RevealAI-Lung

RevealAI-Lung is a computer-aided diagnostic (CADx) software that analyzes lung CT images to characterize incidentally-detected lung nodules and provide a malignancy similarity index (mSI) score to assist radiologists in assessing cancer risk. The mSI score is indicated for evaluating nodules 6-15mm in diameter in patients aged 18 years or above; it is not indicated for lung cancer screening or unassisted reads.

RevealAI-Lung uses machine learning to extract thousands of imaging features from CT data and provides an mSI score (0–1 scale) compared to a reference database of confirmed diagnoses from the National Lung Cancer Screening Trial. It integrates with MIMPS (Medical Image Processing and Management System) viewing systems, whereas the predicate Optellum Software requires web browser access and GPU-equipped servers. Both devices use similar AI-based scoring approaches but differ in deployment architecture and user interface design.

Design, verification, and validation followed FDA Guidance for Software Contained in Medical Devices (Moderate Level of Concern), 21 CFR §892.2060 special controls, FDA Guidance for Human Factors and Usability Engineering, FDA Guidance for Cybersecurity in Medical Devices, and ISO 14971:2019 for risk management.

RevealAI-Lung is substantially equivalent to the predicate Optellum Software because both are CADx software devices classified under 21 CFR §892.2060 for characterizing lung nodules suspicious for cancer. Both use machine learning algorithms to extract imaging features and provide a score relative to a database of known cases, display histograms of benign versus malignant case distributions, and require clinician review for final diagnostic decisions. Nonclinical and clinical performance testing demonstrated RevealAI-Lung performs as safely and effectively as the predicate, with clinical study results showing consistent improvements in radiologist accuracy across diverse patient populations and nodule characteristics.