Accuracy of Analog and Digital Full-Arch Mandibular Impressions: In Vitro and In Vivo Evaluation

<b>Background/Objectives:</b> Accurate full-arch impressions are crucial for predictable prosthodontic outcomes. While intraoral scanners (IOSs) are increasingly adopted, evidence comparing their accuracy with conventional analog impressions across full mandibular arches-particularly under both laboratory and clinical conditions using an objective intraoral reference-is limited. Our study aims to evaluate the in vitro and in vivo accuracy of digital impressions compared to conventional methods in full-arch scans using an intraoral reference tool. <b>Methods:</b> In this study, a custom stainless steel transfer aid carrying four 5 mm steel spheres in a trapezoidal configuration, provided with known reference distances, was used. Ten mandibular Frasaco models (in vitro) and ten healthy young adults (18-30 yrs) with intact lower arches (in vivo) received the bonded spheres. Six inter-sphere distances were defined: intermolar (BL-BR), interpremolar (FL-FR), diagonals (BL-FR, FL-BR), and lateral spans (BL-FL, BR-FR). Each arch underwent a digital scan (Medit i700) and a conventional monophase PVS impression, which was poured in Type IV stone and digitized (GOM Scan 1). The inter-sphere linear distances were measured in GOM Inspect, and trueness (deviation from reference) and precision (SD) were calculated. Data normality and homogeneity were verified; parametric <i>t</i>-tests and one-sample tests (α = 0.05) assessed differences between workflows and against reference values. <b>Results:</b> In vitro, analog impressions closely matched reference distances, with only the long-span BL-BR showing minor deviation (0.053 mm, <i>p</i> < 0.001). Digital scans showed significantly greater deviations across all spans (max 0.117 mm), particularly over long distances. In vivo, both workflows demonstrated comparable accuracy: only BL-BR (analog) and BR-FR (digital) differed significantly from the reference, and all AMEs remained within clinical thresholds (≤0.10 mm), except for BL-BR and BL-FL spans. ICC values ranged from moderate to high. Direct paired comparisons revealed statistically equivalent performance across most spans. <b>Conclusions:</b> Analog impressions outperformed digital scans in vitro, particularly across longer spans, confirming their superior dimensional fidelity under controlled conditions. However, in vivo, both workflows delivered statistically comparable and clinically acceptable accuracy. These findings suggest that while analog impressions remain the gold standard for precision-demanding contexts, modern intraoral scanners-when used correctly-can offer reliable full-arch mandibular impressions. The four-sphere reference system proved valuable for objective, anatomy-independent measurement.