In vitro evaluation of AI-assisted CBCT analysis for detecting additional canals in mandibular premolars

Failure to detect additional canals in mandibular premolars can compromise endodontic treatment outcomes. Hence, this in vitro study compared an artificial intelligence-assisted CBCT analysis system with endodontic specialists for detecting additional canals, using micro-CT as the gold standard. One hundred and fifty mandibular premolars were analyzed. The AI system demonstrated higher sensitivity (86.8%) and accuracy (90.0%) than specialists, with comparable specificity and interpretation time of less than three seconds per case. Thus, AI-based CBCT analysis appears to be a rapid and reliable adjunct for improved canal detection and preoperative endodontic planning.