RESUMO
BACKGROUND: The study compared clinical decisions regarding root angulation correction and root proximity based on the interpretation of Panoramic (PAN) versus Cone-Beam Computed Tomography (CBCT) images. METHODS: A total of 864 teeth from 36 existing, radiographic patient records at a university dental clinic with concurrent PAN and CBCT images were assessed using PANs, then using CBCTs in a blinded manner by two orthodontists. Teeth were rated regarding the need for root repositioning, the direction of repositioning and existence of root proximity. Frequencies, rating time and intra- and inter-examiner Cohen's Kappa were calculated. RESULTS: There was 73.7-84.5% agreement between PAN-based and CBCT-based orthodontists' decisions regarding the need to reposition roots. Root proximity was more frequently reported on PANs than CBCTs by one examiner (p = 0.001 and p = 0.168). Both PANs and CBCTs had moderate to substantial intra-examiner, within-radiograph-type reliability with Kappa values of 0.686-0.79 for PANs, and 0.661 for CBCTs (p < 0.001). Inter-examiner and inter-radiograph-type Kappa values ranged from 0.414 to 0.51 (p < 0.001). Using CBCT decisions as a reference, 78.9% of PAN decisions were coincident, 9.3% would have been repositioned on CBCT but not on PAN, 11.3% would not have been repositioned on CBCT but were on PAN, and 0.3% would have been repositioned in the opposite direction on CBCT versus PAN. Additionally, CBCT images required more time per tooth to assess than PANs (p < 0.001). CONCLUSIONS: PAN-based clinical decisions regarding root angulation had comparable statistical reliability and substantial agreement with CBCT-based clinical decisions.