147 - Evaluating Radiographic Tool Concordance for Crown Sizing and Outcomes

Purpose: To validate stainless-steel crown (SSC) prediction models by Helder et al. (2022)—previously reported achieving 93% accuracy with coronal and 88% cervical measurements—and explore demographic influences on sizing and correlations between matched/non-matched crowns and clinical outcomes.

Methods: Electronic health records (n=111) at Columbia University’s teaching clinic were reviewed for restorative appointments (charted with code D2930) from 2020-23. Data abstracted included radiographs, demographics, postgraduate level of restorative provider, and clinical outcomes (e.g., fit, seating issues, resorption). Measurements followed Helder et al.'s protocol using MiPACS software and were compared to model predictions. Crown matching was evaluated dichotomously (matched vs. non-matched). Chi-square tests assessed associations between patient demographic and provider characteristics on prediction accuracy. Fisher’s exact tests analyzed associations between matching and clinical outcomes (e.g., size, seating, abscesses, etc.).

Results: Children averaged 5.9 years of age (range: 3–11). Over half were male (59:41%, n=66), with 70% (n=78) identifying as Hispanic. Predicted crown sizes matched clinical placements in 92.5% of coronal and 87% of cervical measurements. No significant associations were found between matching and ethnicity, sex, or provider proficiency. Non-matched crowns were more likely oversized, improperly seated, or documented with abscesses. ICC values were 0.98 (coronal) and 0.96 (cervical).

Conclusion: The prediction tool reliably matched crowns with no significant demographic influences, though sample size may have limited power to detect associations. Findings suggest non-matched crowns may be more prone to complications, warranting further evaluation in a larger sample.

Identify Supporting Agency and Grant Number: