2023 World Conference on Lung Cancer (Posters)-EP06.01. Evaluation of Unstained Cytology Slides for Artificial Intelligence Rapid On-Site Evaluation

EP06.01. Evaluation of Unstained Cytology Slides for Artificial Intelligence Rapid On-Site Evaluation - PDF(Abstract)

Pdf Summary

Artificial Intelligence (AI) has been increasingly used in pathology for tissue subtyping and quantification, but there is limited data on its use in pulmonary cytopathology. A recent study validated an AI system for rapid on-site evaluation (ROSE) of pulmonary malignancies, showing high sensitivity, specificity, and accuracy. However, this process requires uniform processing and staining, which can result in tissue loss. In this study, researchers compared a novel label-free microscopy technique that does not require staining to visualize cellular features in cytology specimens to standard slide preparation for AI ROSE.<br /><br />The researchers used a label-free microscopy system to image unstained cytology specimens collected during pulmonary procedures. The specimens included fine-needle aspirate, touch prep, and bronchial brush samples. The unstained images revealed that red blood cells, macrophages, and bronchial cells were well visualized. The novel microscope was able to clearly highlight nucleated cells without the need for staining, providing endogenous contrast between the nucleus and cytoplasm.<br /><br />This label-free microscopy technique allows for the quick digitalization of pulmonary specimens in real time, creating images that allow for the identification and differentiation of tissue and cell types. This eliminates the need for staining the slides, preserving tissue for additional testing, biomarker analysis, and special staining for specimens with low cellularity. Furthermore, standardization of slide preparation will enable the universal application of AI ROSE.<br /><br />Overall, this study demonstrates the feasibility and potential clinical value of using a label-free microscopy technique for AI ROSE in pulmonary cytopathology. The technique can improve the diagnostic accuracy of ROSE and facilitate the application of AI in pathology.

Asset Subtitle

Eric Flenaugh

Meta Tag

Speaker Eric Flenaugh

Topic Pathology & Biomarkers: Artificial Intelligence in Pathology

Keywords

Artificial Intelligence

pathology

tissue subtyping

quantification

pulmonary cytopathology

AI system

rapid on-site evaluation

label-free microscopy

cytology specimens

diagnostic accuracy