Tag
Highlight This unprecedented large-scale multinational study developed a semisupervised contrastive machine learning framework that generates a global
A machine learning model using electronic health records effectively identifies patients at high risk of atrial fibrillation, enabling targeted screening and early diagnosis to prevent stroke.
Machine learning is transforming medical devices but faces challenges like shortcut learning and automation bias. Edge AI and evolving regulation aim to balance innovation with safety.
A machine-learning model using perioperative hemodynamic indices accurately predicted stage 2-3 acute kidney injury after orthotopic heart transplantation in 114 patients, with light gradient boosting machine achieving AUC 0.898.
This study introduces and validates a simplified machine learning-based equation for estimating LDL cholesterol, showing comparable accuracy to the original Martin-Hopkins method and improved ease of use.
Recent research demonstrates that machine learning algorithms can accurately predict patient-reported outcomes one year after breast reconstruction, potentially enhancing shared decision-making and individualized care planning.
The HLH-Risk-Calculator utilizes machine learning to predict initial disease severity and mortality at multiple time points in adult patients with secondary hemophagocytic lymphohistiocytosis, highlighting key biomarkers and advocating for
Machine learning analysis of brief patient-physician conversations using acoustic speech features can effectively screen for cognitive impairment in primary care settings, enabling scalable early detection.
A machine learning algorithm can generate synthetic CT images from temporal bone MRI, enabling visualization of bone and soft tissue simultaneously without radiation, useful for otologic surgical planning and localization.
Unsupervised machine learning identified three distinct preeclampsia phenotypes linked to timing of delivery, placental dysfunction, fetal growth, and complications, suggesting a path toward more personalized risk assessment and management.
Machine learning identified two distinct bilateral hepatocellular carcinoma phenotypes with different tumor burdens, surgical approaches, and outcomes, suggesting preoperative risk stratification may improve patient selection and treatment
A novel machine-learning model incorporating clinical, ultrasonographic, and molecular features predicts occult central lymph node metastasis in papillary thyroid cancer, with RET fusion positivity identified as a key risk factor.
A five-RNA extracellular vesicle signature showed strong diagnostic performance for chronic pancreatitis and may also reflect acinar-to-ductal metaplasia and disease-related metabolic features.
Secondary analysis of the MYRIAD trial reveals that while machine learning models can identify adolescents who may benefit from school-based mindfulness training, the clinical relevance of such predictions remains minimal. Both causal fores
A landmark multicohort study of 1,168 individuals reveals that gut microbiome diversity systematically declines across liver disease stages, with specific bacterial markers like Veillonella and functional metabolic shifts predicting disease
Researchers developed and validated ADAPT-CEC, an artificial intelligence algorithm that can adjudicate cardiovascular events across different trial definitions. A hybrid approach combining AI with selective human review achieved 95.6% accu
This review analyzes a landmark machine learning study identifying four distinct GDM phenotypic clusters, revealing that early-diagnosed, comorbidity-related GDM carries a four-fold higher risk of postpartum diabetes.
A groundbreaking machine learning study of over 37,000 patients identifies four distinct GDM phenotypic clusters. These findings demonstrate that early-diagnosed, high-comorbidity phenotypes carry a four-fold increased risk of postpartum di
A large-scale analysis of the MYRIAD trial indicates that universal school-based mindfulness training does not superiorly reduce depression risk in adolescents. While machine learning identified certain potential responders, the clinical be
This study reveals that combining stress hyperglycemia ratio and glucose variability significantly improves mortality prediction in sepsis, with risk patterns varying significantly based on a patient’s baseline metabolic status (Norma
