Researchers use AI to predict cavities

An international team of researchers is leveraging artificial intelligence to predict who is at high risk for cavities. The findings, published in the journal Caries Research, could help identify population clusters most at risk for tooth decay based on larger social determinants of health rather than on clinical diagnostics at the individual level.

“The emergence of AI offers a transformative opportunity for public health surveillance and risk stratification, including addressing the global burden of cavities,” said Fernando Neves Hugo, chair of the Department of Epidemiology and Health Promotion at NYU College of Dentistry and the study’s senior author. “These are powerful tools that could be helpful for targeting resources and delivering preventive care to specific groups that will benefit the most.”

Dental cavities are the most common health condition globally, affecting billions of people across all age groups. Cavities are driven by a range of factors, many of which occur at the population level rather than the individual, including access to dental care, diet and sugar consumption, and water sources.

AI algorithms allow researchers to analyze large, complex datasets to find patterns that otherwise would be difficult to pinpoint. Hugo and his colleagues applied a series of AI models to existing data from the Brazilian Oral Health Survey, a national epidemiological study coordinated by the Brazilian Ministry of Health. They focused on data from 23,726 people who took part in the 2023 survey, including children, adolescents, and adults. 

The survey captures demographic information (e.g., age, sex, race) as well as social determinants of health that research shows are linked to oral health, including socioeconomic factors (e.g., wealth, household density, private dental insurance status, water source) and behavior (e.g., time since last dental visit). Participants also undergo clinical exams to measure the prevalence and severity of tooth decay. 

Through traditional analyses, the researchers found that the population’s oral health reflects a complex interplay of socioeconomic, structural, and behavioral factors. Wealth and household density were the strongest predictors of cavity risk. Non-White populations had a higher risk for cavities across all age groups, while private dental insurance and recent dental visits were protective against cavities. In addition, water source also played a critical role: in one analysis, 62.5 percent of 12-year-olds without access to fluoridated water had cavities, nearly double the 35.5 percent observed in those with fluoridated water systems. 

The researchers then applied several machine learning algorithms to the dataset and found that the AI models could accurately predict cavity risk and severity by age group by evaluating the relative contributions of the different factors.

“Integrating machine learning into this framework goes beyond merely predicting cavity risk; it addresses the complexity and multidimensional interactions between social determinants that traditional measures often mask,” explained Hugo. “Our machine learning approach acts as a high-resolution sensor that identifies individuals whose profiles signal a heightened risk that would otherwise be invisible within community averages.”

The researchers note that the findings may translate beyond Brazil to other middle-income countries, but that local factors, including access to dental care and benefits, would likely shift the variables that explain cavity risk in other countries. 

More broadly, their goal is to replicate this AI-enabled approach with factors specific to other countries in order to develop tailored prevention strategies. Hugo and his colleagues are preparing to run a similar analysis using US data from the National Health and Nutrition Examination Survey.

“When we think about public health and preventive interventions, we need to be accountable to society and make the best use of the resources we have. A more targeted approach helps us optimize resources—essentially, getting more bang for our buck,” said Hugo.

Additional study authors include Jessye Melgarejo do Amaral Giordani, Orlando Luiz do Amaral Júnior, and Maria Laura Braccini Fagundes of the Federal University of Santa Maria in Brazil, as well as Juliana Balbinot Hilgert of the Federal University of Rio Grande do Sul in Brazil.