Objectives: This study aimed to determine whether Ki-67 index evaluated on cytologic material could reliably grade pancreatic neuroendocrine tumors (PanNETs).
Methods: Cases with adequate cell block and available surgical specimens were included. Ki-67 index was calculated using "eyeballing," "hot spot," and "complete" counting methods.
Results: The overall concordance rates between cytology and surgical specimens were 71%, 73%, and 59%, respectively, by using eyeballing, hot spot, and complete counting approaches. All grade 1 tumors were correctly graded on cytology, but in grade 2 tumors concordance rates were only 36%, 41%, and 9%, respectively. All grade 2 tumors were undergraded when cell blocks contained fewer than 1,000 cells, while concordance rate increased to 57%, 64%, and 14%, respectively, in cases with 1,000 cells or more.
Conclusions: Grade 2 PanNETs can be significantly undergraded when Ki-67 index is evaluated on cell block material. In cases with 1,000 or more cells, the hot spot counting method has better correlation with surgical specimens.
Keywords: Fine-needle aspiration; Ki-67 index; Neuroendocrine tumor; Pancreas.
© American Society for Clinical Pathology, 2019. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.