug discovery, protein design, clinical-trial matching and individualized treatment selection. Superintelligence curing cancer is speculation. AI accelerating biomedical discovery is not.

My own bet is that oncology twenty years from now will be considerably more immune-centered and personalized than oncology today. The patient's immune system increasingly becomes something medicine engineers rather than merely hopes will cooperate.

This creates a peculiar compounding return on prevention. Every year cancer is delayed is another year in which cancer medicine gets a chance to improve.

Someone trying to reach 100 therefore has two jobs: survive with today's tools and hang on long enough to receive tomorrow's.

The Repertoire

The resulting strategy is surprisingly concrete.

Crush the established risks: no smoking, little or no alcohol, low visceral fat, excellent metabolic health, high physical activity, UV protection, radon mitigation, cancer-preventing vaccination, treatment of carcinogenic infections, abundant fiber and whole grains, and an overall dietary pattern associated with lower cancer risk.

Know the genetic hand: obtain high-quality genetic risk assessment, identify actionable inherited cancer syndromes, and concentrate prevention and surveillance accordingly.

Prevent cancer medically: vaccinate, eradicate carcinogenic infections where possible, remove precancers and use colonoscopy and other proven screening aggressively.

Look for cancer before it looks for you: use established screening and consider more aggressive MRI-based or molecular surveillance when the personal tradeoff makes sense. Track emerging multi-cancer detection technologies closely.

Build the terrain: maintain exceptional fitness, muscle, insulin sensitivity, healthy body composition, good sleep and the nutritional substrate for healthy immune and gut function.

Attack aging itself: work on the underlying biology (like the epigenetic clock) rather than merely chasing biomarker scores, while recognizing that the cancer payoff remains unquantified.

Prepare for failure: if cancer still arrives, enter treatment extraordinarily fit, move quickly to expert oncology, do your own research while getting multiple expert opinions, and use the best available conventional and evidence-supported experimental medicine. If the prognosis is grave, go beyond your mental fences of what’s “conventional.” Be willing to investigate the legitimate bleeding edge—clinical trials, off-label prescription drugs, specialist centers, and emerging therapies wherever they are available. Go to the ends of the earth to obtain it, if necessary.

Above all, do what you need to do to buy time.

What We Don't Know Cuts Both Ways

Scientific uncertainty cuts in both directions. An unproven intervention cannot honestly be quantified as a benefit—but lack of proof also does not show that its true effect is zero.

For instance, we don't know the cancer benefit of truly slowing biological aging or rolling back an epigenetic clock, as mentioned above. We also don't know the maximal benefit of decades of unusually high fitness and metabolic health. We don't know how much deliberately optimizing immune function can prevent cancer. We don't know how much microbiome manipulation will matter. We don't know what emerging molecular early-detection tools will accomplish. We don't know what oncology will look like after another twenty years of biotechnology and AI.

Those unknowns cannot be entered into the 86% calculation as established benefits. But neither is there any scientific rule requiring all of them to equal zero. Some will disappoint. Some will matter.

What we don't know is therefore not automatically reason for pessimism. Several enormous biological domains remain potential sources of additional control that today's 40% statistic simply cannot capture.

That is reason for investigation, action and hope.

So, Is Cancer YOUR Fault?

Not in the simplistic moral sense. A healthy person can do everything right and develop glioblastoma. A lifelong smoker can die at 95 without lung cancer. Population probabilities cannot tell us why one particular person got one particular tumor.

But that is very different from saying our choices don't matter very much.

We already have strong evidence that approximately 40% of cancers and 44% of cancer deaths are attributable to known modifiable factors. Low-risk lifestyle comparisons reach 59–67% lower carcinoma mortality against the general population. Distinct populations demonstrate dramatically lower cancer incidence. Vaccines and precancer removal can interrupt carcinogenesis altogether. Imaging and screening tests can find occult disease. Fitness can improve outcomes even after cancer occurs. Beyond that lie immune optimization, biological aging, the microbiome, genomics, increasingly sensitive detection, and future oncology—large domains whose eventual contributions are not yet fully quantified.

I cannot scientifically say that 86% of cancer deaths are preventable today. I can say something more interesting: There is a plausible pathway (demonstrated in table above) to accumulating very high control over cancer mortality because cancer can be attacked repeatedly rather than once.

Our illustrative model reaches 86% without requiring a miracle at any stage. Even deliberately weaker assumptions produce 75%. A future scenario reaches 93%. Maybe the eventual number will be lower. Maybe it will be higher, but the upper end of that range changes the motivational equation.

If 40% is controllable and everything else is fate, enormous sacrifices can seem pointless. If prevention, interception, surveillance, fitness, treatment and advancing technology can compound into something approaching 80–90% control over cancer death, the juice begins to look very much worth the squeeze.

My mother's glioblastoma is why I cannot write that cancer is completely controllable. Her diagnosis is also why I refuse to treat today's unknown and uncontrollable territory as permanently beyond our reach. My response is to keep making that territory smaller: prevent what can be prevented, intercept what can be intercepted, find what can be found, survive what can be treated, and remain alive long enough for medicine to convert more of today's bad luck into tomorrow's controllable risk.

The first goal is never to get cancer.

The more important goal is never to die from it.

And there is considerably more reason for hope in those numbers than “40% preventable” initially suggests.

Sources

1.​   Islami F, et al. Proportion and number of cancer cases and deaths attributable to potentially modifiable risk factors in the United States, 2019. CA: A Cancer Journal for Clinicians. 2024;74(5):405-432. Figures also in the American Cancer Society release of July 11, 2024. https://acsjournals.onlinelibrary.wiley.com/doi/full/10.3322/caac.21858 

2.​   Islami F, et al. Same analysis on 2014 data. CA: A Cancer Journal for Clinicians. 2018;68:31-54. https://pubmed.ncbi.nlm.nih.gov/29160902/ 

3.​   Fink H, et al. Global and regional cancer burden attributable to modifiable risk factors to inform prevention. Nature Medicine. 2026. https://www.nature.com/articles/s41591-026-04219-7 

4.​   Song M, Giovannucci E. Preventable incidence and mortality of carcinoma associated with lifestyle factors among white adults in the United States. JAMA Oncology. 2016;2(9):1154-1161. https://pmc.ncbi.nlm.nih.gov/articles/PMC5016199 

5.​   Tomasetti C, Vogelstein B. Variation in cancer risk among tissues can be explained by the number of stem cell divisions. Science. 2015;347:78-81. 

6.​   Tomasetti C, Li L, Vogelstein B. Stem cell divisions, somatic mutations, cancer etiology, and cancer prevention. Science. 2017;355:1330-1334. Vogelstein's three-mutation explanation is reported in The ASCO Post, May 25, 2017. https://www.science.org/doi/10.1126/science.aaf9011

7.​   Wu S, Powers S, Zhu W, Hannun YA. Substantial contribution of extrinsic risk factors to cancer development. Nature. 2016;529:43-47. https://www.nature.com/articles/nature16166 

8.​   Tindle HA, et al. Lifetime smoking history and risk of lung cancer: results from the Framingham Heart Study. Journal of the National Cancer Institute. 2018. doi:10.1093/jnci/djy041 

9.​   Aminian A, et al. Association of bariatric surgery with cancer risk and mortality in adults with obesity (the SPLENDID study). JAMA. 2022;327:2423-2433. 

10.​ Office of the United States Surgeon General. Alcohol and Cancer Risk: The Surgeon General's Advisory. January 2025. 

11.​ Green AC, et al. Reduced melanoma after regular sunscreen use: randomized trial follow-up. Journal of Clinical Oncology. 2011;29(3):257-263. 

12.​ Moore SC, et al. Association of leisure-time physical activity with risk of 26 types of cancer in 1.44 million adults. JAMA Internal Medicine. 2016. doi:10.1001/jamainternmed.2016.1548 

13.​ International Agency for Research on Cancer. Press release 240: IARC Monographs evaluate consumption of red meat and processed meat. October 26, 2015. https://www.iarc.who.int/en/media-centre/pr/2015/pdfs/pr240_E.pdf 

14.​ Tantamango-Bartley Y, et al. Vegetarian diets and the incidence of cancer in a low-risk population. Cancer Epidemiology, Biomarkers and Prevention. 2013;22(2):286-294. https://pubmed.ncbi.nlm.nih.gov/23169929/ 

15.​ Lei J, et al. HPV vaccination and the risk of invasive cervical cancer. New England Journal of Medicine. 2020;383:1340-1348. 

16.​ Palmer TJ, et al. Invasive cervical cancer incidence following bivalent human papillomavirus vaccination (Scotland). Journal of the National Cancer Institute. 2024. Summarized by Gavi. https://www.gavi.org/vaccineswork/hpv-vaccine-erasing-cancer-heres-proof 

17.​ Ford AC, et al. Eradication therapy to prevent gastric cancer in Helicobacter pylori-positive individuals: systematic review and meta-analysis. Gastroenterology. 2025. https://pubmed.ncbi.nlm.nih.gov/39824392/ 

18.​ Bretthauer M, et al. NordICC trial. Ten-year results: New England Journal of Medicine. 2022;387:1547-1556. Thirteen-year results: The Lancet, May 2026, presented at Digestive Disease Week. https://www.healio.com/news/gastroenterology/20260505/nordicc-trial-update-colonoscopy-reduced-crc-i ncidence-30-no-effect-on-mortality 

19.​ United States Environmental Protection Agency. Health Risk of Radon. 

20.​ Hill W, et al. Lung adenocarcinoma promotion by air pollutants. Nature. 2023;616:159-167. 

21.​ Klein EA, et al. Vitamin E and the risk of prostate cancer: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA. 2011;306(14):1549-1556. 

22.​ Omenn GS, et al. Risk factors for lung cancer and for intervention effects in CARET, the Beta-Carotene and Retinol Efficacy Trial. Journal of the National Cancer Institute. 1996;88(21):1550-1559. https://pubmed.ncbi.nlm.nih.gov/8901853/ 

23.​ Manson JE, et al. Vitamin D supplements and prevention of cancer and cardiovascular disease (VITAL). New England Journal of Medicine. 2019;380(1):33-44. 

24.​ Westman JA, et al. Low cancer incidence rates in Ohio Amish. Cancer Causes and Control. 2010;21(1):69-75. https://pmc.ncbi.nlm.nih.gov/articles/PMC4308689 

25.​ National Highway Traffic Safety Administration. Lap and shoulder belts reduce the risk of fatal injury to front-seat passenger car occupants by 45%. Traffic Safety Facts; summarized by the Insurance Institute for Highway Safety. https://iihs.org/topics/seat-belts 

27. National Cancer Institute. Age and Cancer Risk. NCI identifies advancing age as the most important overall cancer risk factor and gives the incidence figures used above.

28. National Cancer Institute. What Cancer Screening Tests Check for Cancer? NCI notes that colorectal screening can detect and remove precancerous polyps and that cervical screening can detect abnormalities before invasive cancer develops. 

29. National Cancer Institute. Cervical Cancer Causes, Risk Factors, and Prevention. NCI states that virtually all cervical cancers are caused by persistent high-risk HPV and that vaccination, screening, and treatment of precancer make cervical cancer highly preventable.

30. National Cancer Institute. Cervical Cancer Screening (PDQ), Health Professional Version. NCI's evidence review estimates regular Pap screening reduces cervical cancer incidence and mortality by at least 80%. 

31. Courneya KS, et al. Structured Exercise after Adjuvant Chemotherapy for Colon Cancer. New England Journal of Medicine. 2025;393:13–25. The randomized CHALLENGE trial reported 80.3% vs. 73.9% five-year disease-free survival and 90.3% vs. 83.2% eight-year overall survival. 

32. American Cancer Society. Nutrition and Physical Activity Guideline for Cancer Survivors. ACS recommends physical-activity assessment and counseling beginning after diagnosis and summarizes evidence on treatment tolerance, quality of life, recurrence, and survival. 

33. Orlich MJ, et al. Longitudinal associations between vegetarian dietary habits and site-specific cancers in the Adventist Health Study-2 North American cohort. American Journal of Clinical Nutrition. 2025. PMID: 40499906. 

34. Seyfried TN, Flores RE, Poff AM, D'Agostino DP. Cancer as a mitochondrial metabolic disease. Frontiers in Cell and Developmental Biology. 2015;3:43. Review of nuclear-cytoplasmic transfer experiments; interpretation remains debated. PMID: 26217661.

35. National Cancer Institute. HIV Infection and Cancer Risk. NCI fact sheet; cancer-risk estimates for Kaposi sarcoma, lymphomas, cervical cancer and other HIV-associated malignancies.

36. Valenzuela PL, et al. Exercise Training and Natural Killer Cells in Cancer Survivors: Current Evidence and Research Gaps Based on a Systematic Review and Meta-analysis. Sports Medicine - Open. 2022;8:36. PMID: 35244811.

37. Pagès F, et al. International validation of the consensus Immunoscore for the classification of colon cancer. Lancet. 2018;391:2128-2139. PMID: 29754777.

38. Naghshi S, et al. Association of Total Nut, Tree Nut, Peanut, and Peanut Butter Consumption with Cancer Incidence and Mortality: A Comprehensive Systematic Review and Dose-Response Meta-Analysis. Advances in Nutrition. 2021;12:793-808. PMID: 33307550.

39. Spencer CN, et al. Dietary fiber and probiotics influence the gut microbiome and melanoma immunotherapy response. Science. 2021;374:1632-1640. PMID: 34941392.

40. Grzymski JJ, et al. Population genetic screening efficiently identifies carriers of autosomal dominant diseases. Nature Medicine. 2020;26:1235-1239. PMID: 32719484.

41. Bettariga F, et al. Association of muscle strength and cardiorespiratory fitness with all-cause and cancer-specific mortality in patients diagnosed with cancer: a systematic review with meta-analysis. British Journal of Sports Medicine. 2025. PMID: 39837589.

42. Martins da Fonseca J, et al. Whole-body MRI for opportunistic cancer detection in asymptomatic individuals: a systematic review and meta-analysis. European Radiology. 2026;36:1813-1823. PMID: 40884613.

43. Wang F, et al. A systematic review and meta-analysis of 90 cohort studies of social isolation, loneliness and mortality. Nature Human Behaviour. 2023;7:1307-1319. PMID: 37337095.

44. Dugué PA, et al. DNA methylation-based biological aging and cancer risk and survival: pooled analysis of seven prospective studies. International Journal of Cancer. 2018;142:1611-1619. PMID: 29197076.

45. Garcia L, et al. Non-occupational physical activity and risk of cardiovascular disease, cancer and mortality outcomes: a dose-response meta-analysis of large prospective studies. British Journal of Sports Medicine. 2023;57:979-989. PMID: 36854652.

46. Del Fabbro E, et al. Marital status and survival in cancer patients: a systematic review and meta-analysis. Cancer Medicine. 2023. PMCID: PMC9883406. 47. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and breastfeeding: collaborative reanalysis of individual data from epidemiological studies in 30 countries. Lancet. 2002;360:187-195. PMID: 12133652.