← Daily Brief
The Bell BriefFrom Dr. Jennifer Bell, DDS
HIGHClinicalTier 1

Automated Detection of Middle Mesial Canals in Mandibular Molars on CBCT using nnU-Net: A Retrospective Diagnostic Accuracy Study

SourceJournal of EndodonticsTier 1Peer-Reviewed Research

Originally at sciencedirect.com

Summary & scoring by The Bell Brief (Dr. Jennifer Bell) using the Drill-Down Protocol (Drill-Down Score) — not the original publisher.

Why it matters for dental

Endodontists and practice owners who treat mandibular molars can expect improved canal-location accuracy and lower missed-canal retreatment risk once CBCT-based AI algorithms such as nnU-Net are integrated into imaging workflows.

Key points

  • Retrospective study of 1,200 CBCT volumes from a Turkish academic center showed nnU-Net detected middle-mesial canals with 94.3 % sensitivity and 96.1 % specificity versus expert endodontist ground truth.
  • Middle-mesial canals were present in 22.7 % of the mandibular first molars studied—higher than the 10–15 % traditionally cited in textbooks.
  • Algorithm run-time averaged 6.2 seconds per volume on a standard clinical workstation, suggesting compatibility with chair-side CBCT review.
  • Authors note the model has not yet been externally validated on non-Turkish populations or different CBCT vendors, limiting immediate widespread adoption.

Who should care

SpecialistOwner

Read the original on Journal of Endodontics

Full reporting and any paywall content live on sciencedirect.com. We summarize and score; we do not republish.

Open original

Related