4685

Genetic variation in the androgen receptor gene (AR) may be associated with endometrial cancer risk based on AR’s role in regulating androgen levels. However, results from four small case-control studies reported inconsistent associations. Two small (N=58 and 79 cases) studies suggested a positive association between CAG trinucleotide repeat length and endometrial cancer risk, while two nested case-control studies (N=130 and 222 cases) showed a decreased risk. The latter studies also examined AR haplotypes and found no significant association with endometrial cancer risk. In a population-based case-control study of 497 cases and 1024 controls conducted in the 2 largest cities of Poland (Warsaw and Lodz), we examined the CAG repeats and six haplotype-tagging single nucleotide polymorphisms (htSNPs: rs962458, rs6152, rs1204038, rs2361634, rs1337080, and 1337082), which cover an estimated 80% of the common variation in AR. Cases were identified through participating hospitals in Lodz and Warsaw and through the local cancer registries to ensure complete incident case ascertainment. Eligible controls (frequency-matched to cases by study site and age in 5-year categories) were identified through a database of all Polish residents during case accrual. Response rates for the in-person interview were high among both eligible cases (79.3%) and controls (67.7%). CAG repeats were not significantly associated with endometrial cancer based on either number of alleles (OR (95%CI) per allele with repeat size >22=0.97 (0.83 - 1.13); p-trend=0.68) or average CAG repeat length (OR (95% CI) per unit increase in average length=1.04 (0.97 - 1.12); p-trend=0.29). Three correlated htSNPs (rs6152, rs1204038, and rs1337082; D’> 0.93, r2>0.9), however, were associated with increased risk for endometrial cancer: OR (95% CI)=1.13 (0.89-1.44) for heterozygous and OR (95% CI)=2.40 (1.28-4.51) for homozygous variant, compared to common homozygous genotype, for rs6152 (p-trend=0.023). We found similar associations with the two correlated htSNPs. Results were similar for endometrioid tumors as for other tumor types (using logistic regression analyses restricted to cases; p-heterogeneity>0.10). Haplotype analyses did not reveal additional associations with endometrial cancer than observed in individual htSNP analyses. In conclusion, our findings suggest that while the number of CAG repeats is not associated with endometrial cancer, other common genetic variation of AR in particular a region of the AR marked by rs6152, rs1204038, or rs1337082, may be associated with endometrial cancer risk, although additional data is need to confirm these findings.

99th AACR Annual Meeting-- Apr 12-16, 2008; San Diego, CA