Key points are not available for this paper at this time.
Abstract Dental age estimation (DAE) is applicable to forensic/legal and clinical contexts, with the Demirjian staging system being among the most common techniques. However, inter- and intra-observer variability can affect DAE reliability, particularly when evaluating panoramic radiographs influenced by magnification, distortion, and anatomical superimposition. In the staging process, consciously or unconsciously, experts analyze dental development within the available peripheral anatomic features of the tooth and its peripheral region. This study investigated whether isolating teeth as regions of interest (cropped images) influences the reproducibility of Demirjian’s stage allocation. An observational cross-sectional study was conducted using 301 panoramic radiographs from subjects aged 7–14 years. Three experienced forensic odontologists independently staged all seven left mandibular teeth (#31–#37) and, when present, also the left mandibular third molar (#38) using the Demirjian’s staging system. Each tooth was evaluated twice, two weeks apart, in both full panoramic radiograph and cropped formats. Inter- and intra-observer agreement was assessed using kappa statistics. Intra-observer agreement was above 0.81. Cropped images resulted in only marginal changes, with a slight decrease in stage reproducibility (Δ − 0.017 and − 0.023). Inter-observer agreement was nearly identical between non-cropped (AC2 = 0.985) and cropped images (AC2 = 0.986). A slightly lower proportion of cropped images was assessable by all observers (90.7% vs. 91.7%). Isolating teeth as regions of interest did not meaningfully reduce reproducibility and only small intra-observer decreases were observed and concentrated in the anterior teeth. These findings indicate that, among experienced observers, cropping does not degrade the reliability of Demirjian stage allocation. Future DAE staging systems should consider being more descriptive, including not only tooth-related developmental details, but surrounding information, such as the distance between the analyzed tooth and the occlusal plane and its relationship with adjacent developed teeth.
Angelakopoulos et al. (Thu,) studied this question.