Bayesian Beta-Binomial Model Improves MRMC Study Analysis Under Overdispersion

Bayesian Beta-Binomial model showed improved robustness and lower estimation error vs classical linear models in MRMC studies with overdispersion and small sample sizes, applied to chest X-ray bone suppression imaging.

• Simulation study and real-world chest X-ray MRMC evaluation; sample size not reported in source.
• Model explicitly handles class imbalance and incorporates prior information for regularization.
• Limited to one clinical application; generalizability to other imaging tasks not yet established.

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