An improved image analysis method for cell counting lends credibility to the prognostic significance of T cells in colorectal cancer

Numerous immunohistochemically detectable proteins, such as immune cell surface (CD) proteins, vascular endothelial growth factor, and matrix metalloproteinases, have been proposed as potential prognostic markers in colorectal cancer (CRC) and other malignancies. However, the lack of reproducibility has been a major problem in validating the clinical use of such markers, and this has been attributed to insufficiently robust methods used in immunohistochemical staining or its assessment. In this study, we assessed how computer-assisted image analysis might contribute to the reliable assessment of positive area percentage and immune cell density in CRC specimens, and subsequently, we applied the computer-assisted cell counting method in assessing the prognostic value of T cell infiltration in CRC. The computer-assisted analysis methods were based on separating hematoxylin and diaminobenzidine color layers and then applying a brightness threshold using open source image analysis software ImageJ. We found that computer-based analysis results in a more reproducible assessment of the immune positive area percentage than visual semiquantitative estimation. Computer-assisted immune cell counting was rapid to perform and accurate (Pearson > 0.96 with exact manual cell counts). Moreover, the computer-assisted determination of peritumoral and stromal T cell density had independent prognostic value. Our results suggest that computer-assisted image analysis, utilizing freely available image analysis software, provides a valuable alternative to semiquantitative assessment of immunohistochemical results in cancer research, as well as in clinical practice. The advantages of using computer-assisted analysis include objectivity, accuracy, reproducibility, and time efficiency. This study supports the prognostic value of assessing T cell infiltration in CRC.