Optimizing Colocalized Cell Counting Using Automated and Semiautomated Methods.

Abstract

Inflammation within the spinal subarachnoid space leads to arachnoid hypercellularity. Multiplex immunohistochemistry (MP-IHC) enables the quantification of immune cells to assess arachnoid inflammation, but manual counting is time-consuming, impractical for large datasets, and prone to operator bias. Although automated colocalization methods exist, many clinicians prefer manual counting due to challenges with diverse cell morphologies and imperfect colocalization. Object-based colocalization analysis (OBCA) tools address these issues, improving accuracy and efficiency. We evaluated semi-automated and automated OBCA techniques for quantifying colocalized immune cells in human arachnoid tissue sections. Both methods demonstrated sufficient reliability across morphologies (P < 0.0001). While automated counts differed significantly from manual counts, their strong correlation (R² = 0.7764-0.9954) supports their reliability for applications where exact counts are less critical. Additionally, both techniques significantly reduced analysis time compared to manual counting. Our findings support the use of automated and semi-automated colocalization analysis methods in histological samples, particularly as sample size increases.