DS Core Detect

DS Core Detect is a cloud-based computer-aided detection software that uses machine learning to help dentists identify permanent teeth with periapical radiolucencies (root infections) in CBCT (cone-beam CT) images. It is intended for licensed dentists to use on CBCT scans already obtained for other clinical reasons and is not a replacement for clinical judgment. The device is designed for adult patients aged 22 and older with permanent teeth.

The main technological difference from the predicate is that DS Core Detect detects periapical radiolucency on 3D CBCT images, whereas the predicate (Videa Dental Assist) detects multiple pathologies on 2D dental radiographs (bitewing, periapical, and panoramic). The subject device provides a list of flagged teeth and a visual PARL map for user confirmation or rejection, while the predicate outputs bounding boxes. Both are software as a medical device (SaMD) with AI/machine learning, but the subject device is limited to permanent teeth aged 22+ and incidental PARL detection only.

Not stated in this summary. The document references software verification and validation testing per applicable standards and agency guidance but does not cite specific ISO, IEC, or ASTM standards by number.

Both devices are diagnostic aids for periapical radiolucency detection and neither replaces standard clinical diagnosis. Standalone testing achieved sensitivity of 0.66–0.78 and specificity of 0.93–0.97 depending on operating point. Multi-Reader Multi-Case study showed the device improved reader sensitivity by 0.227 (95% CI: 0.124–0.330) with minimal specificity change (−0.016). These performance results, combined with the same regulatory class (Class II, Medical Image Analyzer) and analogous intended function, establish that DS Core Detect is as safe and effective as the legally marketed predicate device.