Detection of DNA stool mutations in colorectal cancer patients.

Abstract

e22074 Background: Colorectal cancer (CRC) is the third leading cause of cancer in the world. Noninvasive screening tests have higher compliance. Changes in tumoral cells are also found in the stools after cancer cell exfoliation. Aim: to detect mutations in the stools of CRC patients and compare to mutations found in CRC tissue Methods: Patients referred for colonoscopy were divided into 3 groups: control, cancer, and polyp. The DNA was extracted from tissue and stool and it was analysed by two microarrays, the KRAS/BRAF/PIK3CA Array and the RanplexCRC Array (Randox Laboratories Ltd). Both based on the simultaneous detection of 20 mutations in KRAS, BRAF and PIK3CA and 28 mutations in KRAS, BRAF, TP53 and APC respectively. Positive tissue sample mutations were confirmed by Sanger sequencing. Results: 90 stools and 112 tissue biopsies were studied. Analysis of tissue samples was initially performed using the KRAS/BRAF/PIK3CA Array. In total mutations were detected in 16(55%) CRC, 12(45%) polyp and 2(6%) control samples. Mutations were confirmed via Sanger sequencing (76%). In the stool the frequency were 17(35%) in CRC, 10(71%) in polyp and 1(4%) in the control. Inconclusive tests were higher among the controls (32.5%) and polyps (22%), being less frequent in the CRC patients (6%). Excluding the inconclusive tests, WT genes were common in the controls (96%). The most common mutation was within KRAS codon 12. Stools samples from 48 subjects were compared to the results obtained by the RanplexCRC Array and mutations were found in 50% of CRC. KRAS/BRAF/PIK3CA Array stool results had a correlation with RanplexCRC Array in 64% of CRC patients and in 47% of polyp patients. The average human DNA in the stools was 15ng/ul in CRC patients and 0.46ng/ul in the controls (p=0.0001). Conclusions: Both arrays had a good correlation and a sensitivity between 30-50% with a high specificity (95-100%). The KRAS/BRAF/PIK3CA Array is currently optimized and recommended for assessment of tissue samples. This initial set of data however does indicate its potential applicability to stool specimens. Inconclusive tests are likely as a result of limited human DNA content within stool control samples. FAPESP (Sao Paulo Research Foundation) project 09/16618-8.