Abstract: PURPOSE: To explore the risk factors for mesioangular and vertical impactions of the mandibular third molars and to construct a predictive model based on logistic regression analysis. METHODS: Clinical data of 243 mandibular third molars collected from June 2021 to December 2023 at Tianjin Stomatology Hospital were classified into the eruption group and the impaction group, with the latter including mesioangular and vertical unilateral impactions. The clinical data were subjected to univariate analysis to screen for statistically significant factors, followed by multivariate analysis using logistic regression to further delineate risk factors for mandibular third molar impaction, with the construction of a nomogram for prediction.SPSS 27.0 software package was used for statistical analysis. RESULTS: Totally 243 mandibular third molars were included, and 75 (30.86%) were in the eruption group and 168 (69.14%) in the impaction group. No significant difference was found between the groups regarding age, gender, number of tooth roots, Co-Go, Co-Cop, W2, W3, W4 and L (P>0.05). Significant differences were observed between the eruption and impaction group concerning Nolla, L-6 missing, L-E missing, Co-Pog, Co-Go/Co-Pog, L6-MP, α and W1(P<0.05). Multivariate regression analysis revealed that Nolla, absence of L-6, absence of L-E, Co-Pog, Co-Go/Co-Pog, L6-MP, α and W1 were independent risk factors for mesioangular and vertical impactions of the mandibular third molars (P<0.05). The construction of nomogram demonstrated high predictive accuracy. Analysis of the receiver operating characteristic curve(ROC) indicated that the area under the curve(AUC) for the joint prediction of mesial and vertical impaction of the mandibular third molar by independent risk factors was 0.924, with a 95%CI of 0.887 to 0.960. The sensitivity was reported to be 86.9%, and the specificity was 86.7%. CONCLUSIONS: Nolla, absence of L-6, absence of L-E, Co-Pog, Co-Go/Co-Pog, L6-MP, α and W1 are major risk factors affecting the impaction of mandibular third molars. The use of logistic regression analysis and nomograms can effectively predict the risk of impaction, providing a scientific basis for clinical treatment.

Key words: Mandibular third molar, Mesioangular impaction, Vertical impaction, Risk factors, Logistic regression analysis