Model Predicted Vertical Root Fracture From Routine Clinical and Radiographic Signs

Edited by Dentaltown staff

A prediction model built from findings available at an ordinary clinical and radiographic examination identified vertical root fracture in root-filled teeth with an area under the curve of 0.84, according to a study published online Sept. 15 in the International Endodontic Journal.

Investigators analyzed 415 root-filled roots from 327 patients who underwent endodontic microsurgery or surgical exploration. Fracture was surgically confirmed in 130 roots, and 285 were free of fracture.

Fourteen candidate predictors covering patient demographics, tooth characteristics, clinical signs and symptoms, and radiographic features were evaluated. Multivariable logistic regression with backward stepwise elimination retained seven: age, tooth type, presence of a sinus tract, probing depth of 5 mm or greater, post type, periapical radiographic status, and the ratio of root canal space to root width.

Five of the seven reached statistical significance. An isolated perilateral radiolucency carried the strongest association, at an adjusted odds ratio of 21.56, followed by probing depth of 5 mm or greater at 7.64, halo radiolucency at 8.67, a root canal space-to-root width ratio greater than one-third at 3.28, and age at 1.03.

The model’s area under the receiver operating characteristic curve was 0.84, with a 95% confidence interval of 0.80 to 0.89. Internal validation by bootstrapping across 1,000 replicates produced an optimism-corrected figure of 0.81, with a confidence interval of 0.76 to 0.85.

Risk stratification based on likelihood ratios sorted roots into five levels. Roots classified as highly suspect carried a positive predictive value of 94.5%.

Jirapat Chalermchaicharoenkit, Kanet Chotvorrarak, Pattaraporn Promchouy, and Sasiprapa Osiri of the Faculty of Dentistry at Mahidol University in Bangkok conducted the study with Phichayut Phinyo of the Department of Biomedical Informatics and Clinical Epidemiology at Chiang Mai University.

Two limits bear on how far the numbers travel. The model has been validated only internally, and its optimism-corrected calibration slope of 0.78 carried a confidence interval running from 0.42 to 1.02, which indicates some overfitting. The cohort also consisted entirely of teeth referred for surgery, where fracture prevalence was 31%, so the positive predictive value would not carry over unchanged to a general practice population with a lower underlying rate.

Sources:
International Endodontic Journal, “Developing a Clinical and Conventional Radiographic Tool for Vertical Root Fracture Prediction in Root-Filled Teeth,” published online Sept. 15, 2026: doi.org/10.1111/iej.70272
PubMed record, PMID 42741958: pubmed.ncbi.nlm.nih.gov/42741958