Abstract
Abstract
Aim
To identify clinical, CBCT-based radiographic, and patient-level predictors of 5-year tooth survival in periodontal-endodontic (perio-endo) combined lesions and develop an internally validated prognostic nomogram.
Methods
This retrospective cohort included 288 teeth from 288 patients with perio-endo combined lesions and ≥ 5 years of follow-up. Time to extraction was analysed using Kaplan–Meier estimation and multivariable Cox regression, with backward elimination from 14 candidate predictors. A points-based nomogram was developed from the final model. Discrimination was assessed using Harrell's C-index with 500 bootstrap resamples, and calibration compared predicted and observed 5-year survival across risk quintiles.
Results
Seventy-nine teeth (27.4%) were extracted, with a median time to extraction of 30 months. Overall 5-year survival was 72.6% (95% CI, 67.4–77.7%). Survival differed by lesion class, mobility, and furcation involvement. Independent predictors of extraction were inadequate root canal obturation (HR 2.41), greater mobility (HR 1.64 per grade), furcation involvement (HR 1.45 per grade), probing depth (HR 1.30/mm), CBCT volumetric bone loss (HR 1.0054/mm³), diabetes mellitus (HR 1.87), and poorer oral hygiene (HR 1.02/%). The nomogram showed good discrimination (apparent C-index 0.812; bootstrap-corrected 0.802) and close agreement between predicted and observed 5-year survival.
Conclusion
A CBCT-integrated nomogram combining seven clinical, radiographic, and patient-level variables provided accurate individualised prediction of 5-year tooth survival in perio-endo combined lesions. External validation is required before routine clinical implementation.