Automated bone age assessment in a German pediatric cohort: agreement between an artificial intelligence software and the manual Greulich and Pyle method - Summary - MDSpire
To validate the agreement between expert radiologist readers and the IB Lab PANDA AI software for bone age assessment in a German pediatric cohort.
Key Findings:
The AI software showed a strong correlation with expert assessments, indicating its reliability.
Mean absolute difference and root mean square error indicated high accuracy of the AI model.
Intraclass correlation coefficients demonstrated good agreement between AI and expert readers.
Interpretation:
The AI software provides a reliable alternative to manual bone age assessment, potentially reducing observer variability.
Limitations:
The study was retrospective and limited to a specific population, which may affect generalizability.
No prior clinical validation of the AI software was available, raising concerns about its application in diverse settings.
Conclusion:
The automated AI assessment of bone age is comparable to the manual Greulich and Pyle method, suggesting its utility in clinical practice and potential to enhance diagnostic consistency.
