Abstract
OBJECTIVES: Compare findings among observers in panoramic images (PAN) and cone beam CT (CBCT); and assess findings in PAN as indicators for marginal bone loss and resorption observed in CBCT. METHODS: 120 impacted maxillary third molars with PAN and CBCT were included. Four observers assessed morphological features: (1) tooth angulation; (2) number of roots; (3) bony impaction (yes/no) and pathology; (4) marginal bone level at the second molar (normal/>3 mm = bone loss); (5) resorption in the second molar (no/superficial/< half way through the dentin/≥ half way through the dentin/involving the pulp); (6) size of follicular space (normal/> 4 mm(cyst)). Percentage accordance and κ statistics described observer variation in PAN and CBCT. Logistic regression analyses tested findings in PAN as indicators for marginal bone loss or resorption observed in CBCT. RESULTS: κ values were fair and interobserver accordance was marginally higher in CBCT than PAN. Agreement between PAN and CBCT was 81-88% for marginal bone loss and 68-81% for resorption. Severe resorption was more often observed in CBCT. Mesio-angulated third molars and marginal bone loss interpreted in PAN significantly indicated marginal bone loss observed in CBCT (odds ration 17-34; p < 0.012; 8.8-52.8; p < 0.02). In contrast, findings in PAN were not significant indicators for resorption observed in CBCT (p > 0.05). CONCLUSION: In general, there was a fair agreement for marginal bone loss between PAN and CBCT, and PAN could significantly predict bone loss observed in CBCT. However, presence of resorption observed in CBCT could not be determined from PAN, and more severe resorption was observed in CBCT. CBCT is indicated if resorption in the second molar needs to be assessed.