Efficacy and safety of artificial intelligence-assisted image-guided navigation in complex endovascular aortic repair: a systematic review and meta-analysis.

BACKGROUND: Complex endovascular aortic repair exposes patients and operators to substantial radiation, contrast, and long procedures. Artificial intelligence (AI)-assisted image-guided navigation, based on cloud-based three-dimensional image fusion, has been proposed to improve intraoperative guidance, but its comparative benefit is uncertain. METHODS: We searched MEDLINE, Embase, the Cochrane Library, and Web of Science (last search July 2025) for studies comparing AI-assisted image-guided navigation with conventional fluoroscopy and digital subtraction angiography during complex endovascular aortic repair. No navigation platform was prespecified. Random-effects meta-analyses yielded mean differences (MD) with 95% confidence intervals (CI); heterogeneity was quantified with I2. Risk of bias was assessed with ROBINS-I and certainty with GRADE. RESULTS: Three retrospective cohorts (332 patients; 189 AI, 143 control), all evaluating one cloud-based platform (Cydar EV), were eligible. AI-assisted navigation was associated with a lower radiation dose (2 studies, 216 patients; MD -1435.85 mGy, 95% CI -2208.84 to -662.86; I2 = 0%). No significant differences were found in fluoroscopy time (MD -11.04 minutes, 95% CI -25.37 to 3.29; I2 = 65%), operative time (MD -16.39 minutes, 95% CI -37.21 to 4.42; I2 = 0%), or contrast volume (MD -30.59 mL, 95% CI -73.95 to 12.76; I2 = 94%). Mortality and major adverse events (single studies) did not differ significantly. CONCLUSIONS: Cydar EV-assisted navigation may reduce radiation dose during complex endovascular aortic repair, but evidence is limited to retrospective studies of a single platform and does not establish a class effect for other AI-assisted navigation systems. Prospective comparative studies across platforms are needed.