| K-number | K251072 |
| Device name | Segmentron Viewer |
| Applicant | DGNCT, LLC |
| Product code | QIH |
| Device class | Class II |
| Decision date | Sep 9, 2025 |
| Decision | Substantially Equivalent |
| Regulation | 892.2050 |
Segmentron Viewer is a software-as-a-medical-device (SaMD) that processes and manipulates maxillofacial radiographic images, specifically CBCT scans. It uses artificial intelligence to automatically segment teeth, tooth pulp, and anatomical structures; generates 3D visualizations and multi-planar reconstructions; and produces editable 3D STL files for educational purposes. The device is intended for use by dental professionals and radiologists treating patients 14 years and older with permanent teeth.
Both Segmentron Viewer and its predicate (Ez3D-i/E3) are software-only, AI-based devices using supervised machine learning to analyze CBCT images and provide comparable tools for dental image processing. Both support DICOM format, offer MPR and 3D visualization, generate segmentation reports, and enable data export. Key differences include: Segmentron Viewer is web-based (predicate is desktop), Segmentron includes pulp segmentation (predicate does not), and the predicate supports broader imaging modalities and includes additional features like implant simulation and bone density profiling.
Not stated in this summary.
Segmentron Viewer meets the substantial equivalence standard because it shares the same fundamental intended use as the predicate—providing dental professionals with AI-based tools to visualize, analyze, and segment maxillofacial radiographic images. Although Segmentron has a narrower range of input imaging sources (CBCT only) compared to the broader modality support of the predicate, this difference does not raise new safety or effectiveness questions because both devices focus on accurate visualization and segmentation of dental and maxillofacial anatomy. Performance validation data demonstrate Segmentron achieves strong segmentation accuracy (Dice Coefficient 0.96 for teeth, 0.88 for pulp) and 100% labeling accuracy against expert radiologist reference standards, demonstrating it functions safely and effectively as intended, comparable to the predicate.
View the full FDA submission: accessdata.fda.gov