InferRead Lung CT.AI

InferRead Lung CT.AI is a computer-assisted reading tool that uses deep learning to detect pulmonary nodules ≥4mm in chest CT scans of asymptomatic patients ≥55 years old. It displays nodule locations, measurements, and density information as CADe marks overlaid on original CT images, requiring both lungs to be in the field of view and the original CT series to be available.

The device performs automated nodule detection, segmentation, and measurement similar to its predicate (K192880), but adds three new capabilities: nodule lobe localization (identifying which lung lobe contains each nodule), nodule density quantification using Hounsfield Units, and automatic nodule registration between prior and current scans with change status tracking.

IEC 62304:2006+A1:2015 (medical device software lifecycle), ISO 14971:2019 (risk management), and AAMI TIR57:2016. Testing followed FDA guidance on premarket submissions for device software and cybersecurity in medical devices.

The core nodule detection and segmentation functions are consistent with predicate K192880 as verified through consistency testing. The three newly added features (lobe localization, density information, and nodule registration) are functional enhancements validated through standalone performance testing demonstrating high accuracy rates (0.970 match rate, 0.957 localization accuracy, 0.966 Dice coefficient), and these additions do not alter the fundamental intended use or introduce new safety risks since the device remains adjunctive and requires original CT images for clinical use.