Optimizing Automated and Semiautomated Cell Counting Methods to Assess Spinal Arachnoid Inflammation

Introduction: Spinal subarachnoid space inflammation leads to subarachnoid and arachnoid hypercellularity. Multichannel immunohistochemistry can assess arachnoid inflammation and hypercellularity. To quantify hypercellularity, manual cell counting (of cells colocalized for nuclear and immune markers) is favored because automated and semiautomated methods can miss colocalized, overlapping cells. However, manual counting is time-consuming, impractical for large datasets, and introduces operator error and bias. Object-based colocalization analysis (OBCA) tools are designed to address these issues. To speed up immune cell quantification, we evaluated semi- and fully automated OBCA techniques for detecting overlapping cells, identifying diverse cell morphologies, and reliability in clinical research.

Methods: Two independent observers analyzed 27 surgical arachnoid tissue specimens from 10 patients with arachnoiditis and syringomyelia—multiplex immunohistochemistry techniques stained nuclear and immune markers in inflammatory cells. Paired t-test and correlation analysis was used to compare semiautomated and automated to manual counts.

Results: The semiautomated and automated techniques detected the relative number of colocalized objects reliably and consistently across observers (P < 0.0001 for all groups). The semi-automated counts were not significantly different from the manual counts; the automated counts strongly correlated with manual counts (R^2 = 0.9954, 0.776, 0.9601, 0.8829, 0.9680 for 1 immune marker each; 0.8210 for two immune markers). The automated (14.8) and semiautomated (98.9) methods had shorter mean analysis times (s) than the manual counting method (159.3; P< 0.0001).

Conclusion : The automated and semiautomated cell counts consistently reflected relative differences in arachnoid cellularity and markers, characterized tissue inflammation levels more rapidly than manual methods, and were sufficiently exact for clinical decision-making and research. Automated and semiautomated colocalization analysis methods can speed up the analysis of cellular inflammatory response in arachnoid membranes and other histological samples labeled by various immunohistochemical markers. The time efficiency of automated counting increases as the sample number rises.