Insulin resistance finds links to cancer

A machine learning model used to predict the insulin resistancea condition related to the onset of diabetes, has found that it is a risk factor for 12 types of cancer, as revealed by a research team led by the University of Tokyo (Japan).

The study, published in ‘Nature Communications‘, has applied a model based on machine learning, a branch of artificial intelligence (IA), for the prediction of insulin resistance through nine clinical parameters. The system, called artificial intelligence-derived insulin resistance (AI-IR), has been used on 500,000 UK Biobank participants.

“Although it has been suggested a possible relationship between insulin resistance and cancer, Large-scale evidence has been limited due to the difficulty of assessing insulin resistance in the clinic. But with AI-IR, we have provided the first evidence on a population scale that insulin resistance is a risk factor for cancer,” explained the researcher. Yuta Hiraikefrom Tokyo University Hospital.

Specifically, IA-IR found a significant association with an increased risk of cancer of the uterus, kidney, esophagus, pancreas, colon and breast, as well as nominal associations with cancer of the renal pelvis, small intestine, stomach, liver and gallbladder, leukemia, and bronchi and lung.

“And since the nine input parameters of AI-IR are obtained through standard medical checkups, AI-IR could be easily implemented to identify high-risk individuals and enable selective detection of diabetes, cardiovascular diseases and cancer,” Hiraike highlighted, underlining the potential of this tool in screening and diagnosis.

It is now common for body mass index (BMI), a measure of body fat, to predict an individual’s insulin resistance and their susceptibility to related cancers. However, this leads to false positives, where some obese people are considered metabolically healthy and do not suffer the harmful effects of obesity to the same extent as others, and false negatives, where people with an ideal BMI end up suffering from insulin resistance or related problems generally associated with obesity.

Part of the challenge Hiraike and his team faced was convincing the paper’s reviewers that AI-IR could overcome these shortcomings in a reliable and repeatable manner. The authors demonstrated not only the predictive power of the machine learning model, but also its robustness under various conditions.

“We are currently working to understand how genetic differences between individuals influence this risk and, ultimately, to link large-scale human data with molecular biology studies to develop better strategies to combat insulin resistance,” Hiraike said.