Using Bland-Altman Analysis to Identify Appropriate Clonogenic Assay Colony Counting Techniques

Area of Science:

• Cell biology
• Cancer research
• Assay development

Background:

• Clonogenic assays are crucial for evaluating cancer cell survival and treatment response.
• Accurate colony counting is essential for reliable assay interpretation.
• Variability in manual counting methods can impact experimental outcomes.

Purpose of the Study:

• To determine the interchangeability of different colony counting methods in clonogenic assays.
• To compare manual, microscopic, automated, and semiautomated counting techniques.
• To identify the most reliable method for colony enumeration in head and neck cancer cell lines.

Main Methods:

• Four counting modalities were evaluated: manual pen, microscopy, a public automated algorithm, and a novel semiautomated algorithm.
• Clonogenic assays using two head and neck cancer cell lines were analyzed (N=24 wells per method).
• Interobserver variability and statistical analyses (correlation, Bland-Altman) were performed.

Main Results:

• Manual pen counting exhibited significant bias compared to microscopic and semiautomated methods.
• The semiautomated algorithm demonstrated high correlation and no bias with microscopic counts (R²=0.970 and 0.931).
• Automated algorithms showed lower correlation with microscopic counts (R²=0.384).

Conclusions:

• Manual pen counting is not interchangeable with microscopic or semiautomated methods due to significant bias.
• Semiautomated algorithms are highly reliable and interchangeable with microscopic counts for clonogenic assays.
• Future algorithm development should prioritize size and shape parameters, and avoid sole reliance on manual counts for training.