Dental caries remains one of the most prevalent chronic conditions globally, yet efforts to quantify a patient′s risk of developing new caries over time remain limited. Existing caries risk assessment (CRA) tools are widely used in clinical settings but typically rely on cross – sectional data, which may limit their predictive accuracy and generalizability across populations. To address these limitations, we developed RT2C, a recurrent neural network model based on bidirectional gated recurrent units (Bi–GRU) to predict time to new caries. The model was trained on longitudinal dental records from 466,782 patients across four diverse dental organizations, encompassing over one million visits between 2019 and 2023. Input features included demographics, preventive procedures, caries history, and previously recorded CRA scores. We benchmarked the model performance against other machine learning models such as logistic regression, light gradient boost machine, random survival forest, and CRA–based baselines using AUROC for binary classification and concordance index (c–index) for survival prediction. Our model demonstrated strong predictive performance, achieving a c–index of 88.52% (95% CI: 88.51–88.54) in survival analysis, significantly outperforming baseline models based on current or past CRA scores by more than 18%. Site–level evaluations confirmed its robust generalizability, with pooled models performing comparably or better than sites–pecific ones. Additionally, decision curve analysis showed that the Bi–GRU model offers greater clinical net benefit across a range of decision thresholds when compared to CRA or treat–all/treat–none strategies. By leveraging longitudinal dental histories and temporal patterns in patient visits, our proposed model provides clinically meaningful improvements in predicting caries risk. These findings support its potential integration into decision support tools for more personalized and preventive dental care.

The authors have declared no competing interest.

This study received funding in part from grant R01DE024166: Implementing Dental Quality Measures in Practice from the US Department of Health and Human Services, National Institutes of Health, and National Institute of Dental and Craniofacial Research and National Library of Medicine grant R01LM014249

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

IRB HSC-DB-21-0471 of University of Texas Health Science Center in Houston McWilliams School of Biomedical Informatics gave ethical approval for this work

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

All data produced in the present study are available upon reasonable request to the authors