Comparing Needles and Methods of Endoscopic Ultrasound-Guided Fine-Needle Biopsy to Optimize Specimen Quality and Diagnostic Accuracy for Patients With Pancreatic Masses in a Randomized Trial

Abstract

Background & aims: Given the lack of procedure standardization, findings vary from analyses of pancreatic tissues collected by endoscopic ultrasound-guided fine-needle biopsy. It is not clear which needle and technique yield the best specimen for analysis. We compared the specimen quality and accuracy of diagnoses made from samples collected by fine-needle biopsy needles using different collection techniques.

Methods: Patients found to have pancreatic masses during imaging (n = 129) were assigned randomly to groups from whom pancreatic tissue samples were collected by reverse-bevel, Menghini-tip, franseen, or fork-tip needles. A second randomization determined the technical sequence of biopsies in each patient (suction, no suction, and stylet retraction). Two independent pathologists, blinded to the type of needle and sampling technique, analyzed all the samples. Final diagnoses of malignancy were made based on surgical resection, death from cancer progression, or findings from radiology or clinical follow-up evaluations (reference standard). The primary objective was to compare the cellularity of the samples collected, defined as the proportion of core tissue in the biopsy sample. Secondary objectives were to compare the accuracy of diagnoses made from biopsy samples and identify factors associated with high cellularity.

Results: One-hundred and nine patients had a final diagnosis of malignancy (84.5%) and 20 patients had benign disease (15.5%). Samples collected by fork-tip or franseen needles had significantly higher cellularity than samples collected by reverse-bevels or Menghini-tip needles (P < .001). Neoplasias were identified with greater than 90% accuracy using samples collected by fork-tip or franseen needles (P < .001 compared with other needles). On multivariable regression analysis, use of franseen needles (odds ratio [OR], 4.42; 95% CI, 2.58-7.58; P < .001) or fork-tip needles (OR, 3.86; 95% CI, 2.24-6.64; P < .001), stylet retraction (OR, 2.13; 95% CI, 1.21-3.72; P = .008), no suction (OR, 2.74; 95% CI, 1.57-4.80; P < .001), and pancreatic mass larger than 3 cm (OR, 1.92; 95% CI, 1.21-3.05; P = .005) were associated with high cellularity of the sample.

Conclusions: In patients with suspected pancreatic cancer, samples with the highest degree of cellularity in a single biopsy, resulting in a diagnostic accuracy of 90% of higher, were collected by fine-needle biopsy using the franseen or fork-tip needle. Clinicaltrials.gov no: NCT04085055.

Keywords: EUS; FNB; Sample Quality; Tumor.