S1939 Prospective Analysis of the Role of CA 19-9 Levels Before and After Endoscopic Decompression of Biliary Obstruction in Differentiating Benign From Malignant Etiologies of Biliary Obstruction

Abstract

Introduction: The whole genomic microarrays can contribute to high-throughput biomarker screening in colorectal diseases, but the further possible diagnostic utilization of the markers requires testing their classificatory efficacy on an independent sample set. Aims: Our aims were to identify characteristic transcript sets in order to develop diagnostic mRNA expression patterns for objective classification of benign and malignant colorectal diseases and to test their classificatory power on an independent biopsy set. Material and methods: Gene expression profiles were evaluated on HGU133plus2 microarrays from colonic biopsies of 22 patients with colorectal cancer (CRC), 20 with adenoma, 21 with IBD and 11 healthy controls. For classification Prediction Analysis of Microarrays were used. The previously determined classificatory transcript sets were tested on independent samples (27 CRC, 29 adenoma, 28 IBD and 38 normal controls). Array real-time PCR validation was done on 68 independent biopsy specimen (24 CRC, 24 adenoma, 20 normal controls). Receiving operating characteristic (ROC) curve analysis was used to evaluate the discriminatory power of the gene panels. Results: Between normal and CRC samples 38 classifiers were identified by microarray analysis including upregulated CXCL1 and CXCL2 oncogenes, osteopontin and downregulated carbonic anhydrase 7. According to these classifiers, the independent CRC and normal biopsies could be clearly separated by 97.4% specificity and 96.3% sensitivity. Adenoma and normal samples could be classified using 20 discriminatory transcripts such as overexpressed cadherin 3, KIAA1199, forkhead box Q1 and downregulated claudin 8, peptide YY. Using these classifiers, independent adenoma and healthy samples could be distinguished with 94.7% specificity and 100% sensitivity. The microarray results were confirmed on independent biopsies using real-time PCR cards. Discriminatory power of the CRC vs. normal gene panel is proved to be considerably high in array real-time PCR (sensitivity: 91.7 %, specificity: 95.0 %). According to the real-time PCR results, adenoma and healthy samples could be clearly separated, only 2 of the 24 adenoma samples were grouped into the normal cluster. The ROC curve analysis showed 95.8% sensitivity and 100% specificity values. Conclusion: Discriminatory transcripts were identified which could correctly classify CRC and adenoma biopsies also on a large independent sample set. These markers can establish the basis of gene expression based diagnostic classification of benign and malignant colorectal diseases. Diagnostic real-time PCR cards can become part of the automated routine procedure.