Comparison Between Conventional and Artificial Intelligence-Assisted Setup for Digital Implant Planning: Accuracy,...

Area of Science:

• Dental implantology
• Medical imaging analysis
• Artificial intelligence in healthcare

Background:

• Computer-assisted implant planning relies on accurate segmentation and registration.
• Conventional manual methods for these tasks can be time-consuming.
• Clinician experience may influence the efficiency and reliability of manual workflows.

Purpose of the Study:

• To compare the reliability and time efficiency of AI-driven versus conventional manual segmentation of the mandibular canal.
• To evaluate the accuracy of AI-automated registration of cone-beam computed tomography (CBCT) with intraoral scan data.
• To assess the impact of clinician experience on these automated and manual processes.

Main Methods:

• Twenty clinicians (10 moderate, 10 high experience) performed manual mandibular canal segmentation and CBCT-model registration.
• The same procedures were repeated using an AI tool for comparison.
• Statistical analysis utilized a mixed model to assess significance.

Main Results:

• AI-automated operations were significantly faster (2.03 min) than manual ones (4.75 min) (p < 0.001).
• Accuracy was comparable between manual and AI-assisted methods for both mandibular canal segmentation (0.71 vs. 0.68 mm RMS error) and CBCT-model registration (0.45 vs. 0.37 mm discrepancy).
• No significant difference in accuracy was observed between manual and AI approaches (p > 0.05).

Conclusions:

• AI-automated tools are feasible for implant planning, offering similar or improved accuracy over manual workflows.
• These AI tools enhance time efficiency for clinicians regardless of their experience level.
• Further research with diverse software and datasets is needed to generalize findings.