Rayvolve LN

Rayvolve LN is a computer-aided detection software that assists radiologists in identifying and marking suspected pulmonary nodules (6-30mm) on frontal chest X-rays (AP/PA views) in patients 18 years or older. It processes DICOM images and provides adjunctive information as a second reader tool, not as a substitute for the original radiograph.

Rayvolve LN uses supervised deep learning (versus the predicate's machine learning), detects nodules from 6-30mm (predicate: 10-30mm), supports both AP and PA views (predicate: PA only), and works with any DICOM Node server rather than being tied to a specific imaging system. Both use DICOM input format and output regions of interest (ROI) marked on duplicated images.

Not stated in this summary.

Both devices share identical intended use (second-reader CAD for pulmonary nodule detection on chest radiographs), regulatory classification (Class II, MYN code), and clinical objective. The predicate's size range (10-30mm) is subsumed within Rayvolve's range (6-30mm); clinical testing demonstrated equivalent accuracy for smaller nodules. Support for AP views (clinically relevant for bedside radiography) was validated to achieve equivalent accuracy to PA views. Performance testing (sensitivity 0.8847, specificity 0.8294, AUC 0.8408 on 2181 radiographs; reader study showing AUC improvement from 0.8071 to 0.8583 when aided) demonstrates comparable or superior performance with no new safety or effectiveness concerns.