Search

Prediction of overall survival in acute myeloid leukemia using supervised machine learning - The Capital Region of Denmark's Research Portal

ML models integrating clinical, cytogenetic, and genetic data improve AML survival prediction versus ELN models in real-world patients; benefit varies by treatment intensity, with strongest OS performance in non-intensively treated patients when using full feature sets.

Emerging applications of machine learning and AI for predictive modeling in precision medicine | npj Digital Medicine

The collection argues that medical AI is maturing into a clinical science, emphasizing interpretable, domain-informed models, portable video diagnostics, lightweight signal devices, multimodal wearable data, and equity at scale as central goals.

Machine Learning Predicts Diabetes Risk in NAFLD - EMJ

A study builds a five-factor machine learning model to predict medium-term diabetes risk in NAFLD patients, validated externally with good discrimination but limited by sample scope; authors call for further multicenter validation before clinical use.

AI predicts cancer survival from one face photo, study finds

A study from Mass General Brigham shows an AI model named FAHR-Face can estimate cancer survival from a single face photo, improving prediction beyond age. The model uses separate components for age and survival, validated across two clinical sites, with noted limitations and calls for further trials before clinical use.

Machine learning models predict common cancer risks using routine health data

Translated title. Original: Les modèles d’apprentissage automatique prédisent les risques courants de cancer à l’aide de données de santé de routine

ML models using routine health data can predict risk for multiple common cancers, showing overall predictive power and variable accuracy by cancer type; the approach relies mainly on blood tests and medical history, with limited gain from genetics or metabolomics, and includes a tool (PRE-Screen) to standardize model use by AI agents.

Artificial intelligence detects the most severe type of heart attack and locates the blocked artery with high precision - Agência SP, the São Paulo news agency

Translated title. Original: Inteligência artificial detecta tipo mais grave de infarto e localiza artéria bloqueada com alta precisão - Agência SP, a agência de notícias de São Paulo

An international collaboration trains an AI to detect acute coronary occlusion from 12-lead ECGs and locate the blocked artery, aiming to speed emergency decisions and catheterization when needed. The model, validated in Brazil and with European and UFMG datasets, is the first of its kind demonstrated in a Brazilian population.

AI allows predicting headaches 24 hours in advance — RexMolón - Science in Action

Translated title. Original: La IA permite anticipar dolores de cabeza con 24 horas de antelación — RexMolón - La Ciencia en Acción

AI can identify early signals that may predict headaches up to a day in advance, based on data from people with persistent post-concussion headaches. This Norwegian study, while related to migraine prediction research, targets a specific condition and should be read alongside the original paper in the Journal of Head Trauma Rehabilitation (Danelakis et al., 2026).

GitHub - Kolawole-Toheeb/diabetes-readmission-prediction · GitHub

This repository trains a machine-learning model to estimate 30-day hospital readmission risk for patients with diabetes and serves predictions through a FastAPI web service. Its best model, XGBoost, achieved a test AUC of 0.653.

Clinically guided machine learning stratifies dental implant risk in

A study presents TBLPSO, a clinically guided, subgroup-aware machine-learning framework for predicting dental implant risk. It uses 55 features to classify patients into four risk groups, achieving 96.1% accuracy with subgroup-specific and some shared features, validated by five-fold temporal cross-validation and multiple baselines; further work aims at clinical decision-support applications.

Artificial intelligence can predict the risk of death in cancer patients from a facial image – News – German Medical Journal

Translated title. Original: Künstliche Intelligenz kann Sterberisiko von Krebspatienten anhand Gesichtsfoto vorhersagen – News – Deutsches Ärzteblatt

A Harvard Medical School team trains FAHR-Face to predict cancer patient mortality from facial images, using 40 million photos and yielding two specialized models; results appear in The Lancet Digital Health (2026).