Abstract
In rectal cancer, a pathologic complete response (pCR) to pre-operative treatment is a favourable prognostic marker, but is reported in a minority of the patients. We aimed at identifying microRNA-related host genetic polymorphisms predictive of pCR. A panel of 114 microRNA-related tagging polymorphisms was selected and analyzed on 265 locally advanced rectal cancer patients treated with neoadjuvant chemo-radiotherapy. Patients were stratified in two subgroups according to the radiotherapy dose (50.4Gy for 202 patients, 55.0Gy for 78 patients). Interactions among genetic and clinical-pathological variants were investigated by recursive partitioning analysis. Only polymorphisms with a consistent significant effect in the two subgroups of patients were selected as predictive markers of pCR. The results were validated by bootstrap analysis. SMAD3-rs744910, SMAD3-rs745103, and TRBP-rs6088619 were associated to an increased chance of pCR (p=0.0153, p=0.0471, p=0.0125). DROSHA-rs10719 and SMAD3-rs17228212 had an opposite detrimental effect on pathological tumour response (p=0.0274, p=0.0049). Recursive partitioning analysis highlighted that a longer interval time between the end of radiotherapy and surgery increases the chance of pCR in patients with a specific combination of SMAD3-rs744910 and TRBP-rs6088619 genotypes. This study demonstrated that microRNA-related host genetic polymorphisms can predict pCR to neo-adjuvant chemo-radiotherapy, and could be used to personalize the interval time between the end of radiotherapy and surgery.