Predictive Analysis of Dental Caries Risk via Rapid Urease Activity Evaluation in Saliva Using a ZIF-8 Nanoporous Membrane

Abstract

Despite a decrease in the incidence of dental caries over the past four decades, it remains a widespread public health concern. The multifactorial etiology of dental caries complicates effective prevention and early intervention efforts, underscoring the need for the development of rapid predictive methods that account for multiple factors. In this study, we selected the activity of urease secreted by Streptococcus salivarius as a metabolic marker for dental caries. This activity was quantified by measuring the diffusion of hydroxide ions generated from the urease catalytic reaction on urea across a ZIF-8-modified nanoporous membrane. The choice of ZIF-8 was based on its preference in transporting hydroxide ions, enabling the accurate detection of urease activity at concentrations as low as 1 CFU/mL. Subsequently, we collected 287 saliva samples to determine the Michaelis constant (K_m) of urease using this method. Logistic regression analysis revealed that both the K_m of urease and the frequency of sugar intake are significant factors influencing the development of dental caries. Furthermore, we developed a machine learning methodology for identifying dental caries, achieving an accuracy rate of 81%. It is expected that increasing the sample size will further enhance the predictive accuracy of the model. This innovative approach provides valuable insights into early intervention strategies in the fight against dental caries.

Keywords: ZIF-8-modified nanoporous; dental caries; ion transport; logistic regression; machine learning; urease activity.