Saige-Dx (3.1.0)

Saige-Dx is an artificial intelligence-powered software designed to serve as a concurrent reading aid for interpreting physicians, assisting in the detection of breast cancer indicators on digital breast tomosynthesis mammograms. It employs AI to analyze DBT image stacks and associated 2D images, assigning a "Suspicion Level" to identified soft tissue lesions and calcifications to enhance diagnostic efficiency.

DeepHealth, Inc. has obtained a determination of substantial equivalence for its Saige-Dx (3.1.0) software, classifying it as Radiological Computer Assisted Detection And Diagnosis Software. This regulatory milestone, achieved after a submission on November 27, 2024, and determination on December 19, 2024, enables the company to market the device, subject to standard regulatory controls. Saige-Dx functions as an AI-powered software device designed to analyze digital breast tomosynthesis (DBT) mammograms, identifying the presence or absence of soft tissue lesions and calcifications indicative of cancer. It processes both DBT image stacks and associated 2D images, assigning a "Suspicion Level" for individual findings and the overall case. Intended as a concurrent reading aid for interpreting physicians during screening mammograms for women aged 35 or older, this version was found substantially equivalent to its predecessor, Saige-Dx v.3.0.0, with updates primarily enhancing cloud deployment, speed, and memory usage without compromising safety or effectiveness.

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 performance metrics (e.g., sensitivity, specificity, false positive rate) did Saige-Dx (3.1.0) demonstrate in the substantial equivalence submission?; How does the 'Suspicion Level' assigned by Saige-Dx correlate with the BI-RADS assessment categories, and what clinical utility does this provide for interpreting physicians?; Are there any specific limitations or patient populations for which Saige-Dx's performance has not been adequately evaluated, despite its intended use for women aged 35 or older?; What, if any, post-market surveillance or clinical validation studies are planned or required for Saige-Dx (3.1.0) following its substantial equivalence determination?; How does the cloud deployment of Saige-Dx v.3.1.0 impact data security, system uptime, and accessibility for clinical users compared to previous versions?