Introduction: In recent years, matched therapy based on comprehensive genomic profiling (CGP) has extended the prognosis of pancreatic cancer patients. Foundation One CDx® (F1CDx) is a representative test of commercially available CGP (CACGP). Tissue samples for CACGP in pancreatic tumors mainly come from surgery and EUS-guided tissue acquisition (EUS-TA), with success rates of 90-100% and 60-70%, respectively. Optimizing EUS-TA techniques is crucial. We defined specimens that meet the criteria for F1CDx as the ideal sample for F1CDx, as determined by pathologists. In our retrospective study, comparing specimens with and without the ideal sample for F1CDx, the success rate of F1CDx was 94.1% vs. 55% (p < 0.01). Furthermore, we reported that for the ideal sample for F1CDx collection, a method involving a 19 gauge (G) Franseen needle for 2 passes, a 22 G Franseen needle for 4 passes, and determining the number of punctures with Macroscopic On-Site Evaluation (MOSE) using a 22 G Franseen needle could be an optimal approach for F1CDx submission.
Aims: The aim of the prospective trial is to investigate whether utilizing these three methods can allow for the collection of ideal samples for F1CDx and enable the success of F1CDx.
Methods: In a randomized, non-comparative, three-arm phase II trial, patients with unresectable pancreatic cancer scheduled for EUS-TA for histological confirmation were randomized into three groups: 19G 2-pass group (19G group), 22G 4-pass group (22G group), and 22G MOSE group (MOSE group). The primary endpoint was the proportion of ideal samples for F1CDx collection, with secondary including puncture number, procedure time, adverse events, and, in F1CDx cases, the proportion classified as passed and qualified (success of F1CDx). Setting the threshold for the rate of ideal samples for F1CDx collection at 60%, the expected rate at 85%, one-sided 5% significance level, and 80% power, the target number of cases for each group was set at 25.
Results: Seventy-five patients (25 in each group) were enrolled, and the prescribed number of punctures was achieved in all cases. The proportion of ideal samples for F1CDx collection were 96%, 92%, and 96% for the 19G, 22G, and MOSE groups, with p-values < 5%. The average number of punctures was 2.0/4.0/2.4 times, with the 22G group significantly higher than others. The average procedure time was 880/1060/854 seconds, with the 22G group notably longer. F1CDx was performed in 10/12/9 cases, and all were successful. Adverse events occurred only in the 22G group, involving severe intra-abdominal bleeding.
Conclusion: In EUS-TA, it is feasible to obtain the ideal sample for F1CDx using either the 19G 2-pass, 22G 4-pass, or 22G MOSE methods. Considering the number of punctures, procedure time, and incidence of adverse events, the use of 19G 2-pass or 22G MOSE is preferable.

Takeshi Kawakami

Appearances