PathoSage is a three-stage framework addressing evidence adjudication in pathology, explicitly separating knowledge retrieval, evidence collection, and evidence adjudication. It employs an agentic system with Structured Evidence Deliberation to independently evaluate heterogeneous evidence and reduce anchoring bias.
ChatCAD is an interactive computer-aided diagnosis system that leverages Large Language Models to analyze medical images. It aims to enhance the accuracy and efficiency of medical diagnosis through artificial intelligence.
This paper introduces MedicalBench, a new benchmark for evaluating Large Language Models in medical concept extraction from electronic health records. It focuses on implicit medical reasoning and evidence grounding, addressing the challenge of identifying concepts not explicitly stated.
The paper introduces HealthCraft, a public reinforcement-learning environment designed to evaluate the safety of frontier language models in emergency medicine. It focuses on trajectory-level safety, tool misuse, and clinical pressure, built on a FHIR R4 world state and offering 195 tasks for comprehensive assessment.
This paper introduces the Convolutional Variational Deep Embedding (Conv-VaDE) model for EEG microstate analysis. It enhances interpretability by jointly learning topographic reconstruction and probabilistic soft clustering, enabling generative decoding of cluster prototypes into verifiable scalp topographies.
This research introduces Med-Stress, a framework to test the epistemic resilience of LLMs in clinical dialogue, revealing that high diagnostic accuracy doesn't guarantee belief stability under escalating pressure. It proposes RBED and R-FT as novel defenses to mitigate this failure mode in medical AI.
The AI market promises significant transformation, with healthcare being a prime target due to financial pressures, labor shortages, and the increasing burden of caring for an aging population. AI developers are targeting a wide range of functions, from curing cancer to streamlining processes.
94% of healthcare executives view AI as critical for the future of medicine. Artificial intelligence assists doctors by scanning radiology images and identifying skin cancers or cancerous cells with speed and accuracy. It serves as a powerful second opinion, combining its speed with human judgment and patient context.
This research investigates the trustworthiness and fairness of nonparametric deep survival models for analyzing Alzheimer's Disease (AD) progression. It addresses the lack of studies considering learned bias in existing deep learning models for AD and proposes novel fairness metrics to ensure reliable predictions.
This paper introduces ClinicalBench, a 400-question benchmark designed to stress-test assertion-aware retrieval for cross-admission clinical QA on MIMIC-IV using real EHR notes. It also presents EpiKG, a patient knowledge graph system that incorporates assertion and temporality tags to route retrieval by question intent, demonstrating significant performance improvements across various LLMs.
DBGL introduces a novel Decay-Aware Bipartite Graph Learning method to address the challenges of irregular medical time series classification. It utilizes a patient-variable bipartite graph to model irregular sampling patterns and variable relationships, alongside a node-specific temporal decay encoding for variable decay irregularity.
The article shares insights from 18 months of building Helios Med, a medical AI with 383 specialist agents designed to assist with diagnostic reasoning. It aims to provide a thorough second opinion for doctors and patients, addressing the limitations of current healthcare practices and diagnostic errors.
This paper introduces OGCaReBench, a new retrieval-focused benchmark aimed at evaluating LLMs' ability to answer clinical questions that go beyond typical medical guidelines. It addresses the gap where most medical LLMs are trained on common, guideline-focused knowledge, while real-world care often involves rare cases not covered by guidelines.
Traj-Evolve is a self-evolving multi-agent system designed to model patient trajectories from electronic health records for early lung cancer detection. It utilizes an Experience Pool for retrieving similar cases and multi-agent reinforcement learning to optimize inter-agent collaboration.
GazeX is a new vision language model trained on radiologists' eye-tracking data and reasoning to improve chest X-ray interpretation. The model learns to emulate expert spatial and temporal attention, aiming to bridge the gap between model outputs and clinical diagnostic reasoning.
O estudo apresenta o EMSDialog, um novo conjunto de dados de 4.414 conversas sintéticas multi-falantes para serviços médicos de emergência, geradas a partir de relatórios reais de pacientes usando uma pipeline de agentes multi-LLM. Este dataset, anotado com diagnósticos e tópicos, demonstra melhorias na precisão e estabilidade da previsão de diagnóstico conversacional.
Despite accurate diagnostic systems from data-centric medical AI, widespread clinical adoption is limited due to insufficient attention to fair performance across diverse patient populations and poor workflow integration. This paper proposes a 'People-Centred Medical Image Analysis' approach to address these interconnected challenges, which prior work has typically examined in isolation.
This research evaluated the impact of prompting language on the diagnostic reasoning and accuracy of large language models (LLMs) in clinical settings. Four out of five models performed better in English, highlighting the uncertainty regarding LLM reliability across different languages.
This article proposes the creation of "Biomedical TransitionNet", a new type of dataset analogous to ImageNet, but focused on biological state transitions for the next generation of medical AI. It argues for the necessity of such infrastructure to build real-world models in biomedicine, moving beyond classification and prediction.
This content details the fine-tuning of a clinical AI model, MedQA, on the AMD ROCm platform. It highlights the ability to perform this task without requiring CUDA, offering a significant alternative for AI development.