A Lucchese1, M Manuelli. 1. School of Dentistry, Department of Pedodontics, Faculty of Medicine and Surgery, University of Ferrara, Ferrara, Italy. aledolci@tiscalinet.it
Abstract
OBJECTIVES: Prediction of third molar eruption carries significant prognostic importance. There have been various reports of methods to predict the probability of impaction of these teeth. We studied the level of agreement between these three prediction methods. MATERIALS AND METHODS: a total of 61 patients, mean age 16.4+/-0.98, of Italian origin with initially retained 121 lower third molars were selected for this study. Panoramic radiographs were taken on each subject. For each third molar, the prognosis of eruption was assessed by three different cephalometric techniques: Olive, Ganss, Venta. The Kappa coefficient was used to compare results between the 3 methods (pairwise agreement) for each individual molar, the McNemar test to assess pairwise imbalance. RESULTS: a weak agreement was found between Ganss and Olive, fair between Venta and Ganss and between Olive and Venta. Additionally, the statistical analyses revealed a significant imbalance within the pairwise comparisons. CONCLUSIONS: the three methods were not found to be reliable or accurate; perhaps, the method of Castella et al. holds better promise.
OBJECTIVES: Prediction of third molar eruption carries significant prognostic importance. There have been various reports of methods to predict the probability of impaction of these teeth. We studied the level of agreement between these three prediction methods. MATERIALS AND METHODS: a total of 61 patients, mean age 16.4+/-0.98, of Italian origin with initially retained 121 lower third molars were selected for this study. Panoramic radiographs were taken on each subject. For each third molar, the prognosis of eruption was assessed by three different cephalometric techniques: Olive, Ganss, Venta. The Kappa coefficient was used to compare results between the 3 methods (pairwise agreement) for each individual molar, the McNemar test to assess pairwise imbalance. RESULTS: a weak agreement was found between Ganss and Olive, fair between Venta and Ganss and between Olive and Venta. Additionally, the statistical analyses revealed a significant imbalance within the pairwise comparisons. CONCLUSIONS: the three methods were not found to be reliable or accurate; perhaps, the method of Castella et al. holds better promise.