Setting objective thresholds for rare event detection in flow cytometry.

Area of Science:

• Immunology
• Biotechnology
• Data Science

Background:

• Accurate identification of rare antigen-specific cytokine-positive cells in intracellular staining (ICS) flow cytometry assays is challenging.
• Manual thresholding by operators is subjective and inconsistent, while clustering methods struggle with rare events and overlapping populations.

Purpose of the Study:

• To introduce and validate a new objective method for identifying rare antigen-specific cytokine-positive cells.
• To compare the performance of the Fβ measure method against manual gating and clustering algorithms.

Main Methods:

• Developed a novel approach based on the Fβ measure for objective threshold determination in ICS flow cytometry.
• Compared the Fβ method with expert visual gating using ICS data from the EQAPOL proficiency program.

Main Results:

• Visually determined thresholds are difficult to reproduce and problematic for cross-operator/laboratory comparisons.
• Clustering algorithms also face challenges in consistently identifying rare event subsets with distributional overlap.
• The Fβ method demonstrates consistent performance across different centers, samples, and instruments, optimizing the precision/recall tradeoff.

Conclusions:

• The Fβ measure offers an objective and reproducible alternative to manual thresholding for rare cell identification in ICS assays.
• This method improves consistency and reliability in flow cytometry data analysis, particularly for rare event detection.