Intraoperative assessment of pancreatic parenchymal margin during pancreatectomies is challenging and misinterpretation by the pathologist is a cause of incorrect frozen section (FS) diagnosis. Although the current literature supports that pancreatic margin FS diagnosis and its accuracy has no impact on the patient outcome for pancreatic ductal adenocarcinoma (PDAC) patients and reexcision in an attempt to achieve a negative intraoperative pancreatic margin after positive FS is not associated with increased overall survival; still it remains a routine practice in many institutions. To this end, we sought to assess the interobserver variation and accuracy of FS diagnosis between subspecialized gastrointestinal/pancreatobiliary (GI) and general pathologists. Seventy seven consecutive pancreatic parenchymal margin FSs performed on pancreatectomies for PDAC from 2010 to 2013 were retrieved at our institution. These were retrospectively evaluated by 2 GI and 2 general pathologists independently without knowledge of the original FS diagnosis or the final diagnosis. The specificity, sensitivity, positive predictive value, negative predictive value, and accuracy of GI versus general pathologist was 97.8% versus 87.5%, 61.1% versus 66.7%, 78.6% versus 41.4%, 95% versus 95.2%, and 93.5% versus 85.1%, respectively. The interobserver agreement between GI and general pathologists was fair (κ = .337, P < .001). The interobserver agreement between 2 GI pathologists was fair (κ = .373, P = .0005) and between 2 general pathologists was slight (κ = .195, P = .042). Although overall accuracy of subspecialized GI pathologists was higher than that of general pathologists, none had an accuracy of 100%. Our study reaffirms the challenging nature of these FSs.
Keywords: accuracy; frozen section analysis; gastrointestinal; general; interobserver variation; pancreatic ductal adenocarcinoma; pathologist; subspecialized.
© The Author(s) 2015.