Identification of Endosonographic Features that Compromise EUS-FNB Diagnostic Accuracy in Pancreatic Masses.

Background: Endoscopic ultrasound-guided fine-needle biopsy (EUS-FNB) is highly accurate for diagnosing pancreatic mass. However, making diagnosis is challenging in 5-20% of patients. This study investigated the challenging features associated with reduced diagnostic performance in EUS-FNB and potential rescue methods that can improve the diagnostic rate.

Methods: This single-center retrospective study included patients with solid pancreatic tumors who underwent EUS-FNB between January 1, 2019, and December 12, 2021. Patients without a computed tomography (CT) scan or definite diagnosis were excluded. Challenging features were features that reduced diagnostic accuracy in EUS-FNB, as determined through multivariate analysis. Rescue methods were methods that assisted operators in assessing lesions in patients with challenging features.

Results: Of 332 enrolled patients, an accurate diagnosis obtained using EUS-FNB was achieved in 286 (86.1%). Univariable analysis revealed that the diagnostic accuracy was lower in cases of pancreatic tumors with isoattenuation in CT images (77.3% vs. 89.8%, odds ratio [OR]: 0.39, p = 0.003), an ill-defined margin on EUS (61.2% vs. 92.5%, OR: 0.13, p < 0.001), or tumor size < 20 mm (65.5% vs. 88.1%, OR: 0.26, p = 0.002). However, only ill-defined margins on EUS (OR: 0.14, p < 0.001) and tumor size < 20 mm (OR: 0.25, p = 0.005) were independent predictors of inconclusive EUS-FNB in the multivariate analysis. The use of contrast (OR: 4.46, p = 0.026) and a highly experienced endosonographer (> 5cases/month; OR: 3.25, p = 0.034) improved diagnostic performance in difficult cases.

Conclusions: Pancreatic tumors with ill-defined tumor margins on EUS or size < 20 mm are challenging features in EUS-FNB. The use of contrast and a highly experienced endosonographer can improve diagnostic performance in difficult cases.