S3371 Diagnostic Efficacy of Endoscopic Ultrasound-Guided Fine-Needle Biopsy (EUS-FNB) in the Evaluation of Solid Pancreatic Neoplasms: A Community Hospital Experience

Abstract

Introduction: Endoscopic ultrasound-guided fine-needle biopsy (EUS-FNB) plays a crucial role in the early diagnosis of pancreatic cancer. Majority of EUS procedures are performed in tertiary care centers, and data from community hospitals is limited. Our goal was to assess EUS-FNB performance and its diagnostic efficacy in a community hospital setting and compare it with specialized centers. Methods: We reviewed electronic health records of patients who underwent EUS-FNB procedure for solid pancreatic lesions at MedStar Southern Maryland Hospital Center between 2019-2020. All cases underwent EUS with linear array echoendoscope, and FNB was performed by an experienced gastroenterologist using a 22-gauge Franseen tip needle. Pathologist performed rapid onsite evaluation (ROSE). EUS impression was correlated with final cytology diagnosis. Follow-up data were obtained on all neoplastic cases by reviewing surgical pathology reports and medical oncology notes. Results: We had 38 patients who underwent EUS-FNB, of which 17 were males and 21 were females. Age group 43-90 years (mean: 66 years). Number of FNA passes 2-6 (average: 3.6). Pancreatic lesions size range 0.8-5.8 cm (mean: 2.8 cm). ROSE was performed on 35/38 cases (92.1%), of which 30/35 cases (85.7%) were deemed adequate. Location of the pancreatic lesions: 20 head, 2 neck 2, 1 uncinate process, 8 body, and 7 tail. By cytology, 20 cases were neoplastic, 12 non-neoplastic, and 6 non-diagnostics. EUS impression was accurate in 18/20 (90%) neoplastic cases (18 adenocarcinomas, 1 Intraductal papillary mucinous neoplasm, and 1 neuroendocrine tumor), and the remaining 2 (10%) cases had initial incorrect EUS impression of pancreatitis. Of 12 non-neoplastic cases on cytology only 7 precisely correlated with EUS impression. The 6 cases deemed non-diagnostic by cytology were excluded from statistical analysis. The overall discrepancy rate between cytology and EUS was 21.9% (7/32) (Table). Conclusion: Calculated sensitivity, specificity, positive predictive value, and negative predictive value for EUS-FNB were 90%, 67%, 82%, and 80%, respectively. The calculated statistical measures of EUS-FNB's performance from our community hospital (sensitivity of 90%) are comparable to published data from larger specialized centers. With a diagnostic efficacy of 90% and higher tissue yield on EUS-FNB, the technique can be safely performed in community hospitals for appropriately triaging patients to specialized centers for a higher level of care as needed.Table 1.: Correlation between Endoscopic Ultrasound (EUS) Impression, Cytology Diagnosis, and Final Diagnostic Surgical Resection/ Medical Oncology. Footnotes: ROSE: Rapid On-site Evaluation; IPMN: Intraductal Papillary Mucinous Neoplasm.