If this is not your name, click here.
| | If this is not your name, click here. | | |
| | Contact Us | Order Now | Journals | Bookstore | Register a colleague | | |
| |  Test Shows High Specificity, Sensitivity in Prognosis and Diagnosis of Colorectal Cancer: Presented at AACR-NCI-EORTC By Crina Frincu-Mallos, PhD SAN FRANCISCO, CA -- November 6, 2007 -- The first transcriptome-based, disease specific microarray for colorectal cancer permits the use of formalin-fixed paraffin-embedded tissue for identifying stage 2 colorectal cancer patients with high risk of tumour remission within five years of surgery, researchers reported here at the 2007 AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics. Peter Kerr, PhD, Scientific Program Manager, Almac Diagnostics, Craigavon, United Kingdom, presented the results obtained with a novel research tool, the Colorectal Cancer DSA (CCDSA), designed to give a prognostic signature for the stage 2 colorectal cancer population. The early development of CCDSA was first reported last year at the American Society for Clinical Oncology (ASCO) conference. Primary tumour specimens from 143 patients in 7 hospitals in Ireland, the United States, and the United Kingdom were collected after obtaining ethical approval, said Dr. Kerr and colleagues. The patients were not subjected to chemotherapy or radiation therapy, either before or after surgery, noted the researchers. On the basis of having disease-free survival five years after surgery, 87 samples were classified as "good prognosis", while 56 samples were labelled "poor prognosis" due to cancer recurrence within five years from surgery, indicated Dr. Kerr in his poster presentation. All analyses were performed on formalin-fixed, paraffin-embedded (FFPE) sample tissues. The 143 FFPE samples were randomly separated into a training set (n = 93, with 56 "good" and 37 "poor prognosis") and a test set (n = 50, with 31 "good" and 19 "poor prognosis"). Total RNA was extracted from 10 micron FFPE sample sections, one for each of the 143 specimens. Biotinylated aRNA targets were generated and hybridised to CCDSAs. For all arrays, gene expression indices were computed using invariant set normalisation and a perfect match-only model. The optimal number of prognostic transcripts was determined using 10-fold cross-validation, according to a known protocol (Ambroise et al., 2002, PNAS, 99: 6562). Dr. Kerr and colleagues derived a 91-transcript prognostic signature from the gene expression data, using the training set of samples (n = 93) described above. The receiver operating characteristic (ROC) curve of the 91 transcript prognostic signature shows a.83 area under the curve (AUC), indicating that the prognostic signature had a high predictive power for patients with stage 2 colorectal cancer, explained Dr. Kerr. Using the independent test set (n = 50), the 91 transcript prognostic signature shows a very good overall performance of the CCDSA research tool. The results indicate that 80% of patients would be correctly classified as having high versus low risk of relapse (80% accuracy), 79% of high risk patients would be correctly identified (79% sensitivity), and 81% of patients with low risk of relapse would be correctly classified (81% specificity), said Dr. Kerr. A hazard ratio (HR) of 7.9 (P <.0001) indicated a higher risk and significantly shorter time to remission in the patients with "poor prognosis" compared to those with "good prognosis," noted Dr. Kerr. Furthermore, applying the CCDSA tool to the whole sample specimens (N = 143), the analysis indicates that the 91 transcript signature would be able to identify a "poor prognostic" cohort with a 58% chance of tumour recurrence and a "good prognostic" cohort with 8% chance, within five years from surgery. Dr. Kerr and colleagues noted that the prognostic signature is currently being refined and would result in an in vitro diagnostic test for prognosis of tumour relapse in this patient population, given that the research tool would meet with regulatory approval. These patients assessed to be at high risk of relapse would benefit from adjuvant chemotherapy, concluded the researchers. Funding was provided by Almac Diagnostics Ltd., Craigavon, United Kingdom.
[Presentation title: Independent validation of a prognostic signature of relapse in Stage II colorectal cancer derived from formalin-fixed paraffin-embedded tissue. Abstract C138]
|
