While skepticism towards cancer prevention is warranted, a landmark study published in Nature Medicine just recently (February 2026) delivered a remarkable headline worthy of further study: nearly 40% of the world's 18.7 million annual cancer cases, roughly 7.1 million diagnoses, are linked to modifiable risk factors.
Tobacco smoking led the list, followed by infections and alcohol consumption.
The research, led by Hanna Fink at the WHO's International Agency for Research on Cancer, examined 36 cancer types across 185 countries, making it one of the most comprehensive analyses of preventable cancer ever conducted.
The number should alarm us and all data regarding such a contentious subject as cancer should be treated with care. But here's another way to think about it: it's likely an undercount. Here is my reasoning why I think that and offer that as at least a thought exercise.
The study evaluated 30 "well-established" risk factors. A methodologically sound choice that also reveals its central limitation. To be included, a risk factor needed robust causal evidence accumulated over decades.
That's a high bar, and it systematically excludes several classes of exposure where the science has matured rapidly in the last two to three years but hasn't yet crossed the threshold of canonical consensus.
When you examine what was left out, the real fraction of preventable cancer starts to look considerably larger than 38%.
Here are four domains of recent research that even many leading oncologists and epidemiologists haven't fully integrated into their thinking. That collectively suggests the true preventable share may be closer to half of all cases, or beyond.
Ultra-processed foods now constitute roughly 60% of caloric intake in the United States and over half in much of Europe. The evidence linking them to cancer has been building with remarkable consistency.
A 2025 IARC study of 266,666 men and women across seven European countries found that higher consumption of ultra-processed foods was associated with elevated risk of cancer and cardiometabolic multimorbidity. Not just one disease, but clusters of them. A meta-analysis pooling multiple cohort studies found statistically significant associations between high ultra-processed food intake and colorectal, breast, and pancreatic cancers, with hazard ratios reaching 1.30 for colorectal and 1.49 for pancreatic cancer when comparing the highest to lowest consumption groups.
Perhaps most striking is a November 2025 study from Mass General Brigham, published in JAMA Oncology, which followed nearly 30,000 women and found that those consuming the highest levels of ultra-processed foods had a 45% higher risk of developing conventional adenomas — the precancerous polyps most associated with early-onset colorectal cancer — compared to the lowest consumers. The dose-response relationship was essentially linear: more ultra-processed food, more precancerous growths.
This matters because early-onset colorectal cancer has been surging in high-income countries for reasons that have puzzled researchers. The ultra-processed food connection is now one of the strongest candidates, and it operates through mechanisms: chronic inflammation, gut microbiome disruption, endocrine interference from additives, that the Fink study's 30 risk factors don't capture directly.
Diet was partially addressed through factors like low fruit and vegetable intake, but the specific industrial processing of food: the emulsifiers, the modified starches, the artificial sweeteners, represents a distinct and additive category of risk.
PFOA, one of the most studied per- and polyfluoroalkyl substances, was upgraded by IARC from "possibly carcinogenic" to a confirmed human carcinogen in 2023. PFOS was classified as possibly carcinogenic. These chemicals have been detected in the blood of virtually every American adult.
A January 2025 study from USC's Keck School of Medicine, the first of its kind, compared PFAS contamination in drinking water against cancer incidence data across U.S. counties from 2016 to 2021. The findings were sobering. Counties with PFAS levels exceeding EPA guidelines showed cancer incidence increases ranging from 2% to 33% depending on the chemical and cancer type, spanning digestive, endocrine, respiratory, and oral cancers. The researchers estimated that PFAS contamination of drinking water alone contributes to approximately 6,864 cancer cases per year in the United States.
NCI investigators have found that individuals with the highest blood concentrations of PFOA were more than twice as likely to develop kidney cancer compared to those with the lowest levels. Elevated PFOS levels have been linked to testicular cancer risk among U.S. Air Force servicemen exposed through contaminated water at military installations.
These chemicals were not among the 30 risk factors in the Fink study. Yet PFAS have been found in roughly 45% of U.S. drinking water supplies, they bioaccumulate for years in human tissue, and their carcinogenic mechanisms: immune suppression, endocrine disruption, chronic inflammation, interference with DNA repair are now well-documented. The global exposure is essentially universal, which means even a modest per-person risk increase translates to an enormous population-level cancer burden that the 38% figure doesn't reflect.
We are the first generation to have microplastics in our blood, our lungs, our placentas, and our colons. Plastic production has skyrocketed since the mid-20th century, and the tiny particles that result from degradation are now present in virtually every ecosystem and organism tested.
A 2025 case-control study published in the Journal of Hazardous Materials provided some of the first direct epidemiological evidence: among 258 colorectal cancer patients and 493 controls in China, fecal microplastic concentrations were significantly higher in cancer cases. Individuals in the highest exposure quartile had an adjusted odds ratio of 11.3 for colorectal cancer compared to the lowest quartile, an extraordinarily strong association for an environmental exposure.
A pooled meta-analysis of 43 studies involving over one million patients, published in the International Journal of Surgery in 2025, found that microplastic pollutant exposure was a risk factor for overall cancer incidence, with phthalates and bisphenol A, chemicals leached from plastics, showing the most consistent associations.
The mechanistic picture is filling in rapidly. Researchers at MD Anderson Cancer Center are studying how microplastics trigger chronic inflammation, oxidative stress, and endocrine disruption in tissues. A 2024 systematic review found that microplastics have a "suspected" role in increasing colon and lung cancer risk. Polystyrene nanoplastics have been shown to trigger cancer-associated changes in healthy lung cells, including DNA damage and activation of cellular survival pathways.
As some researchers have noted, people under 50 were the only demographic group to see rising cancer rates between 1995 and 2021. The explosion of plastic in the environment over precisely this period is difficult to dismiss as coincidence. Yet microplastics appear nowhere in the Fink study's risk factor list, understandably, given how young the epidemiological evidence is, but consequentially, given how ubiquitous the exposure has become.
The IARC has classified night shift work as "probably carcinogenic to humans" (Group 2A) since 2019, based on evidence linking it to breast, prostate, and colorectal cancers. The biological mechanisms are well-characterized: light during the biological night suppresses melatonin, disrupts clock gene regulation, impairs DNA repair, suppresses natural killer cell activity, and deregulates circadian genes involved in cancer-related pathways.
A 2025 systematic review screening over 500 studies found strong evidence linking circadian disruption to breast, prostate, and colorectal cancers, with emerging evidence for melanoma and bladder cancer. The U.S. National Toxicology Program has concluded with "high confidence" that persistent night shift work resulting in circadian disruption can cause breast cancer.
Between 15% and 25% of all workers in Europe and the United States perform night or shift work. A December 2025 review in npj Breast Cancer detailed at least ten distinct biological mechanisms connecting night shift work to breast carcinogenesis, from melatonin suppression and oxidative stress to epigenetic modifications and telomere shortening.
But here's what researchers in this field increasingly emphasize: the problem extends far beyond formal night shift work. Modern society is saturated with circadian disruptors: screen exposure at night, irregular sleep schedules, jet lag, late-night eating. These affect vastly more people than formal shift workers, yet epidemiological tools are only beginning to capture circadian disruption outside occupational settings. The Fink study does not include circadian disruption among its 30 risk factors. The IARC's own classification, while strong, is focused narrowly on occupational night shift work, a fraction of total circadian disruption in the population.
None of this is a criticism of the Fink study. It did exactly what rigorous epidemiology should do: quantify risk from factors with established causal evidence. The researchers used exposure data from 2012, applied it to 2022 cases, and produced the most comprehensive global estimate to date.
But "well-established" is a trailing indicator. It takes decades for a risk factor to move from suggestive association to consensus classification. Tobacco took roughly 30 years from the first epidemiological evidence to widespread policy action. Each of the four domains above is somewhere on that arc, some further along than others, and each involves exposures that are global, involuntary, and growing.
If ultra-processed foods contribute even modestly to colorectal, breast, and pancreatic cancers; if PFAS add thousands of kidney, testicular, and thyroid cases annually; if microplastics are fueling the mysterious rise in early-onset cancers; and if circadian disruption's reach extends well beyond shift workers, then the true preventable fraction of global cancer is not 38%. It is substantially higher.
The implication is not that 38% is wrong. It's more likely that 38% is the floor.
And the distance between that floor and the ceiling is measured in millions of lives that current prevention frameworks aren't yet designed to save. The question for policymakers, clinicians, and the public is whether we wait another two decades for certainty, or begin acting on the weight of evidence we already have.
