Lunit INSIGHT DBT (V1.2)

Lunit INSIGHT DBT (V1.2) is computer-assisted detection and diagnosis (CADe/x) software that analyzes digital breast tomosynthesis (DBT) images using artificial intelligence to help radiologists identify and characterize breast cancer lesions. The software marks suspected lesions with an abnormality score indicating malignancy likelihood and is intended for use by interpreting physicians as an aid, not a replacement for clinical judgment.

The subject device maintains the same core AI/machine learning technology as the predicate (v1.1) but adds expanded compatibility with Siemens and Fujifilm DBT systems (in addition to Hologic and GE Healthcare), introduces an ordinal Case Abnormality Level output with pre-defined likelihood categories, provides user-selectable threshold operating points for different sensitivity levels, includes Current-Prior Comparison for interval change review, and optionally integrates volumetric breast density information.

IEC 62304:2006/A1:2016 (Medical device software – software life-cycle processes) and IEC 62366-1:2015+AMD1:2020 (Medical devices – Application of usability engineering to medical devices).

Lunit INSIGHT DBT v1.2 is substantially equivalent to the predicate device (v1.1, K242652) because both share identical indications for use, regulatory classification (Class II), and fundamental AI/machine learning technology for DBT breast cancer detection. The new features (expanded system compatibility, ordinal output, threshold options, comparative analysis) do not alter the device's intended use or raise new safety/effectiveness concerns. Performance testing on 3,277 DBT exams demonstrated ROC AUC of 0.9388 (95% CI: 0.9304–0.9472), exceeding the primary endpoint threshold of 0.903, confirming safety and effectiveness equivalent to the predicate.