Analytical validation of the Immunoscore and its associated prognostic value in patients with colon cancer.

BACKGROUND: New and fully validated tests need to be brought into clinical practice to improve the estimation of recurrence risk in patients with colon cancer. The aim of this study was to assess the analytical performances of the Immunoscore (IS) and show its contribution to prognosis prediction. METHODS: Immunohistochemical staining of CD3+ and CD8+ T cells on adjacent sections of colon cancer tissues were quantified in the core of the tumor and its invasive margin with dedicated IS modules integrated into digital pathology software. Staining intensity across samples collected between 1989 and 2016 (n=595) was measured. The accuracy of the IS workflow was established by comparing optical and automatic counts. Analytical precision of the IS was evaluated within individual tumor block on distant sections and between eligible blocks. The IS interlaboratory reproducibility (n=100) and overall assay precision were assessed (n=3). Contribution of the IS to prediction of recurrence based on clinical and molecular parameters was determined (n=538). RESULTS: Optical and automatic counts for CD3+ or CD8+ were strongly correlated (r=0.94, p<0.001 and r=0.92, p<0.001, respectively). CD3 and CD8 staining intensities were not altered by the age of the tumor block over a period of 30 years. Neither the position of tested tissue sections within a tumor block nor the selection of the tissue blocks affected the IS. Reproducibility of the IS was not affected by multiple variables (eg, antibody lots, DAB revelation kits, immunohistochemistry automates and operators). Interassay repeatability of the IS was 100% and interlaboratory reproducibility between two testing centers was 93%. Finally, in a case series of patients with stage II-III colon cancer, the relative proportion of variance for time to recurrence was greatest for the IS (53% of prognostic variability) in a model that included IS, T-stage, microsatellite instability status and total number of lymph nodes. CONCLUSION: IS is