Image-guided photogrammetry accuracy: In vitro evaluation of an implant-supported complete arch digital scanning technology

Complete arch digital scanning with intraoral scanners (IOSs) for implant-supported fixed dental prostheses (FDPs) is controversial with current technologies, and accuracy and consistency are insufficient. Stereophotogrammetry (SPG) may capture precise implant coordinates but unrelated to the patient anatomy. Conclusive data comparing IOS versus SPG are lacking. The purpose of this in vitro study was to investigate a novel image-guided photogrammetry (IGP) navigation technology for complete arch digital scanning with an artificial intelligence (AI) driven workflow to superimpose implant coordinates on preoperative patient data sets. After ethical approval had been received for the use of human-derived data, complete arch casts of actual patients with 4 and 6 implants were selected from an archive. Digital scans were made with IGP and a coordinate measuring machine (CMM) to generate standard tessellation language (STL) test and reference files for each cast, which were then superimposed with a best-fit algorithm. For each implant position, linear along x- (longitudinal), y- (lateral), and z- (vertical) axes (ΔX, ΔY, ΔZ) and angular (ΔANGLE) deviations were calculated. Three-dimensional (3D) deviation (ΔEUC) was measured as the Euclidean distance. Sample size was calculated assuming ΔEUC as the primary endpoint, significance level of.05, n=84 to ensure a minimum expected difference of 120 µm (standard deviation 150 µm), and test power of.95. A univariate descriptive analysis of study variables was performed and stratified by implant number (4 versus 6). An independent samples t test was used to assess whether ΔEUC and ΔANGLE differed significantly based on implant number (4 versus 6) and position on the arch (anterior versus posterior for 4-implant casts) and anterior, intermediate, and posterior for 6-implant casts) (α=.05). A total of 26 complete arch casts, mandibular (n=16) and maxillary (n=10) with 4 (n=15) or 6 (n=11) implants, were investigated. A total of 126 implant positions were captured with IGP technology and compared with respective references (n=26 test, n=26 reference files). Mean deviations, standard deviation, range, and 75 percentiles were calculated for ΔX (-0.05 ±29.95; -75.17 - 68.00; 22,00 µm), ΔY (0.01 ±31.65; -68.0 - 93.93; 19,18 µm), ΔZ (-0.02 ±12.95; -39.0 - 43.5; 6,76 µm), ΔEUC (43.02 ±26.24; 3.41 - 190.23; 58,21 µm), and ΔANGLE (0.43 ±0.22; 0.03 - 0.86; 0,61 degrees). Lower deviations were found in 4-implant casts ΔEUC (P=.017), ΔANGLE (P=.003). Implant position did not affect accuracy, with the exception of ΔANGLE on 6-implant casts, and anteriors performed worse than posteriors and intermediate implants (P=.009). Image-guided photogrammetry was feasible for implant-supported complete arch digital scanning with linear and 3D deviations lower than those of conventional SPG. Linear deviations were close to 0 µm and angular deviations about 0.4 degrees, with the 4-implant casts being more accurate. Implant position did not affect accuracy with the exception of anterior implants that showed higher angular deviations in the 6-implant casts. Even though actual patient casts were investigated, allowing the IGP outcomes to be generalized with caution, trials are needed to validate its clinical performance.