Sa1450 YIELD OF PANCREATIC FINE-NEEDLE BIOPSY NOT AFFECTED BY PRIOR BILIARY STENT PLACEMENT

Abstract

Endoscopic ultrasound for suspected malignant pancreatic lesions is useful for confirmation of diagnosis, identifying medically treatable disease, accurate staging and may improve survival for those undergoing surgery. Traditionally, EUS is performed prior to ERCP for those requiring biliary drainage over concern that stenting may affect staging or biopsy yield secondary to inflammation or sonographic distortion. To date studies have been conflicting on whether pre-operative stenting affects biopsy yield. All pancreatic fine needle biopsies performed at our center over the last 7 years (October 2012 - October 2019) were identified by searching the pathology laboratory electronic medical record for the specimen types 'pancreas fine needle biopsy'. Clinical chart review was undertaken and age, sex, location of lesion, size of lesion, final pathologic diagnosis, date and time of biopsy and date and time of biliary stenting, if any was recorded. On-site cytopathology is not used in EUS guided biopsy at our center. Biopsy interpretation was taken from the final pathology report and classified as either diagnostic or non-diagnostic. Two tailed P-value for difference in rate of diagnostic biopsy was performed using Fisher’s exact test. Confidence interval of diagnostic yield was calculated using the modified Wald method. 233 individual results were identified. 9 duplicates were excluded. 54% were male. Mean age was 65. A definitive diagnosis was made in 184 biopsies, 40 biopsies were non-diagnostic. 131 samples were classified as pancreatic head, uncinate process or neck masses. 85 were classified as pancreatic body or tail with 8 from either an atypical site or unreported site. 61 biopsies were performed with a stent in-situ. Diagnostic yield for biopsies on unstented patients was 0.82 (95% CI 0.75-0.87) versus 0.84 (95% CI 0.72-0.91) for biopsies on patients with biliary stents in-situ. This difference was not statistically significant (p=0.8455). Post-hoc power calculations suggest greater than 80% likelihood of identifying a significant difference with the current sample size. Similar analyses were performed using only pancreatic head lesions with a similar result though the test was underpowered to detect a significant difference. In the absence of medical or logistical reasons, EUS guided biopsy prior to biliary stenting should be considered as pathologic diagnosis may affect therapeutic decisions. However, our analysis suggests that in situations when biliary stenting prior to biopsy is indicated or if prior EUS is impractical, clinicians can be reassured that future biopsy yield will likely not be affected.