A retrospective longitudinal assessment of artificial intelligence-assisted radiographic prediction of lower third molar eruption

Abstract

Prediction of lower third molar eruption is crucial for its timely extraction. Therefore, the primary aim of this study was to investigate the prediction of lower third molar eruption and its uprighting with the assistance of an artificial intelligence (AI) tool. The secondary aim was identifying the incidence of fully erupted lower third molars with hygienic cleansability. In total, 771 patients having two panoramic radiographs were recruited, where the first radiograph was acquired at 8-15 years of age (T1) and the second acquisition was between 16 and 23 years (T2). The predictive model for third molar eruption could not be obtained as few teeth reached full eruption. However, uprighting model at T2 showed that in cases with sufficient retromolar space, an initial angulation of < 32° predicted uprighting. Full eruption was observed for 13.9% of the teeth, and only 1.7% showed hygienic cleansability. The predictions model of third molar uprighting could act as a valuable aid for guiding a clinician with the decision-making process of extracting third molars which fail to erupt in an upright fashion. In addition, a low incidence of fully erupted molars with hygienic cleansability suggest that a clinician might opt for prophylactic extraction.

Figure 1

Calculation of lower third molar angulation. (a) Calculation of angulation on molars with fully formed roots where a line was drawn at region of largest coronal diameter and angulation line was drawn from most apical part of pulp chamber or most coronal part of bifurcation area; (b) calculation of angulation on molar with incompletely formed roots where the crown was divided into two equal halves, midpoint of widest diameter was taken and an inclination line was drawn perpendicularly (90°); (c) third molar angle defined based on angular difference (γ) between second (β) and third molar (α), represented by β − α = γ.

Figure 2

Eruption levels of lower third molar and retromolar space. (a) 1: fully erupted with hygienic cleansability where third molar is at level of 2nd molar’s occlusal plane with marginal bone situated beneath the CEJ at distal side, 2: fully erupted without hygienic cleansability where third molar is at level of 2nd molar’s occlusal plane with marginal bone above CEJ at distal side, 3: partially erupted with height of third molar contour above level of surrounding alveolar bone, 4: unerupted with third molar completely encased in bone; (b) available retromolar space, where 1: sufficient space, widest mesiodistal crown width of third molar fits available space measured between distal side of second molar till anterior border of ramus, 2: insufficient space, widest mesiodistal crown width of third molar does not fit available space.

Figure 3

Receiver operating characteristic (ROC) curves of third molar eruption prediction. (a) uprighting with reduced or insufficient retromolar space on training dataset; (b) uprighting with reduced or insufficient retromolar space on validation dataset; (c) uprighting with sufficient retromolar space on training dataset; (d) uprighting with sufficient retromolar space on validation dataset.