Utility of the manual hematology cell counter for Ki-67 assessment in gastrointestinal and pancreatobiliary well-differentiated neuroendocrine tumors

Abstract

In the World Health Organization (WHO) 5th edition, prognosis of gastrointestinal (GI) well-differentiated neuroendocrine tumors (WDNET) depends on proliferation rate, commonly assessed by ki-67 immunohistochemical stain. In daily practice, the gold standard for WHO grade assessment by ki-67 staining, printing a photo of a tumor hotspot, counting the number of ki-67-positive cells out of 500 tumor cells, and calculating a percentage, is time-consuming and many cases are eyeballed. This study investigates the utility of a common tool, the manual cell counter used in hematology smear cell counting, for GI WDNET ki-67 counting. Of 59 resections, the number of cases with a WHO grade difference between gold standard print-and-count and the original report, eyeballing, and hematology counter method, was 23 (39 %), 14 (24 %) and 7 (12 %) cases, respectively. Of 37 biopsies, the number of cases with a WHO grade difference between gold standard print-and-count and the original report, eyeballing, and hematology counter method, was 10 (27 %), 12 (32 %) and 7 (19 %) cases, respectively. For resections, Chi square analysis comparing hematology counter method versus original report, where many cases were likely eyeballed, showed statistically significantly less cases with differing WHO grades from gold standard print-and-count for hematology counter-assessed cases (P = 0.0007), and the same Chi square analysis was marginally not significant (P = 0.09) for hematology counter versus eyeballing. Times taken to perform hematology counter method were statistically significantly lower than times taken for print-and-count. This study suggests the hematology cell counter could strike a reasonable balance between time and accuracy for WDNET resections.