Artificial Intelligence in Healthcare and Education

MedMCP: An Open Agentic Ecosystem for Medical Imaging Workflows

MedMCP provides an agentic framework for medical imaging workflows, enabling non-technical users to apply state-of-the-art image analysis through natural language and shareable workflows.

Facilitating Valid Statistical Inference in Biomedical Image Synthesis

This paper evaluates the statistical validity of synthetic biomedical images and introduces PAMS to improve biological inference.

Version-aware question answering over living oncology guidelines

The paper evaluates how AI systems handle version changes in oncology guidelines, finding that retrieval from multiple versions can lead to outdated answers, and introduces a verification system to reduce errors.

GapDream: Active Synthetic Evidence Acquisition for Long-Tailed Medical Imaging

GAP DREAM improves medical image generation by actively acquiring synthetic evidence that reduces evidence gaps, achieving better performance with fewer samples than random or balanced methods.

Artificial Intelligence-Based Chatbots in Genetic Counseling Practice: Current Uptake, Utilization, and Perspectives

This study surveys genetic counselors on AI chatbot use in clinical practice, finding limited adoption but potential for workflow efficiency with proper training and regulation.

Improving racial fairness in brain age models using style-transfer synthesis

This paper improves racial fairness in brain age models by using synthetic T1 images generated via style-transfer, reducing disparities without requiring race as an input.

Evaluation of AI-assisted summarisation of tertiary clinical genomics reports: results of the QNOMX-VHIR-CPSP-001 Phase 1 study

This study evaluates AI-generated summaries of pediatric cancer genomics reports, finding they are non-inferior to manual summaries in content and presentation.

Turning Domain Expertise into Multi-Dimensional Evaluation of Biomedical AI with Karenina

This paper introduces Karenina, a framework that evaluates biomedical AI using expert knowledge across multiple dimensions beyond traditional scoring.

Evaluating Large Language Models as Tools to Navigate Researchers in Rapidly Evolving Research Landscapes: A Case Study in Cancer Drug Response Prediction

This study evaluates three LLMs for generating a survey on deep learning for cancer drug response prediction, finding they improve reliability but lack human-level analytical depth.

Benchmarking generative AI tools for literature retrieval and summarization in genomic variant interpretation.

This study benchmarks generative AI tools for summarizing genomic variant literature, finding VarChat and Gpt-4o perform best with expert validation needed for reliability.

Diagnostic Accuracy of Large Language Models for Rare Diseases: A Systematic Review and Meta-Analysis

This paper evaluates the diagnostic accuracy of large language models for rare diseases through a systematic review and meta-analysis, finding substantial variation in performance across benchmarks and high risk of bias in all studies.

Artificial intelligence in clinical genetics: current practice and attitudes among the clinical genetics workforce

The paper surveys US-based clinical genetics professionals to assess their knowledge, use, and attitudes toward AI applications in clinical genetics, finding low knowledge and usage but strong support for education.

Ambient AI Documentation in Clinical Genetics: Perspectives on Implementation and Impact on Burnout

This study evaluates how ambient AI affects burnout in genetic counselors, finding reduced burnout and identifying implementation challenges and ethical concerns.

A blinded, counterbalanced rater design for evaluating AI-assisted summarisation of tertiary clinical genomics reports: methodology of the QNOMX-VHIR-CPSP-001 Phase 1 study

This paper describes a blinded, counterbalanced study design to evaluate AI-assisted summarisation of clinical genomics reports using a quality rubric and statistical models.

Trustworthy agentic genomics through versioned skill libraries

This paper shows that trustworthy agentic genome interpretation requires executing validated skills as code rather than relying on models to reason or retrieve information, ensuring deterministic and auditable results.

Development of an Exploratory Taxonomy for Veterinary Professionals' AI Query Patterns Across Clinical Stages: An Expert Panel Study

This study develops a taxonomy of AI query patterns in veterinary practice across clinical stages using expert panel feedback and real-world data.

Cosine Similarity Conflates Clinically Distinct Cancer Variants: A Case for Typed-Graph Retrieval in Precision Oncology Decision Support

This paper shows cosine similarity in RAG systems conflates clinically distinct cancer variants, proposing typed-graph retrieval as a more accurate alternative for precision oncology.

Validation of deep learning enabled web based and smartphone optimized application RadAnalyzer to measure vertebral heart size and vertebral left atrial size in dogs

This study validates RadAnalyzer, a deep learning tool, for measuring VHS and VLAS in dogs, showing strong correlation and minimal bias compared to human observers.

Claim-Level Transparency Analysis of LLM-Generated Diagnostic Reports: A Metabolic and Endocrine Biomarker Study

This study evaluates the factual reliability of LLM-generated diagnostic reports for metabolic and endocrine biomarkers, finding that while hallucination rates are low, many claims lack fully verified evidence.

Bone2Gene: Next-generation Phenotyping of Rare Bone Diseases

Bone2Gene uses deep learning to classify rare bone diseases from hand radiographs, achieving high accuracy for distinct disorders but lower for overlapping ones.

Systematic benchmarking demonstrates large language models have not reached the diagnostic accuracy of traditional rare-disease decision support tools

This study benchmarks large language models against a traditional diagnostic tool for genetic disease diagnosis using 5,213 case reports, finding LLMs lag behind in accuracy.

Performance Characteristics of Reasoning Large Language Models for Evidence Extraction from Clinical Genomics Literature

This paper evaluates large language models for automated PS4 evidence extraction from clinical genomics literature, finding high accuracy but model-dependent performance.

VC-RDAgent: An efficient rare disease diagnosis agent via virtual case construction informed by hybrid statistical-metric and hyperbolic-semantic prioritization

VC-RDAgent improves rare disease diagnosis by generating virtual cases from knowledge bases, outperforming baselines and enabling offline use.

Benchmarking Large Language Models for Predictive Modeling in Biomedical Research With a Focus on Reproductive Health

This study evaluates large language models' ability to generate predictive modeling code for reproductive health using standardized datasets, finding R code generation more successful than Python and some models matching top DREAM challenge teams.

Knowledge Connector: Decision support system for multiomics-based precision oncology

The paper introduces Knowledge Connector, a decision support system that integrates multiomics data with clinical information to aid in precision oncology decision-making.

CARDBiomedBench: A Benchmark for Evaluating Large Language Model Performance in Biomedical Research

CARDBiomedBench evaluates LLM performance in biomedical research using a benchmark focused on neurodegenerative diseases, revealing significant limitations in response quality and safety.

CIViC MCP: Integrating Large Language Models with the Clinical Interpretations of Variants in Cancer

The paper introduces CIViC MCP, a system that uses large language models to interface with the CIViC knowledgebase, enabling natural language queries for cancer variant interpretations.

Query Augmented Generation (QAG) from the Genomic DataCommons for Accurate Variant Statistics

The paper introduces QAG, a method to improve LLM accuracy for somatic variant frequencies by integrating with the Genomic Data Commons.

Benchmarking large language models for genomic knowledge with GeneTuring

This paper benchmarks large language models for genomic knowledge using GeneTuring, finding that a custom GPT-4o setup with NCBI APIs performs best.

What Large Language Models Know About Plant Molecular Biology

This paper benchmarks large language models in plant molecular biology using a new dataset, finding strong performance on multiple-choice questions but limitations in factual accuracy and self-awareness.