DNA repair gene polymorphisms, tumor control, and treatment toxicity in prostate cancer patients treated with permanent implant prostate brachytherapy.

Background Radiotherapy and brachytherapy are common treatments for localized prostate cancer (PCa). However, very few studies evaluated the association of variations in DNA damage response genes and treatment outcomes and toxicity in brachytherapy-treated patients. Purpose To evaluate the association of inherited germline variations in DNA repair-associated genes with tumor control and treatment toxicity in patients treated with low-dose-rate prostate brachytherapy (LDRB). Material and Methods The cohort consists of 475 I-125 LDRB patients with a median follow-up of 51 months after seed implantation. Patients were genotyped for 215 haplotype tagging single nucleotide variations (htSNPs) in 29 candidate genes of DNA damage response and repair pathways. Their association with biochemical recurrence (BCR) was assessed using Cox regression models and Kaplan-Meier survival curves. Linear regressions and analysis of covariance (ANCOVA) between early and late International Prostate Symptom Score (IPSS) with htSNPs were used to evaluate the association with urinary toxicity. Results After adjustment for the established risk factors, six htSNPs in five genes were found to be significantly associated with an altered risk of BCR, with adjusted hazard ratios (HRadj.) ranging between 3.6 and 11.1 (P < .05). Compared to carriers of the ERCC3 rs4150499C allele, patients homozygous for the T allele (n = 22) had a significant higher risk of BCR with a HR of 11.13 (IC95 = 3.9-32.0; P < .0001; q < 0.001). The Kaplan-Meier survival curve revealed a mean BCR-free survival time reduced from 213 +/- 7 to 99 +/- 12 months (log-rank P < .0001) for homozygous T carriers compare to noncarriers. For late IPSS (>6 months after treatment), htSNP rs6544990 from MSH2 showed a statistically significant b-coefficient of 1.85 +/- 0.52 (P < .001; q < 0.1). Homozygous carriers of the MSH2 rs6544990C allele (n = 62) had a mean late IPSS 3.6 points higher than patients homozygous for the A allele (n = 132). This difference was significant when tested by ANCOVA using pretreatment IPSS as a covariate (P < .01). Conclusions This study suggests an association of the intronic variants of the DNA nucleotide excision repair ERCC3 and DNA mismatch repair MSH2 genes with elevated risk of BCR and late urinary toxicity respectively after LDRB. Further validation is required before translational clinical advances.