Second Opinion® 3D

The Second Opinion® 3D is an FDA-cleared radiological automated image processing system designed to serve as a concurrent diagnostic aid for dental professionals reviewing CBCT radiographs. This device leverages advanced computer vision and neural network-based machine learning algorithms to accurately identify and segment key anatomical structures, enhancing diagnostic support.

Pearl, Inc. has successfully obtained a determination from the U.S. Food & Drug Administration (FDA) for its Second Opinion® 3D device, classifying it as a radiological automated image processing system. This clearance, based on the device being found substantially equivalent to existing legally marketed products for its specified indications, grants Pearl, Inc. authorization to market Second Opinion® 3D. The device is now subject to standard general regulatory controls applicable to such medical products, signifying a significant step forward in its commercial availability for dental health professionals. The Second Opinion® 3D is a sophisticated software device engineered to serve as a concurrent, second reader for dental health professionals reviewing CBCT radiographs of patients aged 12 and older. Its core science relies on advanced computer vision and machine learning, specifically utilizing neural network algorithms, to automatically identify and precisely mark a range of clinically relevant anatomical structures. Performance bench testing unequivocally demonstrated the device's safety and effectiveness, achieving statistically significant accuracy in segmenting structures such as dentition, maxilla, mandible, inferior alveolar canal and mental foramen, maxillary sinus, nasal space, and airway, while also adhering to ongoing compliance requirements including quality system regulations and cybersecurity controls.

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 were the specific accuracy, sensitivity, and specificity metrics achieved during bench testing for each identified anatomical structure, and how do these compare to human expert performance?; What was the diversity and size of the dataset used to train the neural network algorithms, particularly concerning variations in patient age, ethnicity, and radiographic quality?; How does the device integrate into existing clinical workflows, and what is its expected impact on diagnostic efficiency and consistency in dental practices?; Are there any known limitations or specific types of pathologies or anomalies that the Second Opinion® 3D device is not designed to detect or that might lead to false positives/negatives?; What post-market surveillance plans are in place to monitor the device's long-term performance, especially regarding cybersecurity and real-world clinical effectiveness?