Influência de filtros de pós-processamento de imagens tomográficas no diagnóstico de desadaptação de coroas protéticas

Abstract

The application of post-processing techniques, such as brightness adjustments, contrast, and the use of digital filters, to improve tomographic images can provide more accurate diagnoses. This study aimed to evaluate the influence of digital filters on the enhancement of cone-beam computed tomography (CBCT) images for diagnosing marginal maladaptation of prosthetic crowns in an in vitro study. For this purpose, tomographic acquisitions were made for 12 teeth restored with ceramic crowns of lithium disilicate, properly adapted, and with maladaptations of 0.30 and 0.50 mm. The images were evaluated by three specialists in oral rehabilitation and three final-year Dentistry students, with a focus on identifying proximal gaps. Three conditions of postprocessing filter application were tested: "normal," "accentuated," and "very accentuated." Intra- and inter-examiner reproducibility was assessed using the Kappa index. The accuracy of gap detection was determined by the area under the ROC curve, and the values for each group of examiners and filters tested were compared using analysis of variance (ANOVA) and Tukey's post-hoc tests. The results showed that the intra- and inter-examiner agreement was considered moderate (p ≤ 0.05), with Kappa indices ranging from 0.32 to 0.79 (mean = 0.52 / SD = ± 0.21) and 0.21 to 0.88 (mean = 0.45 / SD = ± 0.13), respectively. There was no significant difference between the filters (p = 0.914). However, there was a significant difference between the examiners, with the specialists performing better in providing correct answers compared to the academics (p = 0.001). Therefore, it was concluded that, for the tomographic images obtained using the evaluated exposure protocol, the application of post-processing filters does not influence the diagnostic accuracy of marginal maladaptation in indirect restorations based on lithium disilicate.