Abstract: Objective To explore the causal relationship between low grade serous ovarian cancer(LGSOC)and breast cancer(BC). Methods The study employed a bidirectional two sample Mendelian randomization(MR)method, and utilized publicly available genetic data from genome-wide association studies in European populations, with single nucleotide polymorphisms(SNPs)as instrumental variables. The forward study used LGSOC as the exposure variable and BC as the outcome, while the reverse study used BC as the exposure variable and LGSOC as the outcome. In addition, subgroup analysis was conducted on four molecular subtypes of BC. Inverse variance weighted(IVW), MR Egger regression, weighted median, weighted model, and simple model were used to study the causal relationship between LGSOC and BC. The Cochrans Q test was used for heterogeneity of instrumental variables. The MR Egger intercept method was used to test horizontal pleiotropy. Results The positive study included 19 SNPs, and MR analysis showed a positive causal relationship between LGSOC and increased risk of BC(IVW: OR=1.05, 95%CI: 1.01-1.08, P=0.01). Subgroup analysis showed that LGSOC may have a causal relationship with estrogen receptor positive(ER+)BC(P=0.07). The reverse study included 95 SNPs, and the MR analysis results showed that there was no causal relationship between BC and increased risk of LGSOC(P>0.05). No significant heterogeneity or horizontal pleiotropy was found in both heterogeneity and pleiotropy tests(P>0.05). Conclusion There is a positive causal relationship between LGSOC and increased risk of BC in Europeans, which may increase the risk of ER+BC. The causal relationship between BC and increased risk of LGSOC is not supported.

Key words: Low grade serous ovarian cancer, Breast cancer, Molecular subtype, Mendelian randomization, Causal relationship