%0 Journal Article
%T Large Language Model–Driven Automation of Microbiome Diagnostic Reporting in Clinical Laboratories
%A Mateo Alvarez
%A Sofia Herrera
%A Diego Cruz
%A Andres Castro
%J Journal of Medical Sciences and Interdisciplinary Research
%@ 3108-4826
%D 2026
%V 6
%N 1
%R 10.51847/TTj3AtzOb6
%P 71-92
%X Rapid progress in genomic technologies is reshaping laboratory diagnostics by enabling high-throughput analysis of complex biological information, notably microbiome profiles. Large Language Models (LLMs) have exhibited strong potential for uncovering actionable knowledge from extensive datasets. However, their capacity to produce microbiome findings reports that include clinical interpretations and personalized lifestyle advice has not yet been investigated. This study introduces a novel framework that harnesses LLMs to automate the creation of findings reports for microbiome diagnostics. The framework embeds LLMs in an event-driven, workflow-centric system designed to promote scalability and adaptability in clinical laboratory operations. Emphasis is placed on ensuring conformity with established clinical norms and regulatory frameworks, including the In Vitro Diagnostic Regulation (IVDR) and recommendations from the High-Level Expert Group on Artificial Intelligence (HLEG AI). A prototype implementation, designated “MicroFlow”, was developed to illustrate the approach. Deployment of the MicroFlow prototype confirms the practicality of using LLMs to automatically generate findings reports. Early feedback from laboratory specialists indicates that LLM integration yields promising outcomes, with the resulting reports judged to be credible and clinically relevant. Nonetheless, additional validation employing authentic clinical datasets is required to determine the system’s accuracy and robustness. The present work demonstrates a viable pathway for applying LLMs to support findings report generation in microbiome diagnostics. Although initial findings are encouraging, continued refinement and rigorous evaluation are imperative to validate the model’s performance and compliance with clinical requirements. Planned future activities will incorporate expert input from laboratory professionals and extensive testing on real-world patient samples.
%U https://smerpub.com/article/large-language-modeldriven-automation-of-microbiome-diagnostic-reporting-in-clinical-laboratories-joslik18pbyulvl