Accuracy of Augmented Reality-Assisted Navigation in Dental Implant Surgery: Systematic Review and Meta-analysis

Abstract

Background: The novel concept of immersive 3D augmented reality (AR) surgical navigation has recently been introduced in the medical field. This method allows surgeons to directly focus on the surgical objective without having to look at a separate monitor. In the dental field, the recently developed AR-assisted dental implant navigation system (AR navigation), which uses innovative image technology to directly visualize and track a presurgical plan over an actual surgical site, has attracted great interest.

Objective: This study is the first systematic review and meta-analysis study that aimed to assess the accuracy of dental implants placed by AR navigation and compare it with that of the widely used implant placement methods, including the freehand method (FH), template-based static guidance (TG), and conventional navigation (CN).

Methods: Individual search strategies were used in PubMed (MEDLINE), Scopus, ScienceDirect, Cochrane Library, and Google Scholar to search for articles published until March 21, 2022. This study was performed in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines and registered in the International Prospective Register of Systematic Reviews (PROSPERO) database. Peer-reviewed journal articles evaluating the positional deviations of dental implants placed using AR-assisted implant navigation systems were included. Cohen d statistical power analysis was used to investigate the effect size estimate and CIs of standardized mean differences (SMDs) between data sets.

Results: Among the 425 articles retrieved, 15 articles were considered eligible for narrative review, 8 articles were considered for single-arm meta-analysis, and 4 were included in a 2-arm meta-analysis. The mean lateral, global, depth, and angular deviations of the dental implant placed using AR navigation were 0.90 (95% CI 0.78-1.02) mm, 1.18 (95% CI 0.95-1.41) mm, 0.78 (95% CI 0.48-1.08) mm, and 3.96° (95% CI 3.45°-4.48°), respectively. The accuracy of AR navigation was significantly higher than that of the FH method (SMD=-1.01; 95% CI -1.47 to -0.55; P<.001) and CN method (SMD=-0.46; 95% CI -0.64 to -0.29; P<.001). However, the accuracies of the AR navigation and TG methods were similar (SMD=0.06; 95% CI -0.62 to 0.74; P=.73).

Conclusions: The positional deviations of AR-navigated implant placements were within the safety zone, suggesting clinically acceptable accuracy of the AR navigation method. Moreover, the accuracy of AR implant navigation was comparable with that of the highly recommended dental implant-guided surgery method, TG, and superior to that of the conventional FH and CN methods. This review highlights the possibility of using AR navigation as an effective and accurate immersive surgical guide for dental implant placement.

Keywords: accuracy; augmented reality; computer-guided surgery; dental implants; meta-analysis; systematic review.

Figure 1

Positional deviations between the planned and placed implants.

Figure 2

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flowchart showing the results of the study selection process.

Figure 3

Augmented reality image display devices. (A) OST HMD device. (B) VST HMD. (C) IV image overlay device. HMD: head-mounted display; IV: integral videography; OST: optical see-through; VST: video see-through.

Figure 4

Augmented reality (AR) dental implant navigation system. CBCT: cone-beam computed tomography; HMD: head-mounted display device; IV: integral videography.

Figure 5

Risk of bias assessment using the Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) tool. (A) Weighted bar plots showing the distribution of risk-of-bias judgments within each bias domain, and (B) traffic light plots of the domain-level judgments for each study.

Figure 6

Funnel plots for publication bias assessment. (A) Funnel plot. (B) Contour-enhanced funnel plot. (C) Contour-enhanced funnel plot with the trim-and-fill method. Note: The dashed lines represent the random-effects estimate and the corresponding 95% confidence limits; the shaded regions represent different significance levels for the effect size. The filled circles indicate the observed data reported in the included studies, and the empty circles indicate the imputed and added data after Duval and Tweedie trim-and-fill analysis.

Figure 7

Forest plot showing the positional deviation of the implant placed using augmented reality navigation. (A) Lateral deviation (mm). (B) Global deviation (mm). (C) Depth deviation (mm). (D) Angular deviation (°). Mx=maxillary; Mn=mandibular.

Figure 8

Forest plot comparing the positional deviation (mm) of implants placed using augmented reality (AR) navigation versus the freehand (FH), template-based static guidance (TG), and conventional navigation (CN) methods. AD: angular deviation; DD: depth deviation; GAD: global apical deviation; GCD: global coronal deviation; LAD: lateral apical deviation; LCD: lateral coronal deviation.

Figure 9

Forest plot comparing the positional deviations of the implants placed using augmented reality (AR) navigation versus the freehand (FH), template-based static guiding system (TG), and conventional navigation (CN) methods. (A) Lateral deviation (mm). (B) Global deviation (mm). (C) Depth deviation (mm). (D) Angular deviation (°). GAD: global apical deviation; GCD: global coronal deviation; LAD: lateral apical deviation; LCD: lateral coronal deviation.