Articles related to MACHINE LEARNING

A supervised random forest classifier trained on 330 features derived from nocturnal polysomnography achieved areas under the ROC curve approaching or exceeding 0.8 for distinguishing narcolepsy type 2, idiopathic hypersomnia, and controls in a national French reference database.

The study introduced an ML-based, tunable risk stratification strategy tailored for AML patients treated with venetoclax plus azacitidine. By addressing real-world data limitations and incorporating diverse diagnostic features, the approach enables more accurate and adaptable prognostic modeling than traditional systems, supporting personalized treatment decisions and broader clinical applicability.

The method detected 90% of cancer cases, a result very close to the 94% detection rate achieved by colonoscopies and better than all current non-invasive detection methods.

EchoNext AI achieves diagnostic accuracy surpassing human experts for structural heart disease, demonstrating a new frontier for cardiac disease AI applications.

AI-assisted diagnosis of lung cancer showed 87% sensitivity and 87% specificity in a meta-analysis, suggesting potential for improved accuracy in interpreting chest CT scans.

The TRIPOD+AI guidelines mark a significant update in the standards for reporting prediction models, incorporating advancements in AI to support rigorous, transparent research in healthcare settings, facilitating better clinical decision-making tools.