Objective To compare the performance of the modified caries-risk assessment tool for preschool children (PSC-MCAT) and the Cariostat caries activity test in predicting caries risk among young children in Xixiangtang District, Nanning, Guangxi, and to verify the feasibility of PSC-MCAT in predicting caries risk in young children in Guangxi.

Methods A prospective cohort study enrolled 511 three-year-old children from kindergartens in Xixiangtang District, Guangxi. At baseline, their caries risk was assessed with both PSC-MCAT and Cariostat. Comprehensive oral examinations were performed every six months for 1.5 years to compare the effectiveness of the two methods in predicting incident caries.

Results No statistically significant difference was observed in the caries risk levels assessed by the two methods (P > 0.05), and their agreement was poor (P < 0.05). At the 1.5-year follow-up, statistically significant differences in the mean new caries count were found among all three risk levels of the PSC-MCAT (P < 0.05). For the Cariostat, significant differences were only observed between the low-risk and high-risk groups and between the moderate-risk and high-risk groups (P < 0.05); within each baseline caries-risk level, no statistically significant differences in 1.5-year caries incidence or increment were observed between the two CRA methods for low- and moderate-risk children (P > 0.05). In the high-risk group, statistically significant differences were observed in both caries incidence and increment (P < 0.05). Generalized linear model analysis revealed a positive association between the risk levels assessed by both tools and the incidence of new caries. In the low- and moderate-risk levels, PSC-MCAT did not differ significantly from Cariostat (P > 0.05). In the high-risk level, a statistically significant difference was observed between the PSC-MCAT and Cariostat (P < 0.05), and PSC-MCAT was associated with an increased risk of new caries development (P < 0.05). The area under the receiver operating characteristic (ROC) the curve (AUC) for the PSC-MCAT was 0.634 (95% CI: 0.593-0.675), with a sensitivity of 92.55% and a specificity of 32.62%. The Cariostat yielded an AUC of 0.664 (95% CI: 0.622-0.701), with a sensitivity of 24.45%, and a specificity of 94.76%.

Conclusion The PSC-MCAT demonstrates considerable potential for caries risk assessment and prediction of new caries development, showing particular advantage in the initial screening of high-risk individuals.