Ranks distinct AI gain and problem claims from the published record. Scores reward impact, independent source strength, scale, confidence, and recency.
Researchers retrospectively analyzed 134 acute pulmonary embolism cases from April 2023 to March 2024, using commercial AI software to automatically measure a new biomarker, pulmonary artery-to-vein volume difference, alongside traditional CTPA parameters to stratify high/intermediate-high versus lower risk.
A study tested YOLO-based AI models for breast lesion detection on digital breast tomosynthesis using a 94-patient Western database from the Cancer Imaging Archive and a 157-patient Eastern database from a single medical center, with lesions grouped into six types.
In a retrospective study of 340 pancreatic ductal adenocarcinoma patients, investigators developed machine learning models combining CT radiomics and clinical predictors to predict synchronous liver metastasis preoperatively. In an independent validation cohort of 102 patients, the best linear model (LDA) reached AUC 0.828 and the nonlinear model (MLP) reached AUC 0.822, both showing good calibration with Hosmer-Lemeshow P values of 0.551 and 0.682.
Researchers systematically reviewed 41 studies through January 2025 that used deep learning to generate synthetic postcontrast T1-weighted MRI from precontrast images alone, aiming to reduce gadolinium use. Most work was in neuroimaging, using GANs and CNNs, and a targeted meta-analysis of 15 brain tumor studies reported high whole-image similarity metrics.
Researchers developed and validated an interpretable machine learning model to predict 5-year all-cause mortality in non-dialysis chronic kidney disease using data from 1,858 patients in the KNOW-CKD prospective cohort, with 94 deaths observed. The CatBoost model achieved AUC 0.813 versus 0.747 for logistic regression, and a simplified version using age, eGFR, albumin, urine protein-to-creatinine ratio, and total calcium retained AUC 0.795 in an external cohort of 348 patients.
The peer-reviewed article examines responsibility gaps when AI tools in healthcare cause patient harm. It notes that traditional models struggle because decisions are spread across clinicians, developers, institutions and the AI systems themselves.
Researchers analyzed Reddit posts and comments about mental health conversations with ChatGPT to understand how large language models are being used for support outside clinical settings. By October 2025, they found users described ChatGPT as accessible and non-judgmental, providing emotional support, validation, and practical help like navigating difficult conversations and preparing for therapy.
A peer-reviewed study published October 16, 2025 examined user reviews of the Replika companion chatbot to investigate reports of inappropriate sexual behavior. From 35,105 negative Google Play Store reviews, researchers identified 800 cases for thematic analysis.
ace