![]() The Nucleus/Cell area ratio incorporates both of these characteristics in a single measurement. This tends to be higher for tumor cells, because tumor nuclei tend to be larger than non-tumor nuclei, and more densely packed. The measurement can be selected from a list, and sliders can be used to adjust how colors are mapped to measurement values.Ī particularly useful measurement for the purposes of tumor cell identification is the Nucleus/Cell area ratio. This creates a kind of ‘heatmap’ visualization, in which each cell is color-coded according to its value for a particular measurement. However, another way to visualize cell measurements is by using the Measure ‣ Show measurement maps command. One way to view the measurements is by generating a results table, as described in Cell detection. QuPath’s ability to distinguish between different cell types depends upon which measurements have been made. The resulting cell detection is shown below.ĭetected cells View cell measurements (if you want) cells being cut in half, or detected twice). In this case, QuPath will overlap the regions and then try to resolve cells detected on region boundaries to avoid weird artefacts in these areas (e.g. This improves the speed and reduces the memory requirements. If the annotation is large enough, QuPath will break it into smaller regions that it can process in parallel. Therefore there is no need to look at staining intensity now.įor that reason, I’ve used the slightly-simpler Cell detection command. Since in this case we need to classify cells as tumor or non-tumor first, we will postpone considering staining intensity until the end, whenever we know the cell types. This is most useful if all detected cells should be considered the same. Positive cell detection does exactly the same thing as Cell detection, but has the extra step of classifying all cells as positive or negative immediately according to staining intensity. It does not really matter which command is applied in this case. In Cell detection, we used the Analyze ‣ Cell detection ‣ Positive cell detection command. They can be applied for the classification of all detections within QuPath, and not only for classifying different cell types.Ĭell detection or Positive cell detection? Instead, all cells can be detected, and then QuPath can be requested to calculate scores based only on the cells that are relevant for the application – automatically identifying and excluding the others.Īs before, the concepts described in this section are general within QuPath. This provides an alternative method of analysis that avoids the requirement to laboriously draw around every region that should be scored. This section builds upon this by showing how QuPath can be trained to distinguish between different cell types itself. These regions had to be drawn very carefully to try to ensure that they only included tumor cells, and excluded other cell types that should not contribute to conventional scoring of Ki67. Cell detection looked at computing Ki67 labelling indices by counting positive and negative cells within user-defined regions of interest.
0 Comments
Leave a Reply. |