Background and aims: The use of commercially available comprehensive genomic profiling tests, such as the FoundationOne® CDx (F1CDx), has increased. However, the success rate of F1CDx using samples obtained by endoscopic ultrasound-guided tissue acquisition (EUS-TA) is suboptimal. This study aimed to verify the optimal method for obtaining samples suitable for F1CDx among three EUS-TA techniques.
Methods: In a randomized, non-comparative, three-arm trial, patients with unresectable pancreatic cancer scheduled for EUS-TA using a Franseen needle for pathological confirmation were randomized into three groups: two passes with a 19-gauge (G) needle; four passes with a 22-G needle; and macroscopic on-site evaluation (MOSE) using a 22-G needle until the macroscopic visible core length reached ≥40 mm. The primary endpoint was the acquisition rate of an ideal sample meeting the F1CDx quantity criteria; the secondary endpoints were procedure time, adverse events, and F1CDx success rate.
Results: Seventy-five patients were enrolled. The acquisition rates of ideal samples for F1CDx (96%, 92%, and 96% for the 19-G, 22-G, and MOSE groups, respectively; all p<0.05) were significantly greater than the expected 60%. The 19-G and MOSE groups had significantly shorter procedure times than the 22-G group (814 s vs. 896 s vs. 963 s). The 22-G group showed moderate abdominal bleeding. F1CDx was successful in 10, 12, and 9 cases in the 19-G, 22-G, and MOSE groups, respectively, representing a 100% success rate for each group.
Conclusions: Two passes using a 19-G needle and MOSE using a 22-G needle are recommended to collect ideal samples for F1CDx.
Keywords: Biopsy; Needles; Neoplasms; Pathology.
Copyright © 2025 American Society for Gastrointestinal Endoscopy. Published by Elsevier Inc. All rights reserved.