Correlation Engine 2.0
Clear Search sequence regions


Determining the fraction of tumor cells in colon carcinoma samples analyzed for KRAS mutation status is important for choosing the proper testing modality and accurately interpreting the results. However, when asked to determine the tumor cell fraction in tissue samples, different pathologists give considerably different estimations, possibly leading to erroneous interpretation of KRAS mutation analysis results and poor treatment choices. To address this issue, we developed a free, easy-to-use computer program that estimates the tumor cell fraction on colon carcinoma slides that are immune-stained with anti-cytokeratin antibody. The program differentiates between the tumor area and the surrounding stroma on the basis of the immunostaining. Sixty such samples were evaluated by the program. In addition, the actual tumor cell fraction in these samples was measured. The tumor cell fraction estimated by the computer program showed a highly significant correlation with the actual measurements (R=0.64, P<0.001). In addition, we found that a short calibration step before beginning the computer estimation increased the accuracy of the results. In 4 cases (7%), there was some discrepancy between the computer estimation and the actual measurements; however, this was attributed to lower-quality immunohistochemical staining, indicating the importance of this phase in the analysis. In conclusion, we believe that this program can be used for standardizing the evaluation of the tumor cell fraction in colon carcinoma and that its use might aid in making better diagnosis and treatment choices for these patients.

Citation

Tova Hershkovitz, Avshalom Shenhav, Edmond Sabo, Ofer Ben-Izhak, Dov Hershkovitz. Development of a computerized morphometry application for assessment of the tumor fraction in colon carcinoma tissue samples. Applied immunohistochemistry & molecular morphology : AIMM / official publication of the Society for Applied Immunohistochemistry. 2013 Jan;21(1):54-8

Expand section icon Mesh Tags

Expand section icon Substances


PMID: 22595946

View Full Text