This systematic review and meta-analysis evaluated the comparative performance of artificial intelligence (AI) models versus comparators for working-length determination using radiographic or impedance-based inputs. A comprehensive search across seven electronic databases was conducted up to 1 October 2025, identifying five eligible in vitro and ex vivo studies encompassing over 1765 teeth or radiographic images. All the included studies directly compared AI-based approaches with manual or conventional reference standards. An exploratory random-effects meta-analysis demonstrated higher odds of correct working length determination using AI-based methods compared to expert assessment. The risk of bias was moderate to high, primarily because of internal validation and the predominance of laboratory-based study designs. The overall certainty of evidence for the primary outcome was rated low. This first quantitative synthesis suggests that AI-based methods may enhance the consistency of working-length determination under controlled conditions; however, further well-designed clinical studies are required before routine clinical implementation.
Bansal et al. (Sun,) studied this question.