Rapid MLS

Rapid MLS is a software device that measures brain midline shift from non-contrast CT (NCCT) scans using machine learning algorithms. It analyzes adult cases to identify the expected brain midline and any deviation, providing annotated images and quantitative measurements to clinicians. The device is not intended for standalone clinical decision-making but as an adjunct to clinical assessment.

Rapid MLS is an AI/ML-based medical image processing software operating within the Rapid Platform that accepts DICOM-formatted NCCT images and outputs visual overlays and JSON-formatted quantitative measurements. Key differences from the predicate (Rapid RV/LV) include different anatomical targets (brain midline vs. cardiac ventricles), different input modality (NCCT vs. CTPA), and different measurements (midline shift vs. ventricular diameter); however, both use command-line interfaces, AI/ML algorithms, and identical output formats (reports and DICOM secondary capture).

EN ISO 14971:2019 (risk management), IEC 62304:2006 (software lifecycle), IEC 62366:2015 (usability engineering), NEMA PS 3.1-3.20 (DICOM compliance), and UL 2900-1 (cybersecurity). The device was validated per 21 CFR 820.20 verification and validation requirements.

Although Rapid MLS measures a different anatomical structure (brain midline) using different imaging input (NCCT) than the predicate Rapid RV/LV (cardiac ventricles, CTPA), both devices share identical technical architecture (AI/ML, command-line interface, DICOM compliance, identical output formats), identical regulatory classification and product code (Class II, QIH), and identical intended use context (computer-assisted measurement with annotated outputs, not for standalone decision-making). Validation testing demonstrated Rapid MLS mean absolute error (0.7 mm) is non-inferior to expert performance (1.0 mm MAE), with Passing-Bablok analysis showing comparable methods. These commonalities in technological design and regulatory framework, combined with demonstrated performance equivalence despite different anatomical application, establish substantial equivalence.