This section conveys the basic understanding necessary for analyzing and interpreting RT-DC data. If you have the feeling that something is not covered here, please create an issue on GitHub.
In Real-time deformability cytometry (RT-DC), small objects, such as cells or beads, are flushed through a narrow channel at high speeds. The flow profile inside the channel causes hydrodynamic stresses that deform these objects [Mietke2015], [Mokbel2017]. For each event (detected object), an image is recorded and analyzed in real-time [Otto2015]. The analysis consists of the extraction of several features from the event images, such as deformation, projected area, or average brightness. These features can then be used to identify and characterize sample populations; For instance, major blood cells and their pathological changes in disease conditions [Toepfner2018].
A multitude of features can be extracted from the data recorded during an RT-DC measurement. These features are mostly computed live during data acquisition and are stored alongside the raw image data. Here, only the most important features are discussed. A full list of the features available in Shape-Out is maintained in the dclab documentation. Please note that some of the features are only available in expert mode.
Area and porosity¶
The area is the projected object area which is determined via the contour of the binarized event image. Shape-Out differentiates between two types of area, area of the measured contour (“Measured area [px]”) and area of the convex contour (“Convex area [px]” and “Area [µm²]”). The convex contour is the convex hull of the measured contour and enables a quantification of porosity (convex to measured area ratio). The porosity is often used for filtering, e.g. to remove high-porosity dirt particles in a preprocessing step.
A porosity of 1 means that the measured contour is convex. Note that the porosity can only assume values larger than 1. Also note that the convex contour/area is computed on the same pixel grid as the measured contour/area and is, as such, subject to pixelation artifacts.
The bounding box of an event image is the smallest rectangle (with its sides parallel to the x and y axes) that can hold the event contour. The aspect ratio of the bounding box is the rectangle’s side length along x divided by the side length along y. The size of the bounding box along x and y as well as its aspect ratio are often used for filtering.
Brightness within contour¶
Quantifying the brightness values within the image contour yields information on object properties such as homogeneity or density. For instance, it has been shown that the quantities “mean brightness” and “convex area” are sufficient to identify (and count) all major blood cells in a drop of blood [Toepfner2018].
In addition to the average brightness values, Shape-Out also has access to the standard deviation of the brightness in each image.
Deformation and elasticity¶
The deformation describes how much an event image deviates from a circular shape. It is defined via the circularity:
with the projected area \(A\) and the contour length of the convex hull of the event image \(l\). The contour length is computed from the convex hull to prevent an overestimation due to irregular, non-convex event shapes. It has been shown that the knowledge of deformation and area allows to derive a value for elasticity in RT-DC [Mietke2015] [Mokbel2017]. As a convenient measure for elasticity, isoelasticity lines are often employed to visualize stiffness.
Note that it is also possible to directly access the Young’s modulus in Shape-Out.
Real-time fluorescence and deformability cytometry (RT-FDC) records, in addition to the event images, the fluorescence signal of each event [Rosendahl2018]. The raw fluorescence data consists of the one-dimensional fluorescence intensity trace from which features such as peak fluorescence or peak width can be computed. For more advanced applications, RT-FDC also supports multiple fluorescence channels.
The inertia ratio is the ratio of the second order central moments along x and y computed for the event contour. Thus, the inertia ratio is a measure of deformation. In comparison to deformation, the inertia ratio has a low correlation to porosity. Shape-Out also allows to compute the principal inertia ratio which is the maximal inertia ratio that can be obtained by rotating the contour. Thus, the principal inertia ratio is rotation-invariant which makes it applicable to reservoir measurements where e.g. cells are not aligned with the channel. To quantify the alignment of the measured objects with the measurement channel, Shape-Out can additionally quantify the tilt of the contour relative to the channel axis.
Shape-Out can compute the volume from the event contour under the assumption of rotational symmetry, i.e. it is assumed that the projection of the cell volume onto the detector plane does not change when the cell is rotated, with a rotational axis parallel to the flow direction. The computation of the volume is based on a full rotation of the upper and the lower halves of the contour from which the average is then used [Halpern2002]. Volume has the advantage to be less correlated to deformation when compared to the projected area and it is therefore a better measure of cell size in the channel.
|||(1, 2) Detection Of Human Disease Conditions By Single-Cell Morpho-Rheological Phenotyping Of Whole Blood by Toepfner et al., licensed under CC BY 4.0 [Toepfner2018].|