In-house 3D digital workflow for maxillomandibular deformities: segmentation as a key step in visualization and surgical planning

Medical image segmentation plays a central role in the generation of patient-specific anatomical models and represents a critical step in advanced clinical workflows such as surgical planning, computer-assisted surgery, and additive manufacturing. Despite significant technological advancements, segmentation remains a time-consuming and operator-dependent process, with variability that may affect accuracy and reproducibility. The aim of this doctoral thesis is to analyze, compare, and validate different segmentation approaches applied to medical imaging data, with particular focus on their reliability, accuracy, and clinical applicability. The work investigates multiple segmentation strategies, including manual, semi-automatic, and automated methods, applied to volumetric imaging datasets. The resulting three-dimensional models are evaluated using quantitative and qualitative metrics to assess geometric accuracy, consistency, and suitability for clinical use. A systematic validation process is performed, including comparison between different software tools and segmentation workflows. The results demonstrate that while automated and AI-based approaches can significantly reduce segmentation time, careful validation remains essential to ensure accuracy and reproducibility, especially in complex anatomical regions. This thesis highlights the strengths and limitations of current segmentation techniques and proposes a structured workflow for their integration into clinical practice. The findings support the role of validated segmentation pipelines as a key component in personalized medicine and digital surgical planning.