Artificial Intelligence (AI) Cytopathology Trial

Artificial Intelligence for Rapid On-site Evaluation (AI-ROSE) for Endoscopic Ultrasound-guided Fine-needle Aspiration (EUS-FNA) Biopsy of Pancreatic Solid Lesions: A Prospective Double Blinded Study

Purpose The primary objective of the study is to compare interpretation of EUS FNA/FNB samples for adequacy between ROSE and AI at bedside. To compare accuracy of preliminary diagnosis results between ROSE and AI at bedside versus final pathology report.

Research design This is a prospective single center study to compare performance characteristics in the interpretation of EUS FNA/FNB samples between AI and ROSE.

Procedures to be used Eligible patients will undergo EUS guided FNA/FNA of PSLs using standard of care. Sample slides are prepared by a cytopathologist at bedside and observed under a microscope. At the same time, the slides are scanned using a slide scanner and those images are saved for interpretation by AI at a later time.

Inclusion Criteria:

• Have EUS finding of a PSL;
• Do not have contraindications for FNA/FNB.

Exclusion Criteria:

• Inability to provide informed consent for the procedure;
• Contraindication for FNA/FNB eg coagulopathy, lack of avascular window for FNA.