A systematic review of 11 orthopaedic randomized controlled trials involving artificial intelligence reveals critical reporting gaps in data transparency and code accessibility, despite average 81% compliance with CONSORT-AI…

A new study uses deep learning to detect and classify tertiary lymphoid structures—immune cell clusters in tumors—revealing their maturation state predicts patient survival and treatment response across multiple…

A deep learning model called GTBIS stratifies combined SCLC-LCNEC lung cancers based on histopathology, revealing dramatically different survival outcomes.

Automated evaluation metrics can effectively distinguish good from poor AI responses to patient hospitalization questions, providing a scalable alternative to human expert review—though careful reference selection is essential.

A multimodal deep learning model successfully predicts rapid kidney function decline after nephrectomy for renal cancer, potentially guiding surgical decisions in technically challenging cases.

Combining natural language processing analysis of children's essays with conventional risk factors substantially improves early identification of potential mental health disorders, according to a prospective birth cohort study.

Medical AI models exploit unintended shortcuts in training data—subtle biases that work in the lab but fail in clinical practice. G-AUDIT automatically identifies these hidden risks across any…

Pathog-PDx, a machine learning system, diagnoses pediatric respiratory infections within hours of admission—significantly faster than conventional testing and reducing unnecessary antibiotics.

Recent Geriatrics & Aging coverage reveals convergence: aging is measurable across independent modalities—radiographic structure, neural function, microbial ecology. Biomarkers predict mortality and disease before symptoms appear, enabling early…

Early detection transforms fire response. This wireless deep-ultraviolet sensor system detects flames at ignition, before smoke appears.