There has been much controversy regarding the accuracy of grading pancreatic neuroendocrine tumors (PNETs) on fine-needle aspiration (FNA) biopsy. The objectives of this study were to evaluate whether grading according to the fraction of Ki-67-positive tumor cells (the Ki-67 proliferation index) on material from endoscopic ultrasound-guided FNA biopsies correlated with grading on surgical resection specimens and to evaluate the minimum amount of FNA material needed.
A case series of 27 PNETs with FNA biopsies and corresponding surgical resection specimens at the authors' institution were evaluated. Tumors were graded on FNA and surgical specimens with an evaluation of Ki-67 index according to 2010 World Health Organization criteria. Chart reviews were conducted to evaluate recurrence or clinical progression in patients who were being managed conservatively with observation.
The evaluation of grading between FNA and tumor resection specimens revealed that 22 of 26 FNA specimens (84.6%) had Ki-67 results comparable to those in the corresponding surgical resection specimens, thus allowing for accurate grading. Correct FNA diagnosis with the ability to distinguish between grade 1 and 2 tumors had a positive predictive value of 88.9%, with 72.7% sensitivity, 93.3% specificity, and a P value of .00081. In addition, 24 of 26 cases contained less than 2000 cells, of which 20 were correctly graded on FNA material. Seven of 26 FNA samples had less than 1000 cells, of which 6 were correctly graded, including 2 that had only 50 cells.
The current results exhibit good correlation between FNA grade and final grade on surgical resection specimens using Ki-67 index, even in samples with less than the recommended total cell count. Therefore, grading of PNETs on FNA with the Ki-67 proliferation index should be assessed and is a practical parameter to report to clinicians. Cancer Cytopathol 2018;126:170-8. © 2017 American Cancer Society.