InferRead Lung CT.AI

InferRead Lung CT.AI is a software tool that employs advanced image analysis algorithms to aid radiologists in detecting pulmonary nodules on chest CT scans from asymptomatic patients. Its server-based methodology processes images to provide toggleable annotations and quantitative nodule data, with clinical studies demonstrating significant enhancements in detection accuracy and reduced reading times for users.

InferRead Lung CT.AI is a software-as-a-service (SaaS) solution designed to aid radiologists in the detection of pulmonary nodules during the review of CT examinations of the chest on an asymptomatic population. The system utilizes a browser/server architecture; its server processes chest CT images received from PACS, RIS, or CT scanners using integrated algorithm logic and a database, and then provides output annotations regarding lung nodules for viewing on existing PACS workstations. Key software modules include Image reception (Docking Toolbox), Image predictive processing (DLServer), Image storage (RePACS), and Image display (NeoViewer), with the system providing nodule characteristics such as diameter and volume without modifying the original CT scan, instead overlaying toggleable CADe marks. The device's science involves algorithmic identification of solid and sub-solid nodules 4mm and larger, also supporting visualization of smaller ones. Non-clinical evaluations confirmed software integrity, measurement reproducibility, accuracy, and nodule detection sensitivity comparable to established systems. A pivotal multi-reader, multi-case clinical study involving 10 radiologists and 249 scans demonstrated that the use of InferRead Lung CT.AI significantly improved radiologists' nodule detection performance, evidenced by a 0.073 increase in the area under the LROC curve, and also significantly decreased reading times by an average of 23 seconds per scan compared to unaided readings.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. What specific deep learning architectures or algorithms are employed by the DLServer module for nodule detection, and how were they trained and validated prior to the pivotal study?; How does the system's performance vary with different types and sizes of pulmonary nodules (e.g., solid vs. sub-solid, ground-glass opacities, nodules <4mm vs. >4mm)?; What was the range of experience levels of the 10 board-certified radiologists in the pivotal study, and did the device's impact on detection accuracy and reading time differ across these experience levels?; The device provides nodule characteristics such as maximum axial plane longest diameter, mean diameter, and volume; how accurate are these quantitative measurements compared to manual measurements or pathological findings?; Given the SaaS model, what are the data security, privacy (HIPAA compliance), and interoperability protocols implemented, especially when dedicated servers are located at hospitals and connected to hospital networks?