Many factors influence patient response to immunotherapy. In particular, a tumor’s pre-existing immune contexture, or its immunophenotype, is an important consideration. By characterizing the immune cell infiltration into the tumor, it is possible to distinguish three basic immune profiles: a desert (no to very low infiltration), an excluded (infiltrates exclusively in the invasive margin but not in the tumor; or, infiltrates excluded from the intra-epithelial compartment but abundant in the intra-tumoral stroma) and an inflamed phenotype (infiltration within the intra-epithelial compartment). Below are examples of these phenotypes (CD8 expressing T cells in browncarcinoma cells in purple).
(Click or tap on images below for a higher resolution.)

However, most patient tumors rarely present exclusively as one of these phenotypes. Rather, there is heterogeneity that these categories fail to describe comprehensively. Below is an example of core biopsies from a single patient that present all 3 phenotypes. (CD8 expressing T cells in brown.)

Even the use of image analysis tools presented as a single metric to describe the above phenotypes may be insufficient to clearly describe the immune contexture. For example, a large resection containing a small but biologically relevant inflamed region, might nonetheless present low values under semi-automated analysis since the measured densities are typically averaged out over the entire tumor region. 
One way to provide greater phenotype resolution is to describe the underlying immune cell heterogeneity by cataloging the density of the cells of interest as well as their location. The Density Proportion Score (DPS) was developed at HistoGeneX by our pathologists to accommodate the heterogeneous complexity of the tumor microenvironment’s immune contexture. In this method, the pathologist estimates the area of the tumor occupied by positive immune cells at a specific density, ranging from 0 to 3. At density 0, the area contains no to very low densities of immune cells, whereas at density 3, the area has a very high density of positive immune cells. The pathologist conducts this review for both the intra-tumoral stroma (ITS) and the intra-epithelial (IE) compartments. Let’s have a look at an example of a sample stained with our validated panCK-CD8 IHC test. The panCK-positive carcinoma cells stain purple and comprise the intra-epithelial compartment. CD8-positive T cells are stained with brown DAB.

The score can be understood in the following ways. For example, the stroma:

  • 45% of the intra-tumoral stroma contains little to no CD8-positive cells
  • In 25% of the stroma, the CD8-positive cell density is 1 or ‘moderate’
  • In other words, the CD8-positive cell density is at least moderate in 55% of the stroma (ITS1 + ITS2 + ITS3)
  • In 20% of the stroma, the CD8-positive cell density is 2 or ‘high’
  • I.e, in 30% of the stroma, the CD8-positive cell density is 2 or higher 
  • 10% of the tumor stroma is occupied by an area with a very high CD8 cell density

A similar interpretation can be made for the intra-epithelial compartment. Thus, the DPS provides numerical values to describe the immune contexture of the tumor microenvironment.

The training of this method relies on a curated library of reference images showing ITS and IE density examples. We use our PathoTrainer system to train and establish proficiency among pathologists. The DPS-trained pathologists report a strong level of concordance and reproducibility when trained on the DPS.

We used CD8 as the biomarker in this newsletter to illustrate the use of the DPS. Similarly, the DPS can be applied for other relevant biomarkers on immune cells such as LAG3 and TIM3. By systematically applying the DPS, pathology-driven observations are provided in a database-friendly format that can be used to study meaningful reading cutoffs and the impact of heterogeneity of the patient sample.

Another way to characterize the immune contexture is through semi-automated analysis of the CD8 cells by software. We have developed an image analysis algorithm (Visiopharm) to measure CD8-positivity in the intra-epithelial compartment versus the intra-tumoral stroma and in the invasive margin vs. the central tumor region. Outputs can be delivered as relative marker areas or as cell densities.
These density values that are obtained via software analysis serve to complement the DPS values assigned by pathologists. HistoGeneX is currently curating both software-driven density values and pathologist-derived DPS metrics across multiple tumor types. Additionally, we have also developed an application that correlates these values to patient response.
For more information, please This email address is being protected from spambots. You need JavaScript enabled to view it..