Computer system for cell counting in selected brain tumors at Ki-67 immunohistochemical staining

Area of Science:

• Computational Pathology
• Neurosurgical Oncology
• Immunohistochemistry

Background:

• Accurate tumor cellularity assessment is crucial for diagnosing meningioma and oligodendroglioma.
• Tumor cell number variability within histopathologic slides poses challenges for standardization.
• Ki-67 immunohistochemistry is a key marker for cell proliferation in these tumors.

Purpose of the Study:

• To introduce a novel computerized system for quantitative cell counting in meningioma and oligodendroglioma.
• To evaluate the system's performance in standardizing tumor cellularity assessment.
• To analyze cell number variability in Ki-67 stained histopathologic slides.

Main Methods:

• Development of a computer program utilizing mathematical morphology for quantitative slide evaluation.
• Integration of a Support Vector Machine algorithm for classifying cell immunoreactivity.
• Analysis of cell counts in Ki-67 immunohistochemically stained meningioma and oligodendroglioma specimens.

Main Results:

• The system achieved a mean relative discrepancy of 8% compared to human expert scoring.
• Meningioma slides showed a mean of 623 cells per field of view (range: 386-781).
• Oligodendroglioma slides had a mean of 474 cells per field of view (range: 204-736).

Conclusions:

• The developed computerized system is an efficient tool for supporting histopathologic diagnosis.
• The system's sequential thresholding effectively mimics human cell recognition processes.
• Significant variability in tumor cellularity was observed within and between patients, highlighting the need for standardization.