Introduction: Automated whole slide image (WSI) quality control (QC) software is critical for digital pathology workflows. Real-world performance across many tissue types, stains and artifact classes remain incompletely characterized. We evaluated three WSI QC tools on a real-world standardized validation dataset to assess performance variability and deployment considerations.

Methods: A total of 519 WSIs representing all major organ systems, specimen types and stain types were analyzed. Three QC platforms (Tool ?, Tool ?, Tool ?) were evaluated for common artifact detection: tissue folds, pen marks, focus/blur, knife-line artifact and missing tissue. Classification performance metrics were evaluated. Subset analysis was performed on H&E slides (n=248) for direct comparison.

Results: Performance varied substantially by artifact and platform. Tool ? showed highest sensitivity for folds (0.959) and focus (0.813) with strong overall discrimination. Tool ? provided balanced performance with moderate sensitivity and specificity across most artifact types. Tool ? achieved the highest specificity across multiple artifacts (e.g. folds 0.969, focus 0.988) with lower sensitivity. Tool ? achieved overall best performance for pen marks (F1 0.757, accuracy 0.891, AUC 0.859). Knife-line detection showed variability with Tool ? showing perfect sensitivity but zero specificity. Missing tissue detection favored Tool ? for overall performance (AUC 0.806) but Tool ? had highest specificity (0.985). These findings highlight trade-offs between sensitivity and specificity-optimized approaches.

Conclusion: No single QC tool showed uniform superior performance across all artifact classes. Each software exhibited distinct performance profiles suggesting that tool selection should be use-case driven. Hybrid approaches may offer improved robustness in clinical deployment. Standardized validation datasets are essential to benchmark performance and guide implementation in digital pathology workflows.

Learning Objectives:

1. Compare performance of WSI QC tools across diverse artifacts and stains.
2. Interpret sensitivity–specificity trade-offs in real-world QC deployment.
3. Identify considerations for selecting QC tools in clinical workflows.