Pancreatic neuroendocrine tumors (PNETs) are rare; they account for <3 % of all pancreatic tumors [1]. While the survival rate is relatively good for the majority of PNETs, outcomes can vary significantly based on biological heterogeneity [2], [3]. With the updated 2019 World Health Organization (WHO) classification of tumors of endocrine organs, the grading scheme was modified for PNETs: well-differentiated pancreatic NETs (including G1, G2, and G3), poorly differentiated neuroendocrine carcinomas (including small cell type and large cell type), and mixed neuroendocrine-non-neuroendocrine neoplasms [4]. G1, G2, and G3 are graded on the basis of proliferative activity as assessed by the Ki-67 proliferation index and mitotic rate [4]. G1 PNETs have a Ki-67 proliferation index of <3 % and a mitotic rate of <2/2 mm2. G2 PNETs have a Ki-67 proliferation index between 3 and 20 % or a mitotic rate between 2 and 20/2 mm2. G3 PNETs have a Ki-67 proliferation index >20 % or a mitotic rate >20/2 mm2 [4].

Prior studies have shown that larger tumors are more likely to be malignant, have a higher risk of disease recurrence, are associated with worse outcomes, and are more likely to be intermediate rather than low grade [5]. At multivariable analysis, the presence of non-functioning PNETs >2 cm is an independent predictor of malignancy [5], [6]. However, there have been cases of tumors <2 cm, which were malignant, and tumors >2 cm, which were benign, so size alone cannot determine malignant potential. In addition, with the widespread use of high-resolution imaging, an increasing number of non-functional small PNETs are being found. Therefore, in addition to tumor size, a proliferative index such as Ki-67 is used to assign the tumor grade and help predict the biological aggressiveness of PNETs [7], [8], [9], [10]. Using the Ki-67 index evaluated on endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) has a significant impact on clinical management. However, very few studies have assessed the accuracy of the histological PNET grading on preoperative EUS-FNA.

Besides the Ki-67 index, are there any markers that can increase the accuracy of the histological PNET grading on preoperative EUS-FNA? Mitotic counts (MCs) cutoff values were defined by the WHO as an objective criterion. However, substantial variation exists when pathologists manually count mitotic figures. Phosphorylated histone H3 (PHH3) can be used in immunohistochemistry (IHC) assays to identify and quantify dividing cells in tissue samples, which is a marker highly specific to mitotic figures. PHH3 plays a crucial role in regulating gene expression and DNA packing in cells. A lot of research has shown that PHH3 can reliably be used in the routine scoring of mitotic figures and improve diagnostic accuracy, quality, and precision [11], [12], [13].

PNETs development is thought to necessitate both derangement of apoptosis and increased proliferation [14]. BCL-2 is known to be an anti-apoptotic protein [15], [16]. Prior studies suggest that BCL-2 may contribute to tumorigenesis and may be involved in the differentiation of neuroendocrine cells [14], [15], [16]. Therefore, we wondered whether BCL-2 could be a marker that could potentially increase the accuracy of the histological PNET grading on preoperative EUS-FNA.

The present study aimed to retrospectively assess the correlation of the Ki-67 index between EUS-FNA cell block material and subsequently resected surgical specimens. We wanted to study whether the Ki-67 index, evaluated on cytology material, could reliably grade pancreatic neuroendocrine tumors. We also added our experience of immunohistochemical staining for PHH3 and BCL-2 in PNETs.