Daniel Palkovics

Introduction: Three-dimensional virtual planning may be an effective tool during surgical planning of guided bone regeneration (GBR) procedures to reduce complication rates. Alveolar ridge defects can be visualized with 3D models acquired by cone-beam computed tomography (CBCT) segmentation. To increase the efficiency of CBCT segmentation artificial intelligence (AI) based segmentation was previously proposed. The aim of this investigation was to validate the effectiveness of 3D modelling and AI in the planning of GBR procedures. Materials and Methods: Eleven CBCT scans were segmented using an AI model. Thereafter, alveolar ridge defects were digitally “augmented” by the modifying the 3D models. Based on the preoperative plan a customized membrane cutting template was designed and 3D printed for the shaping of the barrier membrane during surgery. 9 months following GBR postoperative CBCT scans were taken, and surgical outcomes were assessed and compared to the original preoperative plans. Results: The volumetric difference between the virtual plan and the actual surgical results showed no statistical significance. The planned volumetric hard tissue gain averaged at 0.75 ± 0.42 cm3, whereas the actual volumetric hard tissue gain averaged at 0.78 ± 0.50 cm3. AI segmentations of the pre- and postoperative CBCT scans did not show a statistically significant difference compared to ground truth segmentations of the same datasets. Conclusion: With the combined application of an AI segmentation model and with the custom membrane cutting templates the digital plans could effectively be transitioned into the clinical practice.

Daniel Palkovics has graduated from the Faculty of Dentistry, Semmelweis University, Budapest in 2016. After graduation he started his residency at the Department of Periodontology. He defended his PhD thesis in 2022 under the supervision of Professor Peter Windisch regarding possible application of 3D technologies in reconstructive periodontal surgery and implant dentistry. He is currently working as an assistant professor and at the Department of Periodontology, Semmelweis University. Besides treating patients on a daily basis, his main interest lies in three-dimensional radiographic image reconstruction, CAD modelling and virtual surgical planning of regenerative-reconstructive surgical procedures. His team has developed a digital workflow to acquire virtual patient models from CBCT and intraoral scan data, allowing to visualize the patient’s anatomy in 3D prior to intervention. Recently he has also turned his focus towards automation and artificial intelligence to simplify the surgical treatment planning process.