Overjet Calculus Assist

Overjet Calculus Assist is an AI-powered software designed to enhance dentists' ability to detect interproximal calculus on bitewing and periapical radiographs by automatically highlighting suspected deposits with bounding boxes. Clinical performance data indicates the system significantly improves reader sensitivity in calculus detection, serving as a valuable adjunctive tool in dental radiographic analysis.

Overjet Calculus Assist (OCalA) is a radiological computer-assisted detection software intended to aid professionally trained and licensed dentists in identifying interproximal calculus deposits on both bitewing and periapical dental radiographs. It functions as a concurrent-read tool, analyzing images to provide supplementary information for diagnosis, rather than replacing a comprehensive dental examination or the clinician's judgment. The system processes various digital radiograph formats (e.g., JPEG, PNG, DICOM) of at least 500x500 pixels, typically imported from existing radiographic devices or practice management systems, and is designed for use by dentists to enhance their diagnostic capabilities for tooth surfaces suspected of having calculus deposits. The device utilizes an artificial intelligence algorithm to automatically identify and outline suspected calculus on radiographs using a bounding box. Its architecture comprises a network layer for image and metadata handling, a decision layer for algorithmic image processing and annotation, and a presentation layer that displays the annotated images in a non-diagnostic viewer, allowing dentists to filter, display, hide, or edit annotations. Performance evaluations included standalone testing, which showed algorithm sensitivity for calculus detection at 74.1% for bitewing and 72.9% for periapical radiographs, with specificity at 99.4% and 99.6% respectively. A clinical reader study with 14 dentists demonstrated that using OCalA improved their detection sensitivity by approximately 9-10% (from around 75% to 84%) for both radiograph types, with a minimal decrease in specificity (0.1-0.2%), and yielded statistically significant improvements in AFROC AUC values.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the Overjet Calculus Assist algorithm differentiate between interproximal calculus and other radiopaque dental structures or artifacts, particularly in images with suboptimal quality?; What is the specific impact of the observed slight decrease in specificity (0.1-0.2%) on clinical workflow, such as the rate of false positives, and are there recommended protocols for dentists to manage these?; Could the 'polygon model scores' used for AFROC curve generation be translated into a clinically intuitive confidence score for each bounding box presented to the dentist?; Beyond pixel dimensions (500x500), what are the optimal image acquisition parameters (e.g., exposure, sensor type compatibility) for the Overjet Calculus Assist to achieve maximal sensitivity and specificity?; What are the computational requirements and processing times for the cloud-native Overjet Calculus Assist to analyze a typical set of bitewing and periapical radiographs in a real-world clinical setting?