Background and objectives: The advent of noninvasive methods of testing for colorectal cancer that have a high level of specificity and sensitivity is eagerly awaited.
Methods: Thirty patients with sporadic colorectal cancer and 11 patients with hereditary nonpolyposis colon cancer (HNPCC) enrolled in this study. We analyzed the loss of heterozygosity (LOH) in matched genomic DNA extracted from blood and surgical specimens (tumor and tumor-free colonic mucosa), and the corresponding oral rinse and stool specimens using seven microsatellite loci (APC, p53, DCC, hMLH1, D9S162, D9S171, and IFNA). To reduce the normal colonocyte DNA contamination of the stool samples, we compared three different methods for fecal genomic DNA extraction. As normal controls, we analyzed the LOH using the oral rinse and stool samples from 15 individuals without cancer.
Results: The LOH determined from the oral rinse and the stool samples matched those determined from the blood and the neoplastic tissue. All patients with HNPCC had microsatellite alterations at hMLH-1 in tumor DNA and corresponding fecal DNA. There were significant differences in the frequency of p53-LOH and D9S171-LOH between the group with sporadic disease and those with HNSCC (P = 0.0031 and 0.0294, respectively). Two cases with D9S162-LOH were detected in individuals without cancer. For the patients with sporadic disease, using p53 and adenomatous polyposis coli (APC), the sensitivity of the fecal DNA analysis was 96.7% (95% CI, 83-100) with a specificity of 100%.
Conclusion: We demonstrate that LOH analysis using oral rinse and stool samples may be a suitable screening tool for colorectal cancer.
Copyright 2002 Wiley-Liss, Inc.