Pharmacogenetic analysis in neoadjuvant chemoradiation for rectal cancer: high incidence of somatic mutations and their relation with response.

Aims: The identification of predictive markers of response to chemoradiotherapy treatment remains a promising approach for patient management in order to obtain the best response with minor side effects. Initially, we investigated whether the analysis of several markers previously studied and others not yet evaluated could predict response to 5-fluorouracil- and capecitabine-based neoadjuvant treatment in locally advanced rectal cancer.

Methods & Materials: We studied germline and tumoral samples of 65 stage II/III rectal patients. A panel of pharmacogenetic markers was genotyped in paired peripheral blood samples and rectal cancer tumors.

Results: Our results seem to confirm the previously described association of thymidylate synthase and the prediction of chemoradiotherapy response in rectal cancer. However, it failed to confirm the clinical utility proposed for XRCC1, ERCC1, ERCC2, MTHFR and EGFR polymorphisms in blood/germline samples. Subsequently, with the aim of improving prediction of individual response and assessing the role of studied polymorphisms in response to treatment, we determined if changes in tumor response to these markers could predict clinical outcome. We found a high degree of changes between germline and tumor samples, mainly somatic mutations without microsatellite instability, and a minor frequency of loss-of-heterozygosity events. In tumoral samples, XRCC1 appeared to be significantly associated (p = 0.006) with downstaging of the tumor (odds ratio: 7.93; 95% CI: 1.03-60.83), but the increasing of TYMS low-expression alleles contradict the previous results observed in germline samples.

Conclusion: The detection of somatic mutations in rectal cancer tumors led us to re-evaluate the utility of the tests performed in blood samples for these polymorphisms in rectal cancer. Furthermore, studies aimed at assessing the influence of pharmacogenetic markers in treatment response performed in blood samples should take into account the particular pattern of hypermutability present in each tumor type. We hypothesize that different patterns of hypermutability present in each tumor type would be related to the different results in association studies related to response to the treatment.