CEPHX3D

CEPHX3D is a cloud-based software program that analyzes dental and craniomaxillofacial CBCT (Cone Beam Computed Tomography) images using deep learning algorithms to automatically segment and reconstruct key anatomical structures including the mandible, maxilla, teeth, and inferior alveolar nerve. It displays these 3D models in an interactive viewer to support diagnostic processes and treatment planning for patients aged 12 and older.

CEPHX3D is a web-based application using 3D U-Net Convolutional Neural Networks for volumetric segmentation, built on Spring Framework (Java) and hosted on AWS Cloud. It accepts CBCT images as input and outputs STL and DICOM formats with measurement and manipulation tools. The predicate device (Relu Creator K233925) also operates as a web application but accepts additional input types (intraoral and facial scans) and outputs additional formats (PLY, OBJ). Both provide 3D viewing, editing, measurement, and treatment planning simulation capabilities.

Not stated in this summary.

CEPHX3D is substantially equivalent to Relu Creator (K233925) because both share identical intended use for analyzing dental and craniomaxillofacial images to support diagnostic and treatment planning processes. Although CEPHX3D emphasizes AI-driven processing while Relu Creator emphasizes image enhancement, both achieve comparable 3D modeling outcomes. Differences in input types and output formats do not raise new safety or effectiveness questions. Performance validation demonstrated exceptional accuracy with skeletal bone DSC of 0.9827, dentition DSC of 0.9993, and IAN DSC of 0.9861, with 100% clinical acceptability in a 53-case review.