Overview
Automated cell counting identifies and enumerates individual cells or validated cellular objects in images. Confluence measures the proportion of an analysed surface occupied by cells, usually expressed as a percentage. Both can describe culture growth, but their biological meaning differs: a change in cell size or spreading can alter confluence without a corresponding change in cell number. [1,2]
How image based counting works
Image analysis first distinguishes relevant objects from the background, a process called segmentation. Depending on the assay, software may identify whole cells, nuclei or other structures using transmitted-light contrast or fluorescent labels. Measurements can then include object count, area, shape and fluorescence intensity. A nuclear count is a proxy for cell number only when the relationship between nuclei and cells is appropriate. Multinucleation, nuclear fragmentation and overlapping nuclei can break that assumption. [1]
How confluence is calculated
Confluence is the area classified as cell-covered divided by the valid analysed area, multiplied by 100. Define exclusions such as image borders, debris and unusable regions consistently. Measurements from selected fields estimate the culture only when the sampling is representative. Near full coverage, confluence has little remaining range, so further increases in cell number may be poorly resolved. [2]
Selecting the right biological endpoint
Cell counts are useful for growth curves and normalisation when individual objects can be resolved reliably. Confluence can be practical for adherent-culture monitoring and wound closure assays. Neither measurement alone establishes viability. Dead cells can remain attached or retain a detectable nucleus, while a treatment that changes spreading can change surface coverage. Viability claims require an appropriate viability or functional readout.
Validating automated analysis
Inspect segmentation overlays across low and high densities, control and treated samples, and early and late time points. Compare a representative subset with an independently reviewed count or another suitable reference method. Check whether touching cells merge, single cells split, or debris is accepted as cells. Keep the analysis consistent across comparison groups unless a justified change is documented. AI-based identification may reduce manual optimisation, but performance still needs validation for the actual sample. Agilent offers an optional Gen5 AI cell identification module for compatible Cytation and Lionheart systems. [3]
Common questions
Can confluence be converted directly to cell number?
Only through an assay-specific relationship validated over the intended range. Cell area, shape, density and treatment can all change that relationship.
Are label-free cell counts suitable for every culture?
No. Contrast, density, morphology and sample background affect whether individual cells can be distinguished. A fluorescent nuclear or cellular marker may be more suitable for some samples.
Need help selecting a platform for your assay? Contact BioSPX to discuss the sample, readout and configuration you need.
References
[2] Agilent Measuring confluence using high contrast brightfield 5994 3446EN