Critical-care AI is moving beyond isolated prediction toward adaptive treatment support, interpretable physiologic surveillance, and workflow-integrated action. The decisive test is now whether these systems improve care, not…

Editing an AI-drafted patient message is not a uniform task: interpreting a radiology report carries a steep per-message time cost, while routine scheduling changes create the largest cumulative…

Orthopedics AI is moving beyond isolated classifiers toward reusable imaging models, accessible measurement, risk phenotyping, and decision support across the surgical pathway. Prospective clinical impact is now the…

Locked artificial intelligence models quantified tumour-infiltrating lymphocytes reproducibly across multiple trials and institutions, and their scores independently predicted outcomes in early-stage triple-negative breast cancer. However, they provided no…

MORGOTH, a foundation model trained to interpret standard clinical EEGs, achieved expert-level performance across seven event-level tasks and 17 recording-level classifications. Its accuracy, calibration, and robustness largely persisted…

Incidental irradiation of the thymus during chemoradiotherapy for locally advanced non-small-cell lung cancer was independently associated with higher risks of distant metastasis and death. The findings identify thymic…

Across 4,262 tumours from the APHINITY trial, greater immune infiltration consistently signalled better outcomes in early HER2-positive breast cancer. AI-derived immune hotspots added information beyond lymphocyte quantity alone.

Recent Oncology coverage shows AI moving beyond lesion recognition toward transferable imaging models, tissue-derived stratification, therapy-outcome prediction, and screening systems designed around real clinical pathways.

NREM sleep spindles appear to play a selective role in consolidating social memory. Across mouse experiments and pediatric EEG studies, their disruption was associated with autism-related social-memory deficits.

DentVLM, a bilingual vision-language model trained across seven dental imaging modalities, outperformed leading AI comparators and improved clinicians’ diagnostic accuracy while shortening reading time.