ADCIS - Software products and custom development in the fields of image processing and analysis

The Stereology Toolkit implements long-standing and accepted stereology techniques that utilize a test grid overlaid on a slide image to characterize its microscopic structures. This extension's techniques can be easily applied to virtual slides acquired using an Aperio ScanScope^® scanning system in order to compute accurate estimates of biomarker parameters, including volume fraction and length density.

The Stereology Toolkit enables the user to define a region of interest (ROI) and a grid that overlay an area of a virtual slide. The type and spacing of the grid can be adjusted by the user, as shown beside, to achieve the best estimate of biomarker parameters in the ROI. After the ROI and grid are defined, the user manually highlights structures of interest (i.e., biomarkers) that are intersected by the test grid. The number of points is then used to automatically compute biomarker parameters using stereological statistical analyses. The extension's results can be exported into Aperio's Spectrum™ environment for display and further analysis of the prognostic and therapeutic impact of biomarkers.

The term stereology was first introduced in 1961 when the International Society for Stereology (ISS) was founded by a small group of scientists, although the basis of stereology theory was defined more than 200 years ago. By definition, stereology is the science that studies the geometrical relationship between a structure that exists in 3D space and a set of images of the same structure that are fundamentally defined in 2D space (images of sections or projections). Note that standard 2D image processing techniques will hardly provide 3D information from sections, except for the volume fraction value.

Pathologists and histologists can use the Stereology Toolkit's techniques to quickly compute unbiased and robust estimates of biomarker estimates on virtual slides. It is an effective alternative to complex image processing algorithms. In addition, while complex algorithms require careful validation of their results, stereology results require no validation since they are derived from a standard statistical analysis of manually counted points. When the structures of interest cannot be highlighted by specific staining (e.g., histochemical or immunohistochemical stainings), or when staining is not optimal, then the techniques of stereology are the best alternative for estimating the parameters of biomarkers.

When specific staining is effective, different image processing and analysis algorithms can be developed that vary in their complexity and efficiency. The Stereology Toolkit is a powerful tool for selecting the algorithm that best achieves accurate estimates of biomarker parameters. The regular sampling process used in stereology science helps overcome the difficult problem resulting from tissue heterogeneity (see adjacent figure). The special strength of the Stereology Toolkit is that it always provides unbiased estimates of biomarker parameters for any complexity of sample tissue. When combining stereology and image processing, the user has a broad set of powerful tools to characterize biomarkers on virtual slides.

• Grid

• Region of interest

• Undesired tissue

• Input

• Output

Figure 1: Region of Interest (ROI)
Define a region-of-interest consisting of one or multiple regions using rectangle, ellipse or freehand drawing tools.

Figure 2: Grid
Select the grid type, choosing from point counting and uniformly sampled squares with forbidden lines, and define the spacing.

Figure 3: Manual Classification
Highlight structures of interest (e.g. biomarkers) on the grid.

Figure 4: Stereology Results
The system automatically computes biomarker parameters such as volume fraction or numerical density of biomarker profiles per unit area based on stereological statistical analysis.