Nerveblox

Nerveblox is software integrated into compatible ultrasound systems (GE Venue models) that assists qualified healthcare professionals in identifying and highlighting anatomical structures during ultrasound-guided peripheral nerve block procedures. It uses AI/ML to apply color overlays, add anatomical labels, and provide image quality scores prior to needle insertion, supporting 12 different nerve block regions for adult patients 18 years and older.

Unlike the predicate device (ScanNav), which runs on dedicated hardware with HDMI input, Nerveblox is software hosted directly on commercial ultrasound systems using a programming interface. Nerveblox provides name labels and quality scores in addition to anatomical highlighting, and highlights veins as a key anatomical structure. Both devices use locked deep learning CNNs, but differ in hardware implementation and output features.

Not stated in this summary.

Both devices share identical intended use—assisting clinicians in identifying and highlighting anatomical structures for ultrasound-guided regional anesthesia prior to needling. Though they support different block regions (10 vs. 12) and differ in hardware architecture and output features, these differences do not raise different safety or effectiveness questions because analytical validation demonstrated anatomical detection accuracy >97%, false-positive rates <1%, and quality score agreement with expert assessments (weighted Kappa ≥0.77). Clinical testing on 80 scans confirmed high accuracy (97%) and expert-assessed risk reduction for adverse events and block failure. Human factors testing with 22 board-certified anesthesiologists found no use errors or UI design issues impacting safe and effective use.