# Artificial Intelligence in Radiology — GCMD Library living collection

Updated: n/a · 4 episodes · 15 cited statements

## Episodes
### Clinical Implementation
- [Are the Pilots Onboard? Equipping Radiologists for Clinical Implementation of AI.](https://library.globalcastmd.com/article/are-the-pilots-onboard-equipping-radiologists-for-clinical-implementation-of-ai-17062) — article · [machine version](https://library.globalcastmd.com/article/are-the-pilots-onboard-equipping-radiologists-for-clinical-implementation-of-ai-17062.md)

### AI-Assisted Diagnosis
- [Artificial Intelligence Algorithm Improves Radiologist Performance in Skeletal Age Assessment: A Prospective Multicenter Randomized Controlled Trial](https://library.globalcastmd.com/article/artificial-intelligence-algorithm-improves-radiologist-performance-in-skeletal-a-17063) — article · [machine version](https://library.globalcastmd.com/article/artificial-intelligence-algorithm-improves-radiologist-performance-in-skeletal-a-17063.md)
- [Automatic Detection of Inadequate Pediatric Lateral Neck Radiographs of the Airway and Soft Tissues using Deep Learning](https://library.globalcastmd.com/article/automatic-detection-of-inadequate-pediatric-lateral-neck-radiographs-of-the-airw-17064) — article · [machine version](https://library.globalcastmd.com/article/automatic-detection-of-inadequate-pediatric-lateral-neck-radiographs-of-the-airw-17064.md)

### Pediatric Datasets & Representation
- [Underrepresentation of children in public medical imaging datasets](https://library.globalcastmd.com/article/underrepresentation-of-children-in-public-medical-imaging-datasets-17056) — article · [machine version](https://library.globalcastmd.com/article/underrepresentation-of-children-in-public-medical-imaging-datasets-17056.md)

## Statements
Every statement is attributed, typed, and timestamped; the link is the citation.
- Measures must be taken to educate radiologists on AI tool performance and failure modes. (recommendation) (no timestamp yet)
- Radiology systems should be designed to avoid automation bias and the potential decline in radiologist performance. (recommendation) (no timestamp yet)
- The mean absolute difference between radiologist skeletal age assessment and a reference panel was 5.36 months when using the AI algorithm compared with 5.95 months without AI assistance (P = .04). (research) (no timestamp yet)
- Hand radiographs were interpreted by 93 radiologists from six centers in this prospective randomized controlled trial. (research) (no timestamp yet)
- A total of 792 examinations were performed with the AI algorithm and 739 without the AI algorithm. (research) (no timestamp yet)
- When using the AI algorithm, radiologists were shown the AI interpretation as part of routine clinical work and were permitted to accept or modify it. (research) (no timestamp yet)
- The primary efficacy outcome was the mean absolute difference between the skeletal age dictated into the radiologists' signed report and the average interpretation of a panel of four radiologists not using a diagnostic aid. (research) (no timestamp yet)
- Interpretation time was measured as a secondary outcome in the study. (research) (no timestamp yet)
- A linear mixed-effects regression model with random center- and radiologist-level effects was used to compare the two experimental groups. (research) (no timestamp yet)
- A total of 1200 lateral airway radiographs obtained in emergency department patients between January 1, 2000, and July 1, 2019, were retrospectively queried and classified by radiologists as adequate or inadequate. (research) (no timestamp yet)
- The training set contained 961 radiographs and the test set contained 239 radiographs. (research) (no timestamp yet)
- The ResNet-50 deep learning classifier achieved sensitivity of 0.90 (95% CI: 0.86, 0.94), specificity of 0.82 (95% CI: 0.76, 0.90), and area under the receiver operating characteristic curve of 0.86 (95% CI: 0.81, 0.91) for detecting inadequate lateral airway radiographs. (research) (no timestamp yet)
- Interrater agreement for technologists in classifying radiograph quality was fair with Fleiss κ of 0.36 (95% CI: 0.29, 0.43). (research) (no timestamp yet)
- Interrater agreement for radiologists in classifying radiograph quality was moderate with Fleiss κ of 0.59 (95% CI: 0.52, 0.66). (research) (no timestamp yet)
- Cohen κ comparing ResNet-50 consensus rating to ground truth was 0.76 (95% CI: 0.63, 0.89), compared to 0.49 (95% CI: 0.37, 0.61) for technologists and 0.66 (95% CI: 0.54, 0.78) for radiologists. (research) (no timestamp yet)

---
Educational content from recorded physician discussions — not medical advice. Cite the canonical URL or the ?t= deep link. Policy: https://library.globalcastmd.com/ai
