Subscribe to RSS
DOI: 10.1055/s-0040-1704040
EUS-FNB WITH VERSUS WITHOUT ROSE: INTERIM ANALYSIS OF AN INTERNATIONAL RANDOMIZED NON-INFERIORITY STUDY
Publication History
Publication Date:
23 April 2020 (online)
Aims To assess if the need of rapid on-site evaluation (ROSE) during EUS-FNB of solid pancreatic lesions (SPLs) can be overcome by the high diagnostic accuracy of new devices.
Methods Randomized non-inferiority study comparing EUS-FNB+ROSE vs. EUS-FNB alone in patients with SPLs. The procedures were done using one of the available FNB needles (SharkCore22G, Acquire22G, ProCore20G) with 3 needle passes performed and the touch imprint cytology technique (TIC) utilized for ROSE. In the ROSE arm when adequacy with TIC was reached, additional passes up to an overall of 3 were performed for histological evaluation. Samples suspicious for malignancy were considered false negative when final diagnosis of malignancy was ascertained. Diagnostic accuracy was measured against definitive surgical pathology or after a clinical course of at least six months congruent with the diagnosis. Secondary endpoints included: safety, presence of tissue core, quality of specimens and time of the sampling procedure.
Results We completed the enrollment of the whole 800 patients needed to finish the study. At present, 327 patients have completed the follow-up (193 males, mean age 66.6±11). Patients’ and lesions’ characteristics in the two arms were similar. Comparable diagnostic accuracies were obtained with or without ROSE [90.3% (95% CI 84.7-94.4) vs 91.3% (95% CI 85.9-95.2), respectively]. Safety, presence of tissue core, and sample quality of histological specimens were similar in the two groups. The mean time of the sampling procedure was significantly longer in the ROSE group (16.1±7.2 vs 11.5±6.0 minutes, p< 0.0001). Significantly higher rate of tissue core procurement was obtained using the end-cutting needles than the side-fenestrated needle (48.9% vs 75% vs 69.7% using Procore 20G, SharkCore 22G, and Acquire 22G, respectively, p< 0.0001).
Conclusions Our preliminary results show that ROSE is not associated with better accuracy of EUS-FNB alone and its utilization increases the duration of the procedure.