MediAI-BA

MediAI-BA is an AI-based software that analyzes 2D posterior-anterior X-ray images of the left hand and wrist to automatically assess and quantify bone age in pediatric patients aged 2 to 18 years. It is intended to aid pediatric radiologists in bone age assessment but should not be relied upon alone for diagnostic decisions. The software uses deep learning to extract and synthesize global and local skeletal maturity features to determine bone age.

MediAI-BA and its predicate device (EFAI BAPXR, K234042) are both Class II automated radiological image processing software using deep learning algorithms for bone age assessment on left-hand X-rays in healthcare settings. Key difference: MediAI-BA extends the age range to 18 years (predicate covers 2-16 years) and includes report creation capability, whereas the predicate does not. Both employ AI algorithms, comply with DICOM standards, and produce quantitative bone age output in years.

Not stated in this summary. The document does not reference specific ISO, IEC, ASTM, or other consensus standards.

MediAI-BA is substantially equivalent because it shares the same intended use, same regulatory classification (Class II, 21 CFR 892.2050, Product Code QIH), same technological approach (AI-based deep learning for bone age assessment), and same anatomical target and image modality as the predicate. Performance testing on 600 independent cases demonstrated equivalence to human radiologist assessments with a mean difference of 0.026 years and 89% of cases within 0.5 years of ground truth, comparable to predicate performance. The device introduces no new risks and demonstrates equivalent safety and effectiveness.