Neuroendocrine neoplasms (NENs) represent a rare subset of tumors, comprising merely 0.5 % of all cancer. Gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs) account for 62 %–70 % of all NENs [1].

The 2019 WHO classification of digestive system tumors underscores the pivotal role of tumor grade, histopathology, and proliferative index determined through mitotic counts or Ki-67 proliferative index in prognosticating GEP-NEN outcomes. This classification divides GEP-NENs into neuroendocrine tumors (NETs) and neuroendocrine carcinomas (NECs), with NETs further stratified into G1 (mitosis <2/2 mm2, Ki-67 < 3 %), G2 (mitosis 2–20/2 mm2, Ki-67 3–20 %), G3 (mitosis >20/2 mm2 or Ki-67 > 20 %). The choice of treatment strategies hinges on proliferative index, alongside expression of somatostatin receptors, tumor growth rate, and extent of disease spread [3].

Accurate Ki-67 proliferative index necessitates the evaluation of at least 500 tumor cells in a designated hotspot. The highest value between mitotic count and Ki-67 proliferative index determines the grade in cases of discrepancy. Given its better sensitivity in identifying proliferative activity within GEP-NENs over mitotic count [2], the Ki-67 proliferative index has become a focal point of this study over mitotic activity.

Although manual counting (MC) is regarded as gold standard for Ki-67 proliferative index counting, its time-intensive nature limits the practical utility in routine diagnostic work [[4], [5], [6], [7], [8]]. Eye estimation (EE) offers a traditional but less precise alternative, whereas digital image analysis (DIA) through scanned slides and software is emerging as a favored method due to its efficiency. Our interest also extends to hotspot selection techniques preceding counting. Gradient map visualization (GM) employs heatmaps to accentuate the intensity of positively immunostained areas, facilitating hotspot identification. While software-based hotspot identification correlates well with eye-selected hotspots, evidence comparing the efficacy of GM to traditional methods remains inconclusive. Thus, this research focuses on comparison of tumor grades between GM-assisted and eye-selected hotspots and between each counting methods.