Abstract:

Aim: Ki67 protein, which shows promise as an immunohistochemical biomarker in breast cancer, is used in cell proliferation assessments since it is present in all active phases of the mitotic cycle. Ki67 index has predictive and prognostic value in patients with breast cancer. Ki67 counting with manual assessment (MD) is sensitive to interobserver variability and is time consuming. In recent years, studies indicating that digital image analysis (DGA) is fast and objective for Ki67 measurements have been increasing, but the routine application of this method requires further studies. In this study, we compared MD, DGA and GK Ki67 measurements in 85 invasive breast cancer cases. Method: Tumor molecular types, mitotic numbers, Ki67 values measured by GK, MD and DGA and their correlations, were determined. DGA analyses were analyzed with ViraPath (Virasoft Yazılım, Istanbul, Turkey) software, and statistical correlations between parameters were analyzed with NCSS (Number Cruncher Statistical System, 2020). Results: The correlations and differences of Ki67 index values determined by all three different methods in terms of age, histological grade, mitotic numbers, and molecular type were found to be compatible with the international literature. The intraclass correlation coefficient between Ki67 indices counted with DGA and MD was measured as 0.974, and the difference between Bland Altman analysis and MD and DGA counts was found to be close to zero. Conclusion: DGA counts give reliable results to replace MD. With multicenter studies that optimize method standards, time savings and high sensitivity can be brought to the practice of pathology.

Keywords: