Artificial Intelligence-Guided Versus Manual CBCT Planning for Immediate Implant Placement

This study is a prospective, parallel-arm, randomized controlled clinical trial designed to evaluate the clinical impact of artificial intelligence (AI)-based CBCT analysis on decision-making for immediate implant placement in molar extraction sites. Following eligibility confirmation and informed consent, participants requiring molar extraction with potential immediate implant placement will undergo standardized preoperative cone-beam computed tomography (CBCT) imaging. Participants will be randomly allocated in a 1:1 ratio to one of two planning workflows: AI-guided planning arm: CBCT scans will be analyzed using a pre-specified, locked AI-based segmentation and socket assessment model. The AI system will quantify interradicular septum dimensions and generate a feasibility classification based on predefined anatomical criteria. Manual planning arm: CBCT scans will undergo conventional manual segmentation and assessment by an experienced radiologist using the same predefined anatomical criteria for feasibility determination. In both arms, feasibility recommendations will be based on identical, prospectively defined decision thresholds to ensure comparability between planning methods. The operating surgeon will receive only the planning report corresponding to the assigned allocation. All surgeries will be performed according to a standardized surgical protocol. The primary outcome is intraoperative feasibility of immediate implant placement, defined as successful implant placement in the extraction socket with achievement of primary stability according to prespecified stability criteria documented in the operative record. Cases in which implant placement is not performed or is aborted due to inability to achieve adequate primary stability will be classified as non-feasible. Secondary outcomes include operative time, need for intraoperative modification of the treatment plan, insertion torque values, and any intraoperative complications. Outcome assessment will be performed by an independent assessor blinded to allocation. AI analysis will be conducted using a locked model without post hoc modification. The study aims to determine whether AI-guided CBCT planning is non-inferior or superior to conventional manual CBCT assessment in supporting immediate implant placement decisions.