Lung is a notoriously difficult organ for conducting reproducible image analysis, due to both the challenges of consistent histology processes in lung tissue and the heterogeneity of various normal and neoplastic pulmonary features.
Below is an example of how histology pattern recognition can be combined with both cell and area based image analysis. Pulmonary adenocarcinomas were stained with pS6-Ser240, and then analyzed with image analysis tools. Below is the first results with histology pattern recognition.
- Pulmonary adenocarcinoma IHC stained with pS6-Ser240 (left) and histology pattern recognition applied using Genie (right)
After pattern recognition has been run, then a choice must be made of conducting either cell-based or area-based image analysis. In IHC, this choice frequently depends on the pattern of the antibody staining. If the cells are distinct, then cell-based approaches are preferred. However, many times the staining is not distinct in individual cells, and an area based approach is better. When we validate a new antibody image analysis process, we frequently run both area and cell based analyses, and then graph the data to compare which approach is more scalable and stable across multiple slides. Below are images illustrating both approaches.
- Two approaches to IHC measurement post-pattern recognition. On the left is an area-based approach, and on the right is a cell-based approach (pS6-Ser240 antibody stain)
