RADAR, a generalist AI trained on over 400,000 abdominal CTs, achieves expert-level diagnosis across 18 structures and 146 findings, with high accuracy and robustness.
- 1 on Bluesky
- 1 opens in JClub
Artificial Intelligence in Healthcare and Education
Moving in Nature, Bioinformatics, Cancer Discovery, Cell, JAMA, Nature Medicine, Science.
- Moved this week
- 8
Rebuilt
Moved this week
AI model Oncoformer accurately predicts cancer development, diagnoses tumors, and stages cancer using health records and imaging data across diverse populations.
- In the news
MIRA, an autonomous AI agent in a sandboxed EHR, surpasses physicians in diagnostic accuracy and makes safe, guideline-concordant clinical decisions across multiple cases.
- Open access
- 29 cites
Large language models like GPT-5.2, Gemini 3.1 Pro, and Claude Opus 4.6 outperform specialized clinical AI tools across multiple medical benchmarks, including real-world queries.
- Open access
- 32 cites
Agentomics autonomously develops state-of-the-art biomedical machine learning models, outperforming existing agentic systems and matching human expertise on 11 of 20 datasets.
- Open access
- 2 cites
A new LLM-based system, AMIE, matches primary care physicians in disease management reasoning and surpasses them in treatment precision and guideline adherence in multi-visit scenarios.
- 1 opens in JClub
Reliability failures in AI for healthcare include erroneous outputs, population disparities, and performance decline, threatening safe deployment across 20+ failure modes.
- Open access
Latest in Artificial Intelligence in Healthcare and Education
Newest first · last 60 days
Synthetic data generation could revolutionize gastrointestinal medicine by enabling more accurate AI-driven diagnosis and treatment with privacy-preserving data.
- 1 opens in JClub
The Multimodal Anonymizer achieves near-complete deidentification of multimodal hospital data with over 98% sensitivity and preserves critical clinical information at rates above 99%.
- Open access
A ReAct agentic AI system achieves 93.4% accuracy in natural language querying and analysis of TCGA clinical data, surpassing rule-based and LLM baselines.
- 1 opens in JClub
- Open access
Physicians using conversational AI in Latin America achieved significantly higher validity scores (median 2.83 vs. 2.46) than usual search methods in clinical case responses.
- Open access
42.7% of surveyed obstetricians and gynecologists in France reported using AI in routine practice, mainly for administrative and informational tasks.
- Open access
General-purpose multimodal large language models in histopathology exhibit errors and hallucinations in over 80% of outputs, impacting diagnostic safety and reliability.
- Open access
Algorithm D achieved the highest AUROC of 0.849 and sensitivity across all AI-positive rates in a large US screening cohort, outperforming other commercial and open-source models.
RoB2-Evaluator, a rule-constrained GPT-5.2 system, achieved 90% agreement with human reviewers and 94.3% stability in Cochrane Risk of Bias 2 assessments across 10 RCTs.
- Open access
Implementing EHR-based risk models with imaging and blood tests detects pancreatic cancer early in 8.2% of high-risk adults aged 50-84.
- Open access
Large language models can serve as human proxies in various scientific and applied contexts, each role requiring distinct validity criteria.
AI is revolutionizing biomedical research by enabling integration of diverse biological data and addressing previously intractable questions.
RADAR, a generalist AI trained on over 400,000 abdominal CTs, achieves expert-level diagnosis across 18 structures and 146 findings, with high accuracy and robustness.
- 1 on Bluesky
- 1 opens in JClub
Reliability failures in AI for healthcare include erroneous outputs, population disparities, and performance decline, threatening safe deployment across 20+ failure modes.
- Open access
Agentic AI systems like Co-Scientist, Robin, and Biomni can autonomously generate hypotheses, design experiments, analyze data, and refine reasoning, advancing scientific discovery.
Single-barrier reforms in algorithmic recourse systems achieve less than 0.02% improvement, with cross-layer interactions accounting for 87.6% of potential gains in digital health contexts.
- Open access
Multi-billion parameter AI models still struggle with surgical tool detection in neurosurgery, showing limited improvements despite extensive scaling efforts.
- Open access
Follow this area in the app
Browse JClub
Journals this month
- bioRxiv5569
- medRxiv1738
- Environ Res483
- PNAS463
- Sci Adv439
- JAMA255
- Cell Rep210
- BMJ199
- Nat Commun177
- Environ Pollution170
- Lancet149
- Science138
- NEJM119
- Applied & Environ Microbio116
- J Exp Bot115
- Curr Biol112
- mBio109
- Nature104
Moving areas, week to 3 Oct 2026
- Single-cell and spatial transcriptomics10
- Artificial Intelligence in Healthcare and Education8
- CAR-T cell therapy research7
- Neuroinflammation and Neurodegeneration Mechanisms6
- Lung Cancer Treatments and Mutations5
- Diabetes Treatment and Management4
- Pancreatic and Hepatic Oncology Research4
- Photosynthetic Processes and Mechanisms4
- Genomics and Chromatin Dynamics4
- Gut microbiota and health4