Artificial intelligence for rapid on-site evaluation of lymph node fine-needle aspiration: improving diagnostic efficiency and accuracy

The study evaluated the clinical value of AI-ROSE, a novel tool for rapid cytological analysis of lymph node samples obtained through CT-guided biopsy. The research included 54 patients biopsied between June 2024 and June 2026. AI-ROSE demonstrated sensitivity of 90.48% and diagnostic accuracy of 89.80%, which were higher than traditional exfoliative cytology. The consistency between AI-ROSE and histopathological gold standard was moderate (κ = 0.647). AI-ROSE provides real-time assessment of specimen adequacy and guidance for repeat puncture decisions. However, researchers emphasize that AI-ROSE should serve only as a supplementary tool and cannot replace final histopathological diagnosis.