AI Ensemble Model Detects Infant Skull Fractures on X-Ray

An ensemble deep learning model combining AP and lateral skull X-rays detected infant skull fractures with an AUC of 0.938 and accuracy of 91.6% on external validation, potentially reducing unnecessary CT scans.

• Retrospective study of 1,184 patients (4,298 internal images, 460 external validation images).
• DenseNet-121 with CLAHE achieved AUC of 0.926 (AP) and 0.911 (lateral); ensemble model improved to AUC 0.938 and 91.6% accuracy externally.
• Expert annotation by neurosurgeons and radiologists helped differentiate fractures from cranial sutures, reducing false positives.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.