The use of Cone Beam Computed Tomography (CBCT) is increasing not only in dentistry, but in the whole field of head and neck surgery. The main advantages of CBCTs are related to the short acquisition time and to the low radiation dose, while keeping an excellent visualization of anatomical structures, especially of hard tissues. In this regard, during the previous year challenge (ToothFairy) we have tackled the segmentation of the Inferior Alveolar Canal (IAC), a noble structure that lies within the mandible, whose identification and preservation represent a primary objective of many surgical interventions. In 2024 challenge, we aims at increasing the number of anatomical structures to be considered in the segmentation, thus including the mandible, the teeth, the maxillary bone, and the pharynx. Their framing is cross-disciplinary, as they are involved in all the head and neck surgical specialties, as well as in clinical and anesthesiological daily practice. In this regards, deep learning models can support medical personnel in the surgical planning procedures by providing an automatic voxel-level segmentation. The challenge we propose aims at pushing the development of deep learning frameworks to segment anatomical structures in CBCTs by incrementally extending the amount of publicly available 3D-annotated CBCT scans and providing the first public-available fully annotated dataset. With respect to ToothFairy this new edition appears as an innovative and multidisciplinary one, expanding the field of view and the tasks of interest of the previous challenge in the perspective of a ever increasing cross-disciplinarity and clinical application.