INTRODUCTION: The advent of endoscopic ultrasound (EUS) has vastly improved access to several gastrointestinal lesions, thereby leading to increased diagnostic capability. Moreover, improved core tissue sampling is now possible with fine needle biopsy (EUS-FNB) needle, as opposed to aspiration (FNA) alone. However, the practice of sample handling is highly individualized due to lack of standardization as to whether core sample should be sent only in histology (in formalin), or cytology (in cytolyte) or both. This study was abstracted to compare the diagnostic yield of cytology versus histology on core tissue obtained from the same lesion, using the same EUS-FNB needle. METHODS: Single endoscopist at a large tertiary referral center conceptualized a prospective EUS database between October 2017 and February 2019, where a 22-gauge FNB needle was used to obtain core tissue samples, which were sent for cytology and histology. After IRB approval, retrospective analysis of this database was performed. Patient’s demographics, medical history, imaging, indication for EUS, EUS findings, and cytology/histology results were analyzed. RESULTS: Out of 226 therapeutic EUS' performed by single endoscopist during study period, 103 samples were collected using 22-gauge FNB needle [solid masses (62), lymph nodes (18) & solid-cystic lesions (23)]. 28 samples were sent for cytology, 29 for histology, and 46 for both (passes equally divided). Out of these 46 samples, 42 were consistent between cytology and histology, but 4 were discordant; in all 4 cases, histology provided the conclusive information. In 3 cases, histology provided the diagnosis of cancer while cytology was reported negative. In the fourth case, cytology was inconclusive while histology was definitely negative for carcinoma. These samples were obtained from 2 pancreatic masses, one gastric subepithelial lesion, and one lymph node, and pathology showed pancreatic adenocarcinoma, GIST, and persistent follicular lymphoma, respectively. 22-gauge needles were used for all these procedures, and no complications were noted. CONCLUSION: Our study demonstrates 8% inconsistency between cytology and histology on core samples obtained using EUS-FNB. In all discordant cases, histology provided more accurate information than cytology. Larger, multi-centric and prospective studies are needed to ascertain the best practice to make the diagnostic process more efficient, accurate, and cost-effective.
Read full abstract