Background: Numerous pharmacogenetic studies have identified genetic polymorphisms (SNPs) associated with an increased risk of anthracycline-mediated cardiotoxicity (AMC). The purpose was to search for SNP-SNP interactions associated with the risk of cardiotoxic manifestations caused by anthracycline therapy in breast cancer patients. Methods: The study included 256 Caucasian patients (median age - 55 years) with a diagnosis of breast cancer with normal cardiovascular system parameters at baseline, who were treated with anthracyclines in 2019-2020. For SNP genotyping of c.4544G > A rs8187710 (ABCC2), c.1744C > T rs11549465 (HIF1A), g.22125504G > C rs1333049 (CDKN2A/B) and c.214T > C rs4673 (CYBA). DNA was extracted from blood by using DNA-sorb-B (AmpliSens, Russia). HRM-PCR was performed on a CFX96 amplifier (Bio-Rad, USA). The presence of polymorphism was confirmed by the Sanger method on a Genetic Analyzer 3500 (ABI, USA). Results: During the follow-up period 21 (8.2%) patients developed signs of subacute (changes developed within several weeks after the last course of therapy) or early chronic anthracycline mediated cardiotoxicity (changes developed within a year after completion of anthracycline therapy). Using the multifactorial dimension reduction method. We obtained a 4-locus model of SNP-SNP interactions c.4544G > A rs8187710 (ABCC2) x g.22125504G > C rs1333049 (CDKN2A/B) x c.214T > C rs4673 (CYBA) x g.23708527G > A rs28714259 which has high prognostic properties. The specificity of the test based on the 4-locus model was 68%, the sensitivity was 100%, and the overall accuracy was 71%. The results of the analysis of SNP-SNP interactions indicate the greatest contribution of rs4673 and rs28714259 to the predisposition to AMC. The first ones yet into antagonistic interactions with rs28714259, rs1333049 and rs11549465, the second with rs8187710, rs11549465 and rs4673. On the contrary rs11549465 and rs1333049 contribute to a synergistic effect. Conclusions: The 4-locus model discovered in this study can form the basis of prognostic tests that predict early the risks of developing AMC in cancer patients, which in the future allow to personalize and select the most optimal treatment regimen.