Detection and diagnosis of mandibular cysts represents a great challenge for doctors in the field of maxillofacial surgery. This paper focuses on the development of a model that will enable automatic detection and identification of cysts in radiological images using deep neural networks, specifically the YOLOv8 model. The aim of the research is to evaluate the performance of the model through four different tasks: detection, multi-class detection, segmentation and multi-class segmentation. The dataset consists of 200 orthopantomograms containing a total of 226 cysts. The images were labeled and the cysts classified into five different categories: radicular cyst, ameloblastoma, keratocystic odontogenic tumor, follicular cyst and residual cyst. Marking of the orthopantomograms was done with the help of a maxillofacial surgeon, a maxillofacial surgery specialist and a radiologist. Advanced augmentation methods were applied to improve the quality and diversity of the available data set. The evaluation of the model's performance was carried out using standard metrics: precision, recall, mAP@50, mAP@50-95. The trained models show superior results in the tasks of detection (precision = 1, recall = 0.945, mAP@50 = 0.966, mAP@50-95 = 0.714) and segmentation (precision = 1, recall = 0.945, mAP@50 = 0.966, mAP@50-95 = 0.714). 50-95 = 0.722) and slightly worse results in the tasks of multi-class detection (precision = 0.64, recall = 0.662, mAP@50 = 0.735, mAP@50-95 = 0.598) and multi-class segmentation (precision = 0.64, recall = 0.662, mAP@50 = 0.735, mAP@50-95 = 0.595). This research points to the utility and potential of deep neural networks in automatic cyst diagnosis, providing physicians with a better basis for making decisions regarding diagnosis and treatment planning.