Tumor location and detection of k-ras mutations in stool from colorectal cancer patients.

Despite the potential effectiveness of colonoscopy and fecal occult blood tests in reducing colorectal cancer mortality, more specific, noninvasive tests are desirable. In this regard, it is possible to detect specific mutations of several genes in stool samples from colorectal cancer patients. Recently, the feasibility of a multitarget assay has been explored (1,2). Because the features of the mutations in the adenomatous polyposi coli (APC) and tumor protein 53 (TP53) make their study in stool difficult and time-consuming (2,3), efforts have focused on detecting the K-Ras gene, which often mutated in patients with colorectal cancer. We analyzed 31 colorectal tumor samples by polymerase chain reaction– restriction fragment length polymorphism (PCR–RFLP) and DNA sequencing and found 16 with K-Ras mutations. No associations were found between these mutations and clinicopathologic variables. To analyze K-Ras mutations in paired stool specimens from the same patients, DNA was extracted from both liquid and solid stool, checked for appropriate quality in PCR, and used in enriched-PCR–RFLP analysis, which has a verified sensitivity of detecting one mutant out of 10 000 wild-type alleles (4). No K-Ras mutations were detected in fecal DNA from liquid stool samples. The same K-Ras mutation was detected in DNA from nine solid stool samples (9/14 62%) and in paired tumors (Table 1). Interestingly, mutations were detected in DNA from solid stool only when the paired tumor was located in the sigmoid colon or the rectum (four of four, and five of five, respectively). No K-Ras mutation was found in DNA from solid stool when the paired tumor was in the ascending colon (none of five; two-sided Fisher’s exact test; P<.001). Our results are consistent with those recently reported (2), where no K-Ras mutations were found in stool samples from patients with tumors in the ascending colon, but K-Ras mutations were found in eight of 13 (61.5%) stool samples from patients with tumors located in the descending colon. We conclude that the possibility of detecting K-Ras in fecal DNA increases when the related tumor is located distally. However, other molecular markers, such as minichromosome maintenance protein 2 (MCM2), locus BAT26, TP53, and APC (2,5,6), could be detected in stool from patients with cancers in the right side of the colon. The difficulty of assessing K-Ras mutations in stool from patients with tumors in the ascending colon may be explained by the possibility that mutant K-Ras alleles are more susceptible than other sequences to DNases present in stool (2), especially considering that their transit through the intestinal lumen is longer than that of DNA released in the rectum (i.e., they have a longer exposure to DNases). These observations are in keeping with the fact that, to our knowledge, K-Ras mutant alleles in fecal DNA from patients with cancers of the ascending colon were found only when a radioactive probe was used (7), suggesting that a technique with high sensitivity can still detect a low amount of nondegraded fecal DNA. In conclusion, our data suggest that the successful assessment of K-Ras mutations in stool mainly depends on tumor location. These results may explain why stool samples from patients with cancer in the ascending colon were negative, despite the presence of K-Ras mutations in the paired tumors. These conclusions may be relevant in the conTable 1. Analysis of K-Ras mutations in tumor and paired stool samples*