Quantitative and Structural Characterization of Streptococcus mutans Biofilms on Dental Restorative Materials Using Scanning Electron Microscopy and Machine Learning

Abstract

This study compared four biofilm fixation protocols for scanning electron microscopy (SEM) imaging and their compatibility with Fiji-based image analysis, and introduced a complementary approach for SEM biofilm cross-section analysis to be used when assessing the antimicrobial activity of dental restorative materials. In the first phase, resin composite discs (8 × 2 mm) were characterized for degree of conversion and surface roughness, then incubated with renilla-reporter Streptococcus mutans for 1 and 24 h. Metabolic activity was assessed by luciferase assay, and biofilms were fixed using four protocols based on buffered formalin and/or glutaraldehyde, with or without graded ethanol dehydration, prior to SEM imaging. In parallel, biofilm cross-sections were analyzed using resin bars incubated under different conditions. In the second phase, the most effective fixation protocol was validated using antimicrobial-containing restorative materials, including two resin composites and a glass ionomer cement. Data were analyzed using ANOVA and Tukey's test (p < 0.05). The results demonstrated that biofilm fixation protocols significantly influence SEM image quality, biofilm architectural features, and quantitative measurements obtained through image analysis. In conclusion, standardized biofilm fixation using glutaraldehyde combined with graded ethanol dehydration and coupled with machine-learning augmented analysis of SEM images is essential for producing superior image resolution to enable clearer discrimination among bacterial cells, biofilm matrix, and material surfaces, and improved sensitivity in detecting architectural differences, particularly in biofilm cross-sections.