L. Celi

Reporting completeness of clinical trials evaluating large language model interventions in medicine: systematic review

Clinical trials of LLM interventions show modest adherence to established reporting guidelines, with persistent underreporting of LLM specific methodological details, reproducibility artefacts, and harms reporting required to support reproducibility and confident appraisal of safety risks before their clinical translat...

Echo Chambers: Bias and Representation in Cardiac Imaging Datasets for Artificial Intelligence

Diabetic retinopathy screening using artificial intelligence: a comprehensive performance comparison on datasets from diverse populations and imaging modalities

The study highlights the impact of imaging equipment, demographics, and image quality on diagnostic performance and underscores the need for benchmarking AI-based DR screening tools using standardized datasets that encompass diverse populations and imaging conditions.

1,357 AI medical devices cleared, 3 actually tested on patient outcomes

The finding that only 0.2% of cleared AI devices have undergone evaluation for patient-centered outcomes reveals a profound validation gap and points to the need for evidence standards capable of keeping pace with the speed of regulatory clearance.

ModaLens: Measuring Image Sensitivity in Report-Conditioned Medical VLMs

ModaLens, a paired image-swap audit, measures how report availability changes image sensitivity and limits conclusions about visual correctness; the direction replicates in two further model lineages.

Towards a Deterministic Math Solver for Clinical Language Models

Adding an executor to the Program-Solve interface helps some open-weight models more than others even under matched formula, variable and note access, and is not a substitute for verified formulas or reliable variable extraction either way.

From policing authors to stewarding science: Governing artificial intelligence across the scientific publishing pipeline in medical journals

It is argued that effective AI governance in scientific publishing cannot be achieved through static rules or journal-centric control alone, and requires a shift toward shared, adaptive oversight of the scientific publishing system.

Agents Catching Agents: Shortcut Cascades and Benchmark Gaming in Clinical Multi-Agent Systems

It is asked whether committees of language-model agents deliberating on a shared workspace can be gamed by shortcuts, cues a benchmark rewards but a clinician would ignore, and what games a committee is social plausibility.

Clinical predictive artificial intelligence evaluation: A narrative review of trial designs and practical considerations

This narrative review outlines the limitations of traditional evaluation frameworks and proposes a paradigm shift toward adaptive, iterative, and context-specific assessment methodologies that can progress from a promising technology to reliable clinical tools that improve patient outcomes, support clinical decision-ma...