- Researchers linked insulin resistance to 12 types of cancer, with uterine cancer showing the strongest connection at 134% increased risk.
- A new AI tool outperformed BMI alone in predicting cancer risk, catching metabolic dysfunction even in people at healthy weights.
- The tool flagged elevated cancer risk in normal-weight patients that standard BMI screening missed.
Insulin sensitivity is a known driver of diabetes progression and other health complications.
Now, insulin resistance has been linked to a 25% higher risk of 12 different types of cancer, according to new findings published February 16 in
Researchers from the University of Tokyo and Taichung Veterans General Hospital in Taiwan developed an artificial intelligence tool to predict insulin resistance.
The AI tool identified patients with insulin resistance and flagged higher risks of developing diabetes, heart disease, and cancer.
The association between insulin resistance and cancer incidence was strongest for uterine cancer, with a 134% increased risk.
“People already believe that insulin resistance is associated with cancer; however, insulin resistance is difficult to evaluate in the clinic because measures to evaluate…are not at all practical in a clinic,” said study author Yuta Hiraike, MD, PhD, of the University of Tokyo Hospital.
Hiraike told Healthline the team approached the research knowing the association with insulin resistance and cancer but also recognizing the practical challenges of diagnosis.
He said the model, named AI-IR, provides a “convincing answer…to demonstrate that insulin resistance is actually a risk factor for cancer.”
Neil Iyengar, MD, an oncologist and director of survivorship at Emory Winship Cancer Institute, wasn’t involved in the research, but said it confirms a large body of evidence linking metabolic health to cancer risk.
Iyengar told Healthline that the findings could help researchers “develop better tools that are more specific and more individualized for risk prediction.”
The AI tool combined nine parameters and uses artificial intelligence (AI) to detect insulin resistance (IR).
In addition to age, sex, race and body mass index (BMI), the model uses five blood tests:
Hiriake said they chose these parameters because primary care physicians routinely capture them. “I believe implementing [the tool] to the real world is not difficult,” he said.
While body mass index (BMI) remains the strongest single predictor of insulin resistance, what sets AI-IR apart is its ability to flag risk in people who would be cleared by standard BMI screening, catching metabolic dysfunction in people with a healthy weight.
“What’s unique about this study is they were able to predict the insulin resistance before the insulin resistance even happened,” Iyengar said.
That’s important because, while obesity has long been recognized as a cancer risk factor, an emerging area of research suggests that metabolic health tells a more complete story than weight alone.
Researchers built the AI model using data sets from U.S. and Taiwanese populations, then tested it with nearly 400,000 participants in the United Kingdom. Because the UK population was predominantly ethnically European, the study’s findings are limited by ethnicity.
People identified as having insulin resistance had a 25% higher cancer risk, the study found. Six cancers showed the strongest correlation with insulin resistance
- uterine
- kidney
- esophagus
- pancreas
- colon
- breast
Six other cancer types were associated with insulin resistance, but not as strongly:
- renal pelvis
- small intestine
- stomach
- liver and gallbladder
- leukemia
- bronchial and lung
The study identified an 134% increased risk of developing uterine cancer for people with insulin resistance.
But the finding was “one of the least surprising,” said Stephen Gruber, MD, PhD, epidemiologist and medical oncologist at City of Hope, a national cancer research and treatment organization. Gruber wasn’t involved in the study.
“Cancer of the lining of the uterus, or what’s called endometrial cancer, is one of the cancers that has been known for the longest to be associated most strongly with excess weight and obesity,” Gruber told Healthline, adding that the results confirmed the AI model was working as intended.
What made the finding significant was that AI-IR gave researchers additional predictive power beyond BMI.
Even after researchers accounted for body weight, AI-IR still flagged elevated uterine cancer risk in some patients, suggesting that metabolic dysfunction itself threatens uterine health independent of body size.
Insulin resistance appears to directly fuel cancer cell growth in the uterine lining through hormonal signaling. Unlike some cancers in the study, experts note that this is a risk of weight loss, and that improved metabolic health can meaningfully reduce it.
The model also successfully predicted heart disease and diabetes risk, both established conditions associated with insulin resistance.
That increased risk extended to a population standard screening consistently misses, and that’s precisely who this tool is uniquely designed to catch.
For lung and bronchial cancer, AI-IR identified elevated risk completely independent of BMI.
Iyengar said that identifying cancer risk in people who do not have classic obesity, which is a BMI over 30, is an important area of research because standard screenings “don’t really recognize their cancer risk.”
“They walk into their doctor’s office and they’re told, ‘oh, you have a normal weight. You’re safe, you’re good to go,’” he said.
“You can have a normal body mass index or a normal weight, and still be at elevated risk for cardiometabolic diseases. And that’s especially those patients who have either high insulin levels or insulin resistance, or patients who have high body fat levels, which is actually called normal weight obesity,” Iyengar said.
Iyengar’s own research found that “women who were normal weight, but who had an elevated total body fat level —above 33% — had more than a doubling in their risk of developing breast cancer.”
“We’re learning classical obesity or BMI is not specific enough to predict cancer risk in everyone so looking at insulin levels and body fat levels, I would add, are a more personalized way of predicting risk,” he said.
This study represents a step toward more personalized cancer risk prediction, a future that looks beyond weight to the metabolic factors driving disease.
“It gives us an opportunity to understand ways in which we can intervene early, to prevent outcomes like diabetes or cancer,” Gruber said
The AI-IR tool is not yet available for clinical practice, but the markers it relies on, including hemoglobin A1C and body fat percentage, are available through routine care today.
Here’s how experts say you can start using them to assess your own risk:
Measure body composition
Iyengar recommended asking your doctor about having your body fat percentage measured.
The most accurate option is a DEXA scan — “it’s the same way that people measure their bone density to predict osteoporosis,” he said, just run under different software.
For most people, a bioimpedance (BIA) scale is the next best option, and many people already own one at home.
“When you go to the gym, and you see those scales with those electrodes, and you hold the things with your thumbs… that’s a BIA machine,” Iyengar said.
He suggested targeting a total body fat level below 30%, but emphasized that the trend over several weeks matters more than any single reading.
“Those BIA machines have a margin of error… don’t be too tied up to the absolute number. Just watch the trend.”
Rethinking bloodwork
Monitoring your A1C levels is key to preventing and also managing diabetes.
“Everybody should be doing standard diabetes screening — checking a hemoglobin A1C — even if it’s in the prediabetes range,” Iyengar said.
“Right now we don’t think of hemoglobin A1C as a risk stratification tool for cancer — we think of it as predicting diabetes.” A level above 5.5–5.7 may signal elevated cancer risk worth discussing with your doctor.
Move more, eat better
The lifestyle habits that protect your heart and prevent diabetes “are really going to help prevent a lot of cancers too,” Iyengar said.
