Clinical and technical validation of a genomic classifier (ColoPrint) for predicting outcome of patients with stage II colon cancer

Publication Name: 2012 ASCO Gastrointestinal Cancers Symposium, JCO Abstract

Author(s): Josep Tabernero, Victor Moreno, Robert Rosenberg, Ulrich Nitsche, Thomas Bachleitner-Hofmann, Giovanni Lanza, Jeroen van den Akker, Paul Roepman, Iris Simon, Ramon Salazar; Vall d'Hebron University Hospital, Barcelona, Spain; IDIBELL, Institut Catala d'Oncologia, L'Hospitalet de Llobregat, Spain; Department of Surgery, Klinikum Rechts der Isar, Technische Universität München, Munich, Germany; Medizinische Universitaet Wien, Vienna, Austria; Department of Experimental and Diagnostic Medicine, Section of Anatomic Pathology, Ferrara, Italy; Agendia BV, Amsterdam, Netherlands; Institut Catala d'Oncologia, Barcelona, Spain

Background: Between 25 and 35% of stage II colon cancer patients will experience a relapse of their disease and may benefit from adjuvant chemotherapy. ColoPrint is a gene expression classifier that can predict disease relapse in patients with early-stage colorectal cancer (Salazar et al. JCO 2011). Methods: ColoPrint was developed using gene expression data from whole genome microarrays and was validated in in-silico datasets and independent patient cohorts from 5 European hospitals. Fresh frozen tissues, clinical parameters, MSI-status and follow-up data for patients were available. Samples were hybridized to Agilent microarrays and the ColoPrint index was determined. Uni-and multivariate analysis was performed on the pooled stage II patient set (n=320) who had a median follow-up of 70 months. Reproducibility and precision studies were performed using clinical and control samples specific for each outcome level (high risk, low risk). Experiments were performed on 20 days with 2 runs per day by multiple technicians reflecting daily diagnostic conditions. Results: Performance of the prognostic classifier was confirmed in reproducibility and stability assays and stringent quality controls were established. In the clinical validation, ColoPrint classified two-third of patients (209/320) as low risk. The 3-year relapse-free survival was 94% for low risk patients and 79% for high risk patients with a HR of 2.74 (95% CI 1.54-4.88; p=0.006). MSI-status and the number of assessed lymph nodes were the only significant clinical parameters in the univariate analysis. Using parameters from the ASCO recommendation (T4, perforation, less than 12 LN assessed and/ or high grade) for the identification of high risk patients was not significant (HR 1.43, 95% CI 0.81-2.55; p= 0.22) and no clinical parameter added power to the ColoPrint classification in multivariate analysis. Conclusions: ColoPrint is available as a routine diagnostic test with a high precision and reproducibility. ColoPrint significantly improves the prognostic accuracy of pathologic factors and MSI and facilitates the identification of low risk patients with stage II disease who may be safely managed without chemotherapy.

See Article