Overjet Charting Assist

Overjet Charting Assist is a medical image processing system that aids dental professionals by automatically analyzing 2D radiographs (bitewing, periapical, panoramic) to detect natural tooth structures, numbering, and various restorative treatments for efficient dental charting. The system utilizes image analysis technology, likely machine learning, and uniquely incorporates a Predetermined Change Control Plan (PCCP) enabling iterative performance enhancements by reducing false positives/negatives through model retraining with real-world data, without altering its core intended use.

Overjet Charting Assist is a Medical Image Management and Processing System (MIMPS) designed to support dental professionals in creating dental charts. It functions by analyzing 2D dental radiographs—specifically bitewing (BW) and periapical (PA) images for patients aged 5 and above with primary and/or permanent teeth, and panoramic (Pano) radiographs for patients with only permanent teeth. The system automatically detects natural dental structures, including tooth anatomy like enamel and pulp, performs tooth numbering, and identifies dental structures from past restorative treatments such as implants, crowns, endodontic treatments, and fillings. The device serves as an assistive tool and is not intended to replace a clinician's comprehensive review or clinical judgment, which should consider all relevant patient information. The core technology relies on image analysis to interpret radiographs and extract relevant dental information. A key feature of this device is the inclusion of a Predetermined Change Control Plan (PCCP). This plan outlines a structured process for implementing modifications aimed at improving performance by reducing false positive and false negative outputs, based on post-market and real-world data which facilitates re-training of its machine learning model. The PCCP ensures modifications are applied manually and uniformly, with algorithms trained, tuned, and locked before release, rather than through continuous field learning. This PCCP is the primary difference from its predicate device (K233590), with other features, indications, and patient populations remaining unchanged, and existing performance data from K233590 considered applicable.

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. What specific machine learning algorithms are employed by Overjet Charting Assist for dental structure detection, and what were their documented performance metrics (e.g., sensitivity, specificity, accuracy) in the validation studies for the predicate device (K233590)?; How does the Predetermined Change Control Plan (PCCP) specifically define the quantitative thresholds for 'major drop in performance' that would prevent a proposed algorithm modification from being implemented?; What are the specific protocols for collecting, curating, and annotating the 'post-market and real-world data' used for re-training the ML model, particularly concerning data diversity and bias mitigation?; How will the updates to the 'Instructions for Use,' detailing algorithm changes and revised performance characteristics under the PCCP, be communicated to end-users, and what mechanisms are in place to ensure users understand these changes?; Beyond reducing false positives and negatives, does the current PCCP allow for the potential expansion of detectable dental conditions or features, or integration with new types of radiographic imaging modalities in future iterations?