Interobserver Variability in Ki-67 Index Assessment of Gastroenteropancreatic Neuroendocrine Neoplasms: A Comparative Evaluation of Three Counting Methods.

Background & Objective: Accurate evaluation of the Ki-67 proliferation index is critical in the grading and prognostication of neuroendocrine neoplasms (NENs). This study aimed to compare three different methods of Ki-67 index assessment in primary gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs). Methods: The Ki-67 proliferation index was assessed by immunohistochemical (IHC) staining in 41 cases of primary GEP-NENs. Two pathologists independently evaluated the Ki-67 index using two visual estimation methods: eyeballing and eye-counting under a light microscope. A third method involved manual counting of Ki-67-positive and -negative nuclei on a captured image using ImageJ software. Interobserver and intermethod reproducibility were analyzed using Pearson's correlation coefficient (R) and the intraclass correlation coefficient (ICC) to assess agreement among the four observers and between each observer and the manual method. Results: The ICC among the four observers was 0.993, indicating excellent interobserver agreement. The ICC between each observer's score and the manual counting method was 0.994, also demonstrating high concordance. Conclusion: All three methods, eyeballing, eye-counting, and manual counting using ImageJ, proved to be comparably effective in assessing the Ki-67 proliferation index in GEP-NENs. Notably, the simpler microscopic techniques of eyeballing and eye-counting showed excellent agreement with the manual image-based method, supporting their reliability in routine practice.