A deep-learning AI model accurately estimates bone mineral density using pediatric chest x-rays, showing potential for opportunistic bone health screening.
Key Details
- 1Developed at Seoul National University using 1,464 chest x-rays paired with DEXA scans from 1,188 pediatric patients (median age 13).
- 2Model trained on ResNet-50 CNN predicts BMD Z scores using x-rays and clinical variables.
- 3Internal dataset: AUC 0.92, sensitivity 60%, specificity 95% for BMD Z score prediction.
- 4External test set: AUC 0.90, sensitivity 82%, specificity 85% for BMD Z score prediction.
- 5For low BMD (Z ≤−2.0): AUC 0.92 (internal), 0.90 (external), both with 82% sensitivity and 85% specificity.
- 6Model interpretability showed focus on spine regions relevant to osteopenia, though study was retrospective and clinical outcomes were not evaluated.
Why It Matters
This work suggests AI could broaden access to pediatric bone health screening without dedicated DEXA scans, potentially enabling earlier detection and treatment of bone deficits in at-risk children. Rigorous future validation and clinical studies are still needed for practice integration.
Source
AuntMinnie
Related News
•AuntMinnie
Radiologists Struggle to Spot AI-Generated Radiology Images
Radiologists correctly identify AI-generated images 75% of the time, with CT and MRI images being particularly challenging to spot.
•AuntMinnie
Radiology Leads FDA AI Device Approvals Over Three Decades
Radiology accounts for 76% of all FDA-cleared AI/ML-enabled medical devices as of the end of 2025.
•Radiology Business
Automation Bias: How AI Can Compromise Radiologist Accuracy
AI decision support can induce automation bias, leading radiologists to accept incorrect interpretations and potentially reduce their diagnostic accuracy.