The diagnostic performance of AI based on dental radiographs in predicting marginal bone loss around dental implants: A systematic review and meta-analysis.

Statement of problem: The variable accuracy of conventional radiographic assessment makes reliable detection of marginal bone loss (MBL) around implants challenging. The diagnostic performance of artificial intelligence (AI) for this purpose remains unclear.

Purpose: The purpose of this systematic review and meta-analysis was to evaluate the diagnostic performance of AI using dental radiographs for detecting MBL around implants.

Material and methods: A systematic search of PubMed, Embase, and Web of Science was conducted up to February 2025, following the Preferred Reporting Items for a Systematic Review and Meta-analysis of diagnostic test accuracy (PRISMA-DTA) guidelines. Studies were included based on the population, intervention, control, outcome, study design (PICOS) framework, focusing on AI analysis of radiographs for detecting implant MBL. The revised Quality Assessment of Diagnostic Accuracy Studies (QUADAS)-2 tool assessed methodological quality. Sensitivity, specificity, and area under the curve (AUC) were pooled using a bivariate random-effects model (α=.05).

Results: Of 240 identified records, 9 studies (7284 participants) met inclusion criteria. The pooled sensitivity of AI for detecting MBL was 0.84 (95% CI: 0.72 to 0.91), specificity was 0.91 (95% CI: 0.81 to 0.96), and AUC was 0.94 (95% CI: 0.91 to 0.96). Compared with dentists, whose pooled sensitivity and specificity were 0.78 (95% CI: 0.70 to 0.84) and 0.70 (95% CI: 0.63 to 0.77), respectively, AI showed significantly higher specificity (Z=4.01, P<.001). The Deeks funnel plot indicated no significant publication bias (P=.20).

Conclusions: AI based on dental radiographs demonstrated excellent diagnostic performance for detecting implant MBL. The preliminary evidence suggests that AI may offer higher specificity than dentists, but this comparison was based on limited data and requires confirmation in future studies with larger sample sizes and more direct, head-to-head comparisons.