Every story in this issue with a short summary, so you can decide what to read before leaving the site. Summaries are written by AI (Claude) only from each publisher's own description and checked by Jev for anything not in that description. They are not full translations. Follow "Read the original" for the full story.
The study, published in BMC Biomedical Engineering, introduces a transfer learning architecture. It integrates custom optimization algorithms and regularization methods to improve stability.
Read the original →A systematic review of 35 studies found that machine learning models analyzing structured clinical, imaging and demographic data reach diagnostic accuracy of up to 98.3 percent for glaucoma.
Read the original →The article explores how AI fracture detection identifies missed radiographic fractures, reduces malpractice claims and limits institutional financial liability.
Read the original →A new machine learning framework uses routine clinical variables to predict medium-term type 2 diabetes risk. It applies to patients with histologically confirmed NAFLD.
Read the original →Machine learning prediction of procalcitonin positivity uses inexpensive routine laboratory tests. It may reduce procalcitonin test volume and support diagnostic stewardship.
Read the original →A new Chinese Medical Journal review says AI and continuous glucose monitoring (CGM) are reshaping diabetes care through glucose prediction, intelligent insulin delivery, personalized nutrition and data-driven phenotyping. It describes a shift beyond traditional blood sugar tracking toward proactive, individualized management. The review notes ongoing challenges in cost, privacy and clinical integration.
Read the original →A scientific review published in BMC Artificial Intelligence evaluates how artificial intelligence, robotics and natural language processing advance telemedicine and remote patient monitoring.
Read the original →Singtel Singapore's chief executive Ng Tian Chong, NUHS chief executive Professor Yeoh Khay Guan and Ericsson's Daniel Ode signed a Memorandum of Understanding. It aims to advance connected healthcare, 5G-Advanced and 6G technology, and connected intelligence.
Read the original →The government-backed program is anchored by wearable sensors, AI-powered eye scans and cough-analysis apps. It targets earlier detection of cardiovascular disease, dementia, traumatic brain injury and respiratory infection.
Read the original →Vienna-based QUANTRO Therapeutics, which develops medicines targeting transcription factors, has secured a 12-month non-dilutive research grant from the Austrian Research Promotion Agency (FFG). The grant is to accelerate development of its AI-powered medicinal chemistry platform, ASCEND.
Read the original →AI-guided agents match responders and verify predictions with organoid tests. The approach is meant to sharpen cancer research in Korea.
Read the original →A study of 163 FDA-cleared AI medical devices found that one-third received algorithm updates. Of the updated devices, 85.7% reported no clinical or organizational benefit data.
Read the original →The article explores how AI and advanced imaging are improving disease detection, diagnostic accuracy and workflow efficiency in modern radiology.
Read the original →According to federal prosecutors in New York, three men used AI to imitate company executives' voices and sometimes faces. They also impersonated lawyers at a major law firm and had accounting staff send money for nonexistent deals. A Missouri hospital group alone lost $4.8 million.
Read the original →AI is now being used in attacks that once required specific skills, such as finding vulnerabilities and breaking in. Defenders must protect a growing number of systems with limited staff and budget, and an annual external vulnerability assessment struggles to keep up with faster attacks. The article also says that leaving defense entirely to AI has a pitfall that is easily overlooked.
Read the original →