A digital health program significantly increased lung cancer screening rates among high-risk individuals, underscoring digital tools' potential to improve preventive care.
Artificial intelligence is reshaping health care, offering vast potential benefits alongside significant challenges. The JAMA Summit emphasizes the need for robust evaluation, equitable deployment, and multi-stakeholder collaboration to ensure AI improves outcomes safely and fairly.
AI demonstrates high accuracy in identifying mpox and other anogenital conditions from clinical images, yet significant research gaps and validation needs remain before clinical adoption.
A cross-sectional study identifies four distinct profiles of nonverbal learning disability (NVLD) in children, revealing significant heterogeneity in visual-spatial deficits, academic skills, and psychiatric diagnoses, highlighting implications for diagnosis and tailored interventions.
This study assesses the educational quality and popularity of Spanish-language medical microvideos on TikTok in Latin America, revealing a discordance between engagement metrics and content quality, emphasizing the need for integrated evaluation tools.
A recent study assessed the quality, reliability, and popularity of anemia-related videos on YouTube, revealing moderate variability and highlighting the importance of health professionals' involvement to ensure accurate online content.
A comprehensive analysis of 903 FDA-approved AI medical devices reveals limited clinical performance data and demographic inclusivity, underscoring the need for ongoing evaluation to ensure safe, effective clinical application.
Integrated radiomic-clinical machine learning models outperform traditional clinical biomarkers in predicting survival and treatment response in unresectable hepatocellular carcinoma patients undergoing atezolizumab plus bevacizumab therapy.
AIRE-CHB, a novel AI-enhanced ECG model, significantly improves prediction of incident complete heart block, outperforming traditional bifascicular block assessment and offering promising clinical utility in risk stratification and management.
AI-assisted digital pathology improves inter-pathologist agreement on fibrosis staging in metabolic dysfunction-associated steatohepatitis (MASH), enhancing clinical trial accuracy and efficiency.
A new AI model, Delphi-2M, predicts the risk of over 1,000 diseases up to 20 years in the future with remarkable accuracy, transforming disease prevention and personalized health management.
A multicenter trial demonstrated that integrating AI-based computer-aided detection with water exchange colonoscopy significantly increases adenomas per colonoscopy without prolonging procedure time or increasing non-neoplastic lesion detection.
This review evaluates AI agent systems built on large language models in healthcare, demonstrating improved clinical task accuracy over base models, especially in complex scenarios. Multi-agent architectures show promise but require further real-world validation.
An RCT evaluated an AI-assisted framework for electrodiagnostic report interpretation, finding no significant improvement over standard physician reporting but highlighting potential workflow benefits for routine cases.
The AID-ME trial demonstrates that an AI-powered clinical decision support system significantly improves remission rates and accelerates symptom improvement in moderate to severe major depressive disorder.
A study assessing the AI-based platform Diagnocat reveals high sensitivity but moderate specificity in detecting periapical radiolucencies on CBCT scans of non-root-filled molars, with decreased accuracy post-root treatment, underscoring the need for clinician oversight.
A randomized controlled trial demonstrates that integrating deep language learning model (DLM) simulations into surgical training significantly improves history-taking skills and communication confidence among senior medical students.
A randomized controlled trial demonstrates that AI assistance moderately improves dentists' accuracy in diagnosing periapical radiolucencies, mainly by reducing false positives, with significant benefits for junior clinicians and a shift towards conservative treatment decisions.
Machine learning models trained on clinical data improve prediction of coronary artery disease on CT scans compared to traditional risk scores, potentially guiding better resource utilization in cardiovascular care.