OBJECTIVE: To analyze intratumoral heterogeneity of immune cells and the resulting impact of heterogeneity on the level of individual patient prediction. STUDY DESIGN: Using whole slide imaging by virtual microscopy, we present the first spatial quantitative study of immune cells in a set of colorectal cancer primary tumors. We generated "tumor maps" based on cell densities in fields of 1 mm2, visualizing intratumoral heterogeneity. In this example, cutoffs of marker-based cell stains identified by tissue microarray (TMA) led to ambiguous decisions in 11 of the 20 patients studied. Classic TMA analysis can be used in large patient cohorts to generate clinically significant predictors. The transfer of these predictors from large-scale TMA to individualized predictions thus far has not been investigated. In colorectal cancer, TMA-based quantitative immune cell counts using immune cell surface molecules (CD3, CD8, Granzyme B, and CD45RO) have been shown to be potentially better predictors for patient survival than the classical TNM system. RESULTS: Our results make clear that for individualized prognostic evaluations, whole slide imaging by virtual microscopy is irreplaceable during identification of prognostic markers as well as in their subsequent application. CONCLUSION: In the future, spatial marker signatures could contribute to individual patient classifiers.
OBJECTIVE: To analyze intratumoral heterogeneity of immune cells and the resulting impact of heterogeneity on the level of individual patient prediction. STUDY DESIGN: Using whole slide imaging by virtual microscopy, we present the first spatial quantitative study of immune cells in a set of colorectal cancer primary tumors. We generated "tumor maps" based on cell densities in fields of 1 mm2, visualizing intratumoral heterogeneity. In this example, cutoffs of marker-based cell stains identified by tissue microarray (TMA) led to ambiguous decisions in 11 of the 20 patients studied. Classic TMA analysis can be used in large patient cohorts to generate clinically significant predictors. The transfer of these predictors from large-scale TMA to individualized predictions thus far has not been investigated. In colorectal cancer, TMA-based quantitative immune cell counts using immune cell surface molecules (CD3, CD8, Granzyme B, and CD45RO) have been shown to be potentially better predictors for patient survival than the classical TNM system. RESULTS: Our results make clear that for individualized prognostic evaluations, whole slide imaging by virtual microscopy is irreplaceable during identification of prognostic markers as well as in their subsequent application. CONCLUSION: In the future, spatial marker signatures could contribute to individual patient classifiers.
Authors: Anna M Sherwood; Ryan O Emerson; Dominique Scherer; Nina Habermann; Katharina Buck; Jürgen Staffa; Cindy Desmarais; Niels Halama; Dirk Jaeger; Peter Schirmacher; Esther Herpel; Matthias Kloor; Alexis Ulrich; Martin Schneider; Cornelia M Ulrich; Harlan Robins Journal: Cancer Immunol Immunother Date: 2013-06-16 Impact factor: 6.968
Authors: Jason Hipp; Jerome Cheng; Liron Pantanowitz; Stephen Hewitt; Yukako Yagi; James Monaco; Anant Madabhushi; Jaime Rodriguez-Canales; Jeffrey Hanson; Sinchita Roy-Chowdhuri; Armando C Filie; Michael D Feldman; John E Tomaszewski; Natalie Nc Shih; Victor Brodsky; Giuseppe Giaccone; Michael R Emmert-Buck; Ulysses J Balis Journal: J Pathol Inform Date: 2011-10-29