Agreement and reproducibility of intraoral scan- versus CBCT-based three-dimensional implant position assessment in the posterior mandible — a paired in vitro method-comparison study

Background Three-dimensional postoperative implant assessment is increasingly performed using intraoral scan (IOS) superimposition as a radiation-free alternative to cone-beam computed tomography (CBCT). However, it remains unclear to what extent IOS- and CBCT-based measurements agree and how reproducibility changes across posterior mandibular configurations. Methods In this paired in vitro method-comparison study, based on the experimental platform of a previous study, 60 implants placed at Fédération Dentaire Internationale (FDI) positions 35, 36, 45, 46 and 47 in bounded edentulous spaces (BES) and distal extension situations (DE) underwent three IOS-based and three CBCT-based postoperative assessments relative to the virtual plan using the Treatment Evaluation Tool of the planning software. Angular, mesiodistal, buccolingual, apico-coronal, and three-dimensional (3D) deviations at implant base and apex were recorded, resulting in 180 paired measurements. Agreement was assessed by Bland-Altman analysis, and within-method reproducibility by the within-implant standard deviation (Sw). IOS variability was additionally separated into within-scan and between-scan components. Non-parametric analyses were used for comparisons by indication and FDI position. Results Mean absolute IOS–CBCT differences were 0.73° ± 0.55° for angular deviation, 0.22 ± 0.20 mm for the 3D base offset and 0.26 ± 0.22 mm for the 3D apical offset. Bland-Altman bias (CBCT − IOS) was small (− 0.35°; −0.05 mm; −0.07 mm) with 95% limits of agreement of [− 2.01°; +1.31°], [− 0.62; +0.53] mm and [− 0.73; +0.58] mm. Within-method reproducibility was superior for IOS in 8 of 9 parameters (e.g., Sw 0.20° vs. 0.55° for the angle), and IOS variability was dominated by scan-to-scan rather than within-scan re-evaluation. Discrepancies were larger in DE than in BES sites in 7 of 9 parameters (all p < 0.001 for the principal apical and buccolingual components) and were larger at position 47 than at position 35 (3D base p = 0.004; 3D apex p = 0.011). Conclusions In this in vitro setting, IOS- and CBCT-based implant assessment in the posterior mandible showed good agreement with small overall differences, and IOS demonstrated lower within-method variability than CBCT. IOS-based superimposition may therefore serve as a radiation-free complement to CBCT for the postoperative evaluation of single implants in bounded posterior sites. Agreement decreased in distal extension situations and at more distal positions; in these configurations, and whenever clinical decisions depend on sub-millimeter or sub-degree precision, CBCT remains the reference technique. Confirmation in vivo is required before routine clinical use.